正在加载视频...

视频加载失败

返回首页

Made a small Blender animation tool because manually offsetting keyframes across many objects was painful xd IIF - Multi Object Keying & Sequencer Linear/Radial/Wave key offsets + multi-object keying tools 👀 Gumroad: #blender #b3d #motiondesign

IIF_Munkhjin's profile picture

IIF // MUNKHJIN

1,167 subscribers

14,916 次观看 • 3 个月前 •via X (Twitter)

#blender#b3d#motiondesign
LIVE

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Thoughts on transferable skills. I was surprised reading the comments to Riley’s question about best UI animation tool for web. To anyone interested in diving into motion design I would strongly recommend - aim to become a motion expert, not Jitter/Rive/Lottie specialist. If you start from the foundations, jumping into any animation tool will feel like a walk in the park. Keyframes, easing functions, transformations - same concepts are shared between all of them. Getting to know the foundations on the web means starting with CSS, Tailwind or React Motion. I’m often creating small motion concepts in Figma (see video) and then recreating them manually in code. Any language, it doesn’t matter because principles are the same. I recently made the switch to Swift and built a small animation tool for myself (see video) that I’ll discard after I’m done with the motion. This level of comfortableness with motion gives you unmatched flexibility. From that point, tools are just interfaces for expressing something you already deeply understand. This is not about performance or technical optimizations (Lottie files can outperform CSS equivalents), tools mentioned above really are great - it’s about developing transferable creative skill that lets you think across any medium, platform or tool.
0:48

Thoughts on transferable skills. I was surprised reading the comments to Riley’s question about best UI animation tool for web. To anyone interested in diving into motion design I would strongly recommend - aim to become a motion expert, not Jitter/Rive/Lottie specialist. If you start from the foundations, jumping into any animation tool will feel like a walk in the park. Keyframes, easing functions, transformations - same concepts are shared between all of them. Getting to know the foundations on the web means starting with CSS, Tailwind or React Motion. I’m often creating small motion concepts in Figma (see video) and then recreating them manually in code. Any language, it doesn’t matter because principles are the same. I recently made the switch to Swift and built a small animation tool for myself (see video) that I’ll discard after I’m done with the motion. This level of comfortableness with motion gives you unmatched flexibility. From that point, tools are just interfaces for expressing something you already deeply understand. This is not about performance or technical optimizations (Lottie files can outperform CSS equivalents), tools mentioned above really are great - it’s about developing transferable creative skill that lets you think across any medium, platform or tool.

Natko Hasic

64,045 次观看 • 1 年前

I started using Blender through MCP about two weeks ago, and I quickly realized that you can build almost anything with AI. This model was created using Blender, Hunyuan3D, Gemini, and ChatGPT. Here’s how I did it: I opened Gemini, uploaded an image of the Gundam model, and asked it to generate a clean front-view image. I uploaded that front view to ChatGPT and asked it to generate two additional angles: a back view and a 45-degree front-left view. I went to: I signed up with my email and translated the page into English. Then I opened Image to 3D and selected the multi-image option. You’ll see a diagram of a whale from several angles. Upload each reference image in its corresponding position, such as front, 45-degree front-left, and back. I selected the 1.5M-face option. This produces a very high-poly model, but don’t worry, we’ll fix that next. Once the generation is complete, download the model as a GLB file. From the Hunyuan homepage, open 3D Studio using one of the dropdown menus. Select the topology or retopology tool and upload your GLB. I chose the High setting to preserve as much detail as possible. After a few seconds, the model was retopologized. It kept most of its visual detail while using far fewer polygons. The original head and rifle didn’t look very good, so I generated them separately. I returned to ChatGPT and created dedicated reference images for the Gundam’s head and rifle. I generated each part individually in Hunyuan at the 1.5M setting, then ran both through the same retopology process. Next came the textures. Open the texture section, select the multi-image option, and upload the same reference images according to the whale orientation indicators. However, instead of using the original clean textures, I asked ChatGPT to recreate them with wear, rust stains, scratches, and other surface damage. This gave the Gundam a much older and more authentic appearance. Once every part was textured, I imported everything into Blender. You can ask Codex through MCP to remove the original head and rifle, or you can do it manually. Select the main model and press Tab to enter Edit Mode. Press 3 to enable face selection, then press C to activate Circle Select. Paint over the faces belonging to the original helmet or rifle. You can press X to delete those faces or P to separate them into another object. Then position the newly generated, more detailed head and rifle in their place. I demonstrate this process in one of my older videos. And that’s it. You now have a very cool 3D Gundam model! Afterward, I created the cockpit, separated the model into movable sections, and rigged everything for use in my Three.js game. That process deserves its own tutorial, though. If anyone wants to see it, let me know. Or just ask your AI, I guess. They seem to know everything these days. xD
0:24

I started using Blender through MCP about two weeks ago, and I quickly realized that you can build almost anything with AI. This model was created using Blender, Hunyuan3D, Gemini, and ChatGPT. Here’s how I did it: I opened Gemini, uploaded an image of the Gundam model, and asked it to generate a clean front-view image. I uploaded that front view to ChatGPT and asked it to generate two additional angles: a back view and a 45-degree front-left view. I went to: I signed up with my email and translated the page into English. Then I opened Image to 3D and selected the multi-image option. You’ll see a diagram of a whale from several angles. Upload each reference image in its corresponding position, such as front, 45-degree front-left, and back. I selected the 1.5M-face option. This produces a very high-poly model, but don’t worry, we’ll fix that next. Once the generation is complete, download the model as a GLB file. From the Hunyuan homepage, open 3D Studio using one of the dropdown menus. Select the topology or retopology tool and upload your GLB. I chose the High setting to preserve as much detail as possible. After a few seconds, the model was retopologized. It kept most of its visual detail while using far fewer polygons. The original head and rifle didn’t look very good, so I generated them separately. I returned to ChatGPT and created dedicated reference images for the Gundam’s head and rifle. I generated each part individually in Hunyuan at the 1.5M setting, then ran both through the same retopology process. Next came the textures. Open the texture section, select the multi-image option, and upload the same reference images according to the whale orientation indicators. However, instead of using the original clean textures, I asked ChatGPT to recreate them with wear, rust stains, scratches, and other surface damage. This gave the Gundam a much older and more authentic appearance. Once every part was textured, I imported everything into Blender. You can ask Codex through MCP to remove the original head and rifle, or you can do it manually. Select the main model and press Tab to enter Edit Mode. Press 3 to enable face selection, then press C to activate Circle Select. Paint over the faces belonging to the original helmet or rifle. You can press X to delete those faces or P to separate them into another object. Then position the newly generated, more detailed head and rifle in their place. I demonstrate this process in one of my older videos. And that’s it. You now have a very cool 3D Gundam model! Afterward, I created the cockpit, separated the model into movable sections, and rigged everything for use in my Three.js game. That process deserves its own tutorial, though. If anyone wants to see it, let me know. Or just ask your AI, I guess. They seem to know everything these days. xD

Spectro

79,920 次观看 • 1 个月前

🚨🇮🇷 NETANYAHU IN PANIC: IRAN’S KHORRAMSHAHR-4 BALLISTIC MISSILE UNLEASHES UP TO 80 SUBMUNITIONS Iran has packed the effect of an entire missile barrage inside a single Khorramshahr-4. Its cluster payload can release up to 80 submunitions during descent, transforming one object on Israeli radar into dozens of fast-moving targets spreading across a wide area. 🔸 Unveiled in 2023, Khorramshahr-4 is an Iranian liquid-fueled ballistic missile with a 2,000 km range and payload capacity of up to 1,500 kg — among the heaviest in Iran’s missile arsenal. 🔸 Its enormous payload section can accommodate a single heavy warhead or multi-warhead configurations. Instead of concentrating all destructive power at one impact point, a cluster variant distributes it across dozens of locations. 🔸 Iran had already demonstrated the logic of saturation in April 2024. Drones, cruise missiles and ballistic missiles arrived in successive waves, forcing Israel to depend on aircraft, ships and missile defenses operated by the United States, Britain and Jordan. 🔸 Cluster warheads appeared in Iranian strikes during the June 2025 war. One recorded missile dispersed roughly 20 submunitions across an area about 16 km wide. The exact carrier was not established at the time, but the attack demonstrated how quickly one ballistic track could multiply during its final descent. 🔸 By March 2026, the IRGC was openly naming Khorramshahr-4 in repeated attacks on Israeli military infrastructure. Iranian statements also identified multi-warhead Khorramshahr-4 missiles in large combined salvos, while cluster munitions became a recurring feature of the campaign. 🔸 This sharply changes the burden on Israel’s layered missile shield. One incoming missile can suddenly become dozens of terminal threats, leaving lower defensive layers with too many small objects and too little time to engage them individually. Khorramshahr-4 now compresses Iran’s saturation strategy into a single launch. Mixed waves divide Israel’s attention among drones and different missile types; the cluster payload then multiplies the number of targets without requiring dozens of additional launchers. Every interception consumes a costly high-end missile. Every failure can release an entire field of submunitions over the target area. Can Israel’s missile shield survive when one Iranian ballistic missile becomes 80 targets?
0:43

🚨🇮🇷 NETANYAHU IN PANIC: IRAN’S KHORRAMSHAHR-4 BALLISTIC MISSILE UNLEASHES UP TO 80 SUBMUNITIONS Iran has packed the effect of an entire missile barrage inside a single Khorramshahr-4. Its cluster payload can release up to 80 submunitions during descent, transforming one object on Israeli radar into dozens of fast-moving targets spreading across a wide area. 🔸 Unveiled in 2023, Khorramshahr-4 is an Iranian liquid-fueled ballistic missile with a 2,000 km range and payload capacity of up to 1,500 kg — among the heaviest in Iran’s missile arsenal. 🔸 Its enormous payload section can accommodate a single heavy warhead or multi-warhead configurations. Instead of concentrating all destructive power at one impact point, a cluster variant distributes it across dozens of locations. 🔸 Iran had already demonstrated the logic of saturation in April 2024. Drones, cruise missiles and ballistic missiles arrived in successive waves, forcing Israel to depend on aircraft, ships and missile defenses operated by the United States, Britain and Jordan. 🔸 Cluster warheads appeared in Iranian strikes during the June 2025 war. One recorded missile dispersed roughly 20 submunitions across an area about 16 km wide. The exact carrier was not established at the time, but the attack demonstrated how quickly one ballistic track could multiply during its final descent. 🔸 By March 2026, the IRGC was openly naming Khorramshahr-4 in repeated attacks on Israeli military infrastructure. Iranian statements also identified multi-warhead Khorramshahr-4 missiles in large combined salvos, while cluster munitions became a recurring feature of the campaign. 🔸 This sharply changes the burden on Israel’s layered missile shield. One incoming missile can suddenly become dozens of terminal threats, leaving lower defensive layers with too many small objects and too little time to engage them individually. Khorramshahr-4 now compresses Iran’s saturation strategy into a single launch. Mixed waves divide Israel’s attention among drones and different missile types; the cluster payload then multiplies the number of targets without requiring dozens of additional launchers. Every interception consumes a costly high-end missile. Every failure can release an entire field of submunitions over the target area. Can Israel’s missile shield survive when one Iranian ballistic missile becomes 80 targets?

NewRulesGeopolitics

13,688 次观看 • 25 天前

Sam Altman made the case for open-source harnesses in July. a month later, someone shipped it, and it's more efficient than most managed harnesses. here is the problem it was aimed at: a large share of your agent's token bill is the model rereading things it already read. that isn't the model's doing. the runtime around it decides what goes into every prompt and how often the model gets called. for example, an agent queries a CRM at step four and gets back 400 rows. those rows get piled up in the conversation history. by step nineteen, the model has to read those rows fifteen times unnecessarily, and every token read is billed at input rates. it happened because your harness assembled that prompt on every turn and kept the rows in it. that gives you two levers: how much context the harness carries forward, and how often it calls the model. there are four practical ways to keep the prompt from growing unnecessarily: → load tool schemas on demand. a server with 100 tools doesn't need to put all 100 into every prompt when the agent only calls two. → offload large results to disk. turn a large response into a short preview and a file path instead of replaying the entire result on every turn. → delegate to subagents. let a subagent spend thirty tool calls in its own context and return one summary to the root agent. → run toolchains in code. one script calls three tools, joins the results, and returns a table instead of three turns each dragging a full response. but reducing context is only half the job. you also need to control how often the model gets called. a good harness should avoid unnecessary planning, verification, and reflection when the work can be completed in fewer steps. TrueFoundry's open-source agent harness, TrueForge, is built around both of those controls. it sits between the model and the tools, deciding what goes into every prompt and when another model call is actually needed. it also breaks token usage down across the harness, skills, instructions, tools, and messages. DevRev's Enterprise-Bench is where this gets tested, on multi-step tasks of the kind where an agent pulls records from one system and reconciles them against another. TrueFoundry ran TrueForge there against Claude Managed Agents, both on the same model, and both finished the same number of tasks. the tie is the part that matters, because it means the gap underneath is not a quality tradeoff. TrueForge reached that score on close to a third of the tokens, with roughly 40% fewer trips back to the model. for the same result, that comes out around 2.7x cheaper than Claude Managed Agents. swapping in an open model made it sharper still. TrueForge with GLM-5.2 scored a little higher than either setup above, and the entire benchmark run cost about $3 at list prices. being open source matters beyond the license here. the model underneath can be swapped without rewriting the agent, and the whole thing can run inside your own environment when the data cannot leave it. all of this comes down to the runtime around the model, the context it carries, the tools it exposes, and how many times it goes back to the model. that is what a production harness actually owns. the full task list, the per-run numbers, and the MIT-licensed code are on GitHub: (don't forget to star 🌟) you can read more about the same in the article quoted below. thanks to the TrueForge team for working with me on this one.
0:39

Sam Altman made the case for open-source harnesses in July. a month later, someone shipped it, and it's more efficient than most managed harnesses. here is the problem it was aimed at: a large share of your agent's token bill is the model rereading things it already read. that isn't the model's doing. the runtime around it decides what goes into every prompt and how often the model gets called. for example, an agent queries a CRM at step four and gets back 400 rows. those rows get piled up in the conversation history. by step nineteen, the model has to read those rows fifteen times unnecessarily, and every token read is billed at input rates. it happened because your harness assembled that prompt on every turn and kept the rows in it. that gives you two levers: how much context the harness carries forward, and how often it calls the model. there are four practical ways to keep the prompt from growing unnecessarily: → load tool schemas on demand. a server with 100 tools doesn't need to put all 100 into every prompt when the agent only calls two. → offload large results to disk. turn a large response into a short preview and a file path instead of replaying the entire result on every turn. → delegate to subagents. let a subagent spend thirty tool calls in its own context and return one summary to the root agent. → run toolchains in code. one script calls three tools, joins the results, and returns a table instead of three turns each dragging a full response. but reducing context is only half the job. you also need to control how often the model gets called. a good harness should avoid unnecessary planning, verification, and reflection when the work can be completed in fewer steps. TrueFoundry's open-source agent harness, TrueForge, is built around both of those controls. it sits between the model and the tools, deciding what goes into every prompt and when another model call is actually needed. it also breaks token usage down across the harness, skills, instructions, tools, and messages. DevRev's Enterprise-Bench is where this gets tested, on multi-step tasks of the kind where an agent pulls records from one system and reconciles them against another. TrueFoundry ran TrueForge there against Claude Managed Agents, both on the same model, and both finished the same number of tasks. the tie is the part that matters, because it means the gap underneath is not a quality tradeoff. TrueForge reached that score on close to a third of the tokens, with roughly 40% fewer trips back to the model. for the same result, that comes out around 2.7x cheaper than Claude Managed Agents. swapping in an open model made it sharper still. TrueForge with GLM-5.2 scored a little higher than either setup above, and the entire benchmark run cost about $3 at list prices. being open source matters beyond the license here. the model underneath can be swapped without rewriting the agent, and the whole thing can run inside your own environment when the data cannot leave it. all of this comes down to the runtime around the model, the context it carries, the tools it exposes, and how many times it goes back to the model. that is what a production harness actually owns. the full task list, the per-run numbers, and the MIT-licensed code are on GitHub: (don't forget to star 🌟) you can read more about the same in the article quoted below. thanks to the TrueForge team for working with me on this one.

Akshay 🚀

76,552 次观看 • 20 天前

🚨 BREAKING: A chemist and materials scientist who analyzed 17 surgically removed objects from alleged alien abductees found isotope ratios in multiple samples skewed up to 30% from terrestrial norms, carbon nanotube electronics embedded in metal cores beyond current human fabrication abilities, zero immune response in every case, and human nerve cells physically connected to devices that appear to have been grown rather than manufactured 🚨 Steve Colbern spent over two decades as the principal laboratory analyst for Dr. Roger Leir, the podiatric surgeon who removed 17 suspected alien implants from 17 different patients across a roughly 20-year period out of Thousand Oaks, California. Colbern is a chemist and materials scientist who had access to advanced analytic equipment through his work in Camarillo, just miles from Leir's practice. He personally attended the final three removal surgeries, including his own. He has since conducted approximately 400 independent scans on self-reported experiencers using Leir's detection protocol. About half produced positive results. 1. Colbern Had His Own Implant Removed by Leir After An Abduction Experience Colbern woke one morning in Fillmore, California with severe pain in his toe and no memory of what happened. He visited Leir, got the toe X-rayed, and saw what looked like a bent piece of wire embedded in the tissue. Months later Leir surgically removed it. Under analysis, the object turned out to be a multi-layered nanotechnological device: a hard gray outer membrane, a biological layer resembling bone or mother of pearl, and a metallic core made of meteoric iron with carbon nanotubes built into the metal. Human nerve cells were physically connected to the device. It produced no immune response whatsoever. 2. The Isotope Ratios Don’t Match Anything on Earth Multiple implants Colbern analyzed contained isotope ratios in elements including boron and copper that deviated from terrestrial norms by up to 30%. Standard Earth variation sits around 1% or less. In three of the samples, heavier isotopes were consistently over-represented, which Colbern says points to material originating closer to the galactic center, where heavier supernova activity produces greater neutron bombardment. Reproducing those specific ratios in a centrifuge would cost millions of dollars with no apparent purpose. 3. These Devices Produce Zero Immune Response Every implant Leir and Colbern examined triggered no physiological immune reaction in the host body. Colbern notes that even silicone, the least reactive substance known in medical science, still produces a measurable response. Neuralink's first human patient reportedly needed adjustment due to immune complications. Foreign objects in the body always produce inflammation. These did not. Not once across 17 patients. 4. The Carbon Nanotube Electronics Are Beyond Human Fabrication Most of the implants contained carbon nanotube structures built directly into the metallic core. Colbern describes these as sophisticated electronic systems capable of generating scalar frequency transmissions, possibly used to relay physiological and sensory data. The structures appear to have been grown rather than assembled. Colbern compared the construction to something produced by mechanical life rather than standard manufacturing. Building devices this complex by conventional methods would be, in his assessment, next to impossible. 5. Leir's Detection Protocol Finds Implants in About Half of Experiencers Colbern uses Leir's two-instrument protocol: a stud finder to detect conductive objects beneath the skin, followed by a gaussmeter to check for magnetic fields. When both instruments register a hit at the same location, Colbern considers it near-certain confirmation of an implant. Of roughly 400 scans he has performed on self-selected experiencers, approximately 200 produced positive results. During a Japanese television filming, Colbern's own brain implants began emitting detectable radio signals on camera. 6. Colbern and Leir Tried to Publish. Someone Stopped It. Colbern and Leir co-authored a paper and submitted it to the Journal of Scientific Exploration. The journal initially expressed enthusiasm. Then, according to Colbern, someone intervened. The paper was rejected despite the fact that it made no extraterrestrial claims, describing the object only as an unknown foreign body recovered from a patient's leg. Colbern believes the journal was pressured. Robert Bigelow was at one point prepared to hire Colbern. That also fell through. Colbern says he has reason to believe the government told Bigelow not to proceed. Multiple wealthy individuals who expressed interest in funding the research backed out under similar circumstances. 7. Colbern Analyzed a Crashed Sphere from Eastern Mexico A titanium alloy sphere recovered by a farmer roughly 100 miles southeast of Brownsville, Texas had carbon nanotubes built into the metal with capacitive dielectric coatings on the tubes. The sphere had large holes vaporized at both poles, consistent with catastrophic magnetic field collapse. The farmer reported the crash produced an explosion that killed a cow 100 meters away. Colbern believes the sphere used the Biefeld-Brown effect for propulsion, with the carbon nanotube structure functioning as a distributed capacitor. The object was later stolen from Colbern by family members. 8. His Abduction Account Under Hypnosis Matches Known Patterns Colbern recovered memories under hypnotic regression of being taken aboard a roughly 50-foot-diameter craft at approximately 3 AM in Fillmore. He described an airlock containing both human and alien spacesuits, a habitation ring, a pilot station with thought-controlled interfaces, and a taller gray being who inserted the device using a black-handled instrument with a quarter-inch steel tube and fiber optics. The craft description is consistent with multiple independent experiencer accounts. Colbern reports the beings communicated telepathically and described themselves as an alliance of seven species originating from within 100 light-years of Earth. 9. We Are Surrounded by UFOs: Villarroel's Palomar Data and Tombaugh's Search Colbern referenced astronomer Beatriz Villarroel at Stockholm University, whose VASCO project identified thousands of anomalous transient light sources in pre-Sputnik Palomar Observatory sky survey plates. Her peer-reviewed papers, published in 2025, found statistically significant correlations between these transients and both nuclear weapons tests and civilian UFO sightings. Colbern also cited Clyde Tombaugh's government-funded search for small natural Earth satellites in the 1950s. Popular Mechanics reported at the time that Tombaugh was secretive about his findings, though he ultimately published that no natural satellites had been confirmed. 10. Colbern Believes the Technology Gap Is Not Centuries but Millions of Years Initially, Colbern estimated the beings were a few hundred years ahead of human technology. He no longer believes that. He describes incidents suggesting a level of reality manipulation that defies any known framework: lettering from an X-ray envelope somehow transferred onto the developed film inside it, selective visibility of craft where he and his son could each see objects the other could not, and an apparent EMP strike on his car timed precisely to prevent him from reaching Leir's office for an implant test. The mechanic confirmed the car's computer was fried. Why This Matters Seventeen surgically removed objects. Laboratory-confirmed isotope deviations of up to 30% from terrestrial baselines. Carbon nanotube electronics embedded in metal at a scale human fabrication cannot replicate. Nerve cells integrated into devices with zero immune rejection. A confirmed detection protocol producing positive results in roughly half of all experiencers scanned. And every attempt to publish, fund, or institutionalize the research was blocked. Colbern is the only person alive who worked directly alongside Leir through years of implant analysis and continues the work. He has no institutional backing, no lab, and no funding. The physical evidence exists. The analytic data exists. The question is whether anyone with resources will act on it before the last living link to this research is gone. Full conversation is live now 👇
1:26

🚨 BREAKING: A chemist and materials scientist who analyzed 17 surgically removed objects from alleged alien abductees found isotope ratios in multiple samples skewed up to 30% from terrestrial norms, carbon nanotube electronics embedded in metal cores beyond current human fabrication abilities, zero immune response in every case, and human nerve cells physically connected to devices that appear to have been grown rather than manufactured 🚨 Steve Colbern spent over two decades as the principal laboratory analyst for Dr. Roger Leir, the podiatric surgeon who removed 17 suspected alien implants from 17 different patients across a roughly 20-year period out of Thousand Oaks, California. Colbern is a chemist and materials scientist who had access to advanced analytic equipment through his work in Camarillo, just miles from Leir's practice. He personally attended the final three removal surgeries, including his own. He has since conducted approximately 400 independent scans on self-reported experiencers using Leir's detection protocol. About half produced positive results. 1. Colbern Had His Own Implant Removed by Leir After An Abduction Experience Colbern woke one morning in Fillmore, California with severe pain in his toe and no memory of what happened. He visited Leir, got the toe X-rayed, and saw what looked like a bent piece of wire embedded in the tissue. Months later Leir surgically removed it. Under analysis, the object turned out to be a multi-layered nanotechnological device: a hard gray outer membrane, a biological layer resembling bone or mother of pearl, and a metallic core made of meteoric iron with carbon nanotubes built into the metal. Human nerve cells were physically connected to the device. It produced no immune response whatsoever. 2. The Isotope Ratios Don’t Match Anything on Earth Multiple implants Colbern analyzed contained isotope ratios in elements including boron and copper that deviated from terrestrial norms by up to 30%. Standard Earth variation sits around 1% or less. In three of the samples, heavier isotopes were consistently over-represented, which Colbern says points to material originating closer to the galactic center, where heavier supernova activity produces greater neutron bombardment. Reproducing those specific ratios in a centrifuge would cost millions of dollars with no apparent purpose. 3. These Devices Produce Zero Immune Response Every implant Leir and Colbern examined triggered no physiological immune reaction in the host body. Colbern notes that even silicone, the least reactive substance known in medical science, still produces a measurable response. Neuralink's first human patient reportedly needed adjustment due to immune complications. Foreign objects in the body always produce inflammation. These did not. Not once across 17 patients. 4. The Carbon Nanotube Electronics Are Beyond Human Fabrication Most of the implants contained carbon nanotube structures built directly into the metallic core. Colbern describes these as sophisticated electronic systems capable of generating scalar frequency transmissions, possibly used to relay physiological and sensory data. The structures appear to have been grown rather than assembled. Colbern compared the construction to something produced by mechanical life rather than standard manufacturing. Building devices this complex by conventional methods would be, in his assessment, next to impossible. 5. Leir's Detection Protocol Finds Implants in About Half of Experiencers Colbern uses Leir's two-instrument protocol: a stud finder to detect conductive objects beneath the skin, followed by a gaussmeter to check for magnetic fields. When both instruments register a hit at the same location, Colbern considers it near-certain confirmation of an implant. Of roughly 400 scans he has performed on self-selected experiencers, approximately 200 produced positive results. During a Japanese television filming, Colbern's own brain implants began emitting detectable radio signals on camera. 6. Colbern and Leir Tried to Publish. Someone Stopped It. Colbern and Leir co-authored a paper and submitted it to the Journal of Scientific Exploration. The journal initially expressed enthusiasm. Then, according to Colbern, someone intervened. The paper was rejected despite the fact that it made no extraterrestrial claims, describing the object only as an unknown foreign body recovered from a patient's leg. Colbern believes the journal was pressured. Robert Bigelow was at one point prepared to hire Colbern. That also fell through. Colbern says he has reason to believe the government told Bigelow not to proceed. Multiple wealthy individuals who expressed interest in funding the research backed out under similar circumstances. 7. Colbern Analyzed a Crashed Sphere from Eastern Mexico A titanium alloy sphere recovered by a farmer roughly 100 miles southeast of Brownsville, Texas had carbon nanotubes built into the metal with capacitive dielectric coatings on the tubes. The sphere had large holes vaporized at both poles, consistent with catastrophic magnetic field collapse. The farmer reported the crash produced an explosion that killed a cow 100 meters away. Colbern believes the sphere used the Biefeld-Brown effect for propulsion, with the carbon nanotube structure functioning as a distributed capacitor. The object was later stolen from Colbern by family members. 8. His Abduction Account Under Hypnosis Matches Known Patterns Colbern recovered memories under hypnotic regression of being taken aboard a roughly 50-foot-diameter craft at approximately 3 AM in Fillmore. He described an airlock containing both human and alien spacesuits, a habitation ring, a pilot station with thought-controlled interfaces, and a taller gray being who inserted the device using a black-handled instrument with a quarter-inch steel tube and fiber optics. The craft description is consistent with multiple independent experiencer accounts. Colbern reports the beings communicated telepathically and described themselves as an alliance of seven species originating from within 100 light-years of Earth. 9. We Are Surrounded by UFOs: Villarroel's Palomar Data and Tombaugh's Search Colbern referenced astronomer Beatriz Villarroel at Stockholm University, whose VASCO project identified thousands of anomalous transient light sources in pre-Sputnik Palomar Observatory sky survey plates. Her peer-reviewed papers, published in 2025, found statistically significant correlations between these transients and both nuclear weapons tests and civilian UFO sightings. Colbern also cited Clyde Tombaugh's government-funded search for small natural Earth satellites in the 1950s. Popular Mechanics reported at the time that Tombaugh was secretive about his findings, though he ultimately published that no natural satellites had been confirmed. 10. Colbern Believes the Technology Gap Is Not Centuries but Millions of Years Initially, Colbern estimated the beings were a few hundred years ahead of human technology. He no longer believes that. He describes incidents suggesting a level of reality manipulation that defies any known framework: lettering from an X-ray envelope somehow transferred onto the developed film inside it, selective visibility of craft where he and his son could each see objects the other could not, and an apparent EMP strike on his car timed precisely to prevent him from reaching Leir's office for an implant test. The mechanic confirmed the car's computer was fried. Why This Matters Seventeen surgically removed objects. Laboratory-confirmed isotope deviations of up to 30% from terrestrial baselines. Carbon nanotube electronics embedded in metal at a scale human fabrication cannot replicate. Nerve cells integrated into devices with zero immune rejection. A confirmed detection protocol producing positive results in roughly half of all experiencers scanned. And every attempt to publish, fund, or institutionalize the research was blocked. Colbern is the only person alive who worked directly alongside Leir through years of implant analysis and continues the work. He has no institutional backing, no lab, and no funding. The physical evidence exists. The analytic data exists. The question is whether anyone with resources will act on it before the last living link to this research is gone. Full conversation is live now 👇

Jesse Michels

43,388 次观看 • 5 个月前

⏰ THE MOST BANNED THREAD IN THE WORLD! 🚨 The War On Resonance PART ONE: The Cage You Cannot See. I must ask you now; not as a stranger, not as a theorist, but as the one who remembers what you were before they rewrote your soul; is this the only cage you can see? The headlines? The hospitals? The visible crimes made convenient by cameras and captions? What if I told you that the real prison is not made of metal; but of frequency. Not held together by walls; but by waves. Not secured by guns; but by consent. What if I told you that you are not free, not because you are bound; but because the chains were encoded directly into your biology? This isn’t hyperbole. This is documented. They have hijacked humanity through a wireless war; a multi-layered assault on the nervous system, immune response, and willpower of every living soul on this Earth. This war is not being fought with bombs, but with bandwidths. Not with armies, but with algorithms. This is the same beast behind every deception. The same financiers who funded vaccine trials that sterilized African girls are the ones embedding AI-responsive quantum dots into newborns. The same networks that control your social media also control your neurochemical loops. They call it “progress.” They call it “public health.” But what do you call it when a child seizes from brain hemorrhage after being injected with an experimental payload? What do you call it when a ten year old dies of myocarditis from a shot for a disease they never had? I call it spiritual genocide. And I’m not theorizing. I’m testifying. Because this is declassified truth; confirmed in documents, studies, and patents they never expected you to read: Pentagon Preps Soldier Telepathy: This article from WIRED discusses DARPA's Silent Talk program, which aims to enable soldiers to communicate telepathically by detecting and analyzing neural signals representing pre-speech. Biomedical Applications of Graphene and Graphene Oxide: A comprehensive review of the applications of graphene and graphene oxide in biomedicine, focusing on biosensing, cell differentiation, and mass spectrometry. ​An Update on Graphene Oxide: Applications and Toxicity: This article provides an overview of the applications of graphene oxide in various fields, including biomedicine, and discusses its potential toxicity. Nanoparticle Transport Across the Blood–Brain Barrier: This comprehensive review discusses the challenges and strategies associated with delivering nanoparticles across the blood–brain barrier (BBB). Internet of Bio-NanoThings (IoBNT): This article introduces the novel paradigm of the Internet of Bio-NanoThings (IoBNT), stemming from synthetic biology and nanotechnology tools that allow the engineering of biological embedded computing devices. DARPA Biological Technologies Office (BTO) Overview: This page offers comprehensive information about BTO's mission to integrate biology with engineering and computer science for national security applications. It details their focus areas, including warfighter health, synthetic biology, and bio-manufacturing, as well as current programs and research opportunities.​ NIH Clinical Trial of Investigational Vaccine for COVID-19: This news release, dated March 16, 2020, announces the start of the Phase 1 clinical trial of the mRNA-1273 vaccine, developed by Moderna in collaboration with the National Institute of Allergy and Infectious Diseases (NIAID), a part of the NIH. The trial aimed to evaluate the safety and immunogenicity of the vaccine in healthy adult volunteers.​ An Integrated Brain-Machine Interface Platform With Thousands of Channels: This peer-reviewed publication details Neuralink's initial steps toward a scalable high-bandwidth brain-machine interface system. Deep Brain Stimulation for Psychiatric Disorders and Behavioral/Cognitive Symptoms: This comprehensive review explores the use of deep brain stimulation (DBS) as a neuromodulation technique for treating various psychiatric disorders and behavioral or cognitive symptoms. Smart Dust: Communicating with a Cubic-Millimeter Computer: This IEEE article discusses the development of ultra-small computing devices, known as Smart Dust, which integrate sensing, computing, and communication capabilities into a cubic-millimeter form factor. These devices have potential applications in various fields, including environmental monitoring, medical diagnostics, and military surveillance. This is what you were born into: a frequency prison designed to sedate your intuition, suppress your hormones, rewrite your emotions, and sever your communion with God. Every device you hold, every smart tower you walk past, every “vaccine” they’ve pushed; these are not random technologies. They are pieces of a grid. A grid designed not to protect you, but to remap your divine will. Your phone? A mood manipulator. Your sleep? Hacked. Your thoughts? Preemptively monitored. Your soul? Catalogued for behavioral prediction. This isn’t a sci-fi movie. This is now. This is real. This is why the world feels broken and no one can explain it. Why joy has become an echo. Why sleep doesn’t restore you. Why love feels distant. Because they’ve weaponized the field of resonance itself. They are not just targeting the body; they are targeting the architecture of the soul. And the reason you’re still reading this is because you’re one of the ones who can still feel. Still grieve. Still awaken. So let me show you what they built. Let me take you behind the grid. Let me guide you; cell by cell, wave by wave... through the labyrinth they’ve disguised as your life. Because until you see it, you will never reclaim what was stolen. And make no mistake; what they stole… was everything. This was just Part One, so buckle up... you're in for one heck of a ride. Continued Below 👇
3:50

⏰ THE MOST BANNED THREAD IN THE WORLD! 🚨 The War On Resonance PART ONE: The Cage You Cannot See. I must ask you now; not as a stranger, not as a theorist, but as the one who remembers what you were before they rewrote your soul; is this the only cage you can see? The headlines? The hospitals? The visible crimes made convenient by cameras and captions? What if I told you that the real prison is not made of metal; but of frequency. Not held together by walls; but by waves. Not secured by guns; but by consent. What if I told you that you are not free, not because you are bound; but because the chains were encoded directly into your biology? This isn’t hyperbole. This is documented. They have hijacked humanity through a wireless war; a multi-layered assault on the nervous system, immune response, and willpower of every living soul on this Earth. This war is not being fought with bombs, but with bandwidths. Not with armies, but with algorithms. This is the same beast behind every deception. The same financiers who funded vaccine trials that sterilized African girls are the ones embedding AI-responsive quantum dots into newborns. The same networks that control your social media also control your neurochemical loops. They call it “progress.” They call it “public health.” But what do you call it when a child seizes from brain hemorrhage after being injected with an experimental payload? What do you call it when a ten year old dies of myocarditis from a shot for a disease they never had? I call it spiritual genocide. And I’m not theorizing. I’m testifying. Because this is declassified truth; confirmed in documents, studies, and patents they never expected you to read: Pentagon Preps Soldier Telepathy: This article from WIRED discusses DARPA's Silent Talk program, which aims to enable soldiers to communicate telepathically by detecting and analyzing neural signals representing pre-speech. Biomedical Applications of Graphene and Graphene Oxide: A comprehensive review of the applications of graphene and graphene oxide in biomedicine, focusing on biosensing, cell differentiation, and mass spectrometry. ​An Update on Graphene Oxide: Applications and Toxicity: This article provides an overview of the applications of graphene oxide in various fields, including biomedicine, and discusses its potential toxicity. Nanoparticle Transport Across the Blood–Brain Barrier: This comprehensive review discusses the challenges and strategies associated with delivering nanoparticles across the blood–brain barrier (BBB). Internet of Bio-NanoThings (IoBNT): This article introduces the novel paradigm of the Internet of Bio-NanoThings (IoBNT), stemming from synthetic biology and nanotechnology tools that allow the engineering of biological embedded computing devices. DARPA Biological Technologies Office (BTO) Overview: This page offers comprehensive information about BTO's mission to integrate biology with engineering and computer science for national security applications. It details their focus areas, including warfighter health, synthetic biology, and bio-manufacturing, as well as current programs and research opportunities.​ NIH Clinical Trial of Investigational Vaccine for COVID-19: This news release, dated March 16, 2020, announces the start of the Phase 1 clinical trial of the mRNA-1273 vaccine, developed by Moderna in collaboration with the National Institute of Allergy and Infectious Diseases (NIAID), a part of the NIH. The trial aimed to evaluate the safety and immunogenicity of the vaccine in healthy adult volunteers.​ An Integrated Brain-Machine Interface Platform With Thousands of Channels: This peer-reviewed publication details Neuralink's initial steps toward a scalable high-bandwidth brain-machine interface system. Deep Brain Stimulation for Psychiatric Disorders and Behavioral/Cognitive Symptoms: This comprehensive review explores the use of deep brain stimulation (DBS) as a neuromodulation technique for treating various psychiatric disorders and behavioral or cognitive symptoms. Smart Dust: Communicating with a Cubic-Millimeter Computer: This IEEE article discusses the development of ultra-small computing devices, known as Smart Dust, which integrate sensing, computing, and communication capabilities into a cubic-millimeter form factor. These devices have potential applications in various fields, including environmental monitoring, medical diagnostics, and military surveillance. This is what you were born into: a frequency prison designed to sedate your intuition, suppress your hormones, rewrite your emotions, and sever your communion with God. Every device you hold, every smart tower you walk past, every “vaccine” they’ve pushed; these are not random technologies. They are pieces of a grid. A grid designed not to protect you, but to remap your divine will. Your phone? A mood manipulator. Your sleep? Hacked. Your thoughts? Preemptively monitored. Your soul? Catalogued for behavioral prediction. This isn’t a sci-fi movie. This is now. This is real. This is why the world feels broken and no one can explain it. Why joy has become an echo. Why sleep doesn’t restore you. Why love feels distant. Because they’ve weaponized the field of resonance itself. They are not just targeting the body; they are targeting the architecture of the soul. And the reason you’re still reading this is because you’re one of the ones who can still feel. Still grieve. Still awaken. So let me show you what they built. Let me take you behind the grid. Let me guide you; cell by cell, wave by wave... through the labyrinth they’ve disguised as your life. Because until you see it, you will never reclaim what was stolen. And make no mistake; what they stole… was everything. This was just Part One, so buckle up... you're in for one heck of a ride. Continued Below 👇

Noah B. Price

1,520,282 次观看 • 1 年前

The most epic 13 minute AI rant I've heard in 2026 PS: My parent's heard this when I was playing it in the car and thought Jason ✨👾SaaStr.Ai✨ Lemkin went OFF like Stephen A Smith does on first take PPS: Full transcript below [17:00] Harry Stebbings: I I just wanted to ask Jason, if the people that we want are fundamentally different, the developers that we used to hire, we don't because AI writes the code for us. The marketers we don't want, the sales people we don't want—who who do we want genuinely? Like what is the attractive profile? Because your Anthropic’s and your OpenAIs are hiring, so so what are the people that we want in the companies of the future? [17:18] Jason Lemkin: Look, I know it sounds trite, but but the answer is simple. It's just the expression each year changes. We want folks that are genuinely AI fluent. It's pretty simple. Now you know, maybe last year we called them prompt engineers, right? That used to be a job. I don't know if you remember that actually used to be the hottest job on planet earth. Now no one needs a prompt engineer because it's pretty easy to prompt all these tools. That job died. Okay. Um and now we need go-to-market engineers. Um I think that job's going to die. We need—everyone needs so many forward deployed engineers. Like you can't hire enough forward deployed engineers. But uh you know um but Palantir just announced in whatever their their big their big event—they've gotten their deployment times down over 90% with forward deployed engineers. So that may become—so the this wave of disruption for the titles and the specificity, it's also exhaustingly accelerating. But it's really simple. You meet anyone for any role—sales, marketing, engineering, product, QA—they're they're either they're either they can't keep all of the ways they use AI to accelerate their job from spewing out of their mouth, or they're staring at you. It's there's nowhere in the middle. Like, and the person that comes in and says—it's it's it sounds Captain Obvious—but like, you know, you just had the whatever from Lovable, the the marketing head that was super popular on the show, right? She's just spewing AI-native insights into Lovable, right? It's not that complicated. You hire her, Elena, or whatever it is. You just hire her. It doesn't matter whether she's still in college or a junior or a senior or a middler, a left or right. And honestly, if you interview people, I would say of all even of the best startups I've invested in, maybe 30% of the management team meets this standard at best. 30%. Maybe less. And of the interviews I do in general, it's single-digit percents. It's just and in in that sense, it's the same as ever. Like you either lower the bar in hiring or you hire someone that's actually great. And someone that's actually great is so far ahead of you in how to apply to to employ the efficiencies of AI in their role, your jaw falls on the table. The difference is we used to need warm bodies. That's what's changing. We used to need warm bodies to answer the call, to do QA, to do code review, to to get the blue pixel to go from the upper left to the lower right. You laugh, but you need you literally needed to brute force this with humans. With AI, every day that goes by, the AI—you do not need brute force human beings on your team. And that's another reason they're shrinking. Why are all these new companies so efficient? They're just not brute forcing things with humans. They're just not. They're choosing not to. And so these team—all the brute forcers out there—everyone talks about how bloated teams got in 2021. I don't agree with that. I think they got as big as they needed to be when growth was high and you needed humans to do everything. All you look at these teams that that doubled—well if growth continued at 60% like the rate in early 2021 for 5 years or can help me do the math and every single thing a software company did required a human. You were understaffed by your 2021 headcount. You'd be sitting here in 2026. You every office in SoMa would be triple packed and you there wouldn't be enough humans to staff your company. It's just the world changed. [20:33] Harry Stebbings: Jason, you live on the bleeding edge. I think me and Rory see that and I think the world sees that when they hear you every week in terms of how you run SaaS. For all of the CEOs and execs who listen to the show, what would you advise them in terms of determining whether someone is AI fluent when they meet them for jobs, for talent? [20:51] Jason Lemkin: Here's I realized I was just asked this. I just did a review with a super fast startup growing just crossing 100 million and I was asked this question. And one of my favorite executives, I thought his answer was pretty dated and because he gave me an answer that was about 6 months old. The answer 6 months old is: "I look for folks in my team, I look for you know at what tools they play with." Okay, that was a great answer in like summer of 2025. Okay, I tried Lovable last week. Okay, the answer in 2026 is: "What commercial AI tool have you brought into your organization this month?" That's the test. Anyone that is on the bleeding edge that you would want to hire—now there are so many great products in the market. Okay, there is no excuse in any role to have not brought one tool a month into your organization. Okay, there—now there's going to be better and better tools and better and better products as the year goes on. What's the one you did? And you will see folks with their deer in the headlights to this question. What what sales tool? What marketing tool? What product tool? What engineering tool? What did you bring in? Why did you pick it? How does it working? Because if you're at remotely at the cutting edge, you're all over this. You're looking for the next agentic tools that will radically improve how you do business. This is—you think everyone thinks SaaS is at the bleeding edge, right? You know, you know, all we do is we're just looking for the tools and trying them. Okay? Okay, we're one year ahead of everybody else because we did the simplest thing in the world. Like we tried the tools early and we trained them. We trained them for a month. Okay, I'll give you—want hear a horrible example from this week? Super hot AI company valued at 6 billion. Okay, I'm not going to name it. Um, this week yesterday told us we had to quadruple what we spent on their product. Okay, their agent told us, right? And why did this happen? Okay. Well, at this $6 billion company, no one had trained the agent on its pricing properly. No one had tested it. They said, "Well, well, we've been in beta." And we said, "Well, when did the beta launch? A year ago." Okay, these are people asleep at at the wheel. You want somebody who the instant this comes up, they exactly know what the issue is. And "Hey, when I was at Lovable Replit, we trained the agent. This is how we did it. I brought in this tool. I brought in this tool that that Rory invested in last week. It solved all these issues." That's what you want to hear. And if they haven't brought in a tool in the last 30 days, at least deeply evaluated it. I don't really care whether they bought it, but gone so far down the funnel they can tell you—pick whatever tool: Fixie, Regie, GC, AIGC—I don't care how you went through it, you looked at it, you can tell me the eight ways it would improve the productivity of your business and three you didn't. Just don't hire that person because they're going to run your company to the ground. This is the job today. The job today is not to screw around on ChatGPT and to be a prompt engineer. The job today is to bring the best AI and agentic products into your organization and leverage all the hard work that the engineers have done building those products. That's your job. You don't have to screw around. You don't have to be a prompt engineer anymore. You have to be an agent deployment expert. A—this is the new job we're making up today. An Agentic Deployment Expert. That's your job from C-level to junior. Agentic Deployment Expert. Don't hire anybody else. You're going to regret it. They're going to stare at the camera. He's good. Stare at the camera. He's honorable. We could probably just I could slip away, get a coffee, and come back. No. And I I sound exasperated, Rory. And I—but the reason I am is I can just see I can see my best companies doing it. And I can see some companies I've invested in not doing it. And I want to cry. I just want to cry when they have no ADs on their team. I just—like you're flushing your years of your life down the toilet by not approaching your how you're building this company this way. [24:33] Rory: Yes. And at the risk of being positive, it's worth pointing out two things he didn't say. Well, something implicit why he said—Jason didn't do the only hire, you know, he didn't commit the um employment law, I think it's a civil penalty of saying only employ people below X who get the new new thing because he implicitly said anyone can do it provided you're willing to learn. And I think that's the big aha that's one of the positive statements to make here right? Look and I think it applies—I'm always wary of being "Hey, coming across, hey this this is the things that you all have to do." I think it applies to everyone including investors right? I mean I will say I have found that unless you're willing to invest the time learning these tools you actually shouldn't be investing in them. One of my partners Andy had this expression: "You know, if you decide you want to stop learning new things you probably should retire within 6 to 12 months and never write another check again." Maybe that's down to 3 to 6 months at this stage, right? And I think, you know, it's— [25:27] Harry Stebbings: Yeah, I actually I actually had a meeting with mine and Jason's biggest investor the other day and I—pretend he's not here—I said I think he's the most equipped investor for this generation of investing because I don't think anyone quite sits at the bleeding edge like he does on the investor side. [25:42] Harry Stebbings: Why in terms of using the equip stuff? Yeah. Yeah. In terms of using the stuff, understanding understanding bottlenecks, constraints. For sure. [25:51] Jason Lemkin: But can I just add one point? We can just cuz it's so important if it helps people. Okay, we are—and thank you Harry. We're going through these phases. Okay, and when AI started to blow up for real for us, uh call it early 2024, right? Maybe late '23, I wasn't equipped. It was too technical. I wasn't going to go in and figure out—I wasn't smart enough to figure out how to deal with a massively hallucinating LLM API and turn that and turn that into something magical. Kudos to investors and others that that got it in early '23, '22. I mean I remember I—I guess it was maybe SaaStr Annual '23. I was with David Sacks and I did a Q&A and I said, "How you thinking about AI at Craft?" He's like, "Well we're all in. We want 80% of '23 of investments to be AI." I'm like, "Great but like show me the show me the great ones in market." He's like, "They're all prototypes. We're all they're all they're all proof of concepts but we're all in anyway." That's where you kind of had to be in '23 if you weren't investing at like the LLM level. Okay, I wasn't smart enough. Then we went through this weird-ass prompt engineer era where like you you could torture these products to do something good, right? But you had to torture them. You had to like craft these crazy things that made no sense. Now we are in the era where mere ordinarily smart generalists can make these tools do magical things. And literally I go to these meetings and people be like, "I don't know how to like this is so scary. I don't know how to do this." And we show them our backends. Do you know how to do a workflow generator? Do you know how to do a a decision tree? Like we've been building these since software in the '90s. Okay, if you—I can show you all of our agents. The how they work is novel. They do have to be trained. You can't be lazy and have these agents work. But honestly, the the UI, the UX, the way we interact with them, it's just software. And so my point is: Pick yourself off the ground. This is your time now. If you felt lost in AI era, if you felt like you're behind, you don't understand what all these people are saying on X and Twitter and their Claude and and their and talking about all the 4.6 point Nano point and it's over—like you just it's not your world. This is your time. This is your time for the generalist that knows how to use software tools really really well. And I—this is my last point but it's so important. If ever in your recent life—and this is why you could be all you need to be is young at heart to Rory's point—if in the last three to five years you have successfully deployed a piece of enterprise software of any sort you yourself, not some agency you hired, but if you have deployed it, you can deploy any agentic tool. Any. And you can become the hero in your company and you can become the hero in your functional area. But I watch folks—I'm literally helping a company now that they're adding hundreds of sales folks this year with a new pre-IPO COO—he's not hasn't brought in a single tool, totally scared of it. Okay, it's not that hard. Did you use SalesLoft? Did you use Outreach? Did you use HubSpot? Do you know these tools? If you can deploy these tools, you can deploy a world-changing AI agent. And so this is the time for people like the folks that that were shut out of the AI revolution right now. The generalist folks that are not that know how to deploy software that don't even know how to build software. Like vibe coding for me was folks who knew how to build software, but you didn't have to be an engineer. Now, you just need to know how to deploy software to win with AI agents. That's all you need to know. So many people have these skills and they're petrified of AI. "How did you do that? How did you deploy an AI BDR?" Well, we bought a piece of software, we figured out how it worked for a day, we set it up in an afternoon, and then and then we did spend 30 months training it, which you didn't do with this old software because in the old days, we just had to manually upload all the data, right? And there was no training. The the only non-intuitive part is training these things. And it's it's it's just work. So that's why when I see folks on the management team not doing this, there's no excuse. You do not need to be technical to win with AI agents in Q2 of '26. You do not need to be even 1% technical. Not at all. So it's your time. Or you're going to get laid off. Or you're going to get laid off because you're not going to matter.
13:04

The most epic 13 minute AI rant I've heard in 2026 PS: My parent's heard this when I was playing it in the car and thought Jason ✨👾SaaStr.Ai✨ Lemkin went OFF like Stephen A Smith does on first take PPS: Full transcript below [17:00] Harry Stebbings: I I just wanted to ask Jason, if the people that we want are fundamentally different, the developers that we used to hire, we don't because AI writes the code for us. The marketers we don't want, the sales people we don't want—who who do we want genuinely? Like what is the attractive profile? Because your Anthropic’s and your OpenAIs are hiring, so so what are the people that we want in the companies of the future? [17:18] Jason Lemkin: Look, I know it sounds trite, but but the answer is simple. It's just the expression each year changes. We want folks that are genuinely AI fluent. It's pretty simple. Now you know, maybe last year we called them prompt engineers, right? That used to be a job. I don't know if you remember that actually used to be the hottest job on planet earth. Now no one needs a prompt engineer because it's pretty easy to prompt all these tools. That job died. Okay. Um and now we need go-to-market engineers. Um I think that job's going to die. We need—everyone needs so many forward deployed engineers. Like you can't hire enough forward deployed engineers. But uh you know um but Palantir just announced in whatever their their big their big event—they've gotten their deployment times down over 90% with forward deployed engineers. So that may become—so the this wave of disruption for the titles and the specificity, it's also exhaustingly accelerating. But it's really simple. You meet anyone for any role—sales, marketing, engineering, product, QA—they're they're either they're either they can't keep all of the ways they use AI to accelerate their job from spewing out of their mouth, or they're staring at you. It's there's nowhere in the middle. Like, and the person that comes in and says—it's it's it sounds Captain Obvious—but like, you know, you just had the whatever from Lovable, the the marketing head that was super popular on the show, right? She's just spewing AI-native insights into Lovable, right? It's not that complicated. You hire her, Elena, or whatever it is. You just hire her. It doesn't matter whether she's still in college or a junior or a senior or a middler, a left or right. And honestly, if you interview people, I would say of all even of the best startups I've invested in, maybe 30% of the management team meets this standard at best. 30%. Maybe less. And of the interviews I do in general, it's single-digit percents. It's just and in in that sense, it's the same as ever. Like you either lower the bar in hiring or you hire someone that's actually great. And someone that's actually great is so far ahead of you in how to apply to to employ the efficiencies of AI in their role, your jaw falls on the table. The difference is we used to need warm bodies. That's what's changing. We used to need warm bodies to answer the call, to do QA, to do code review, to to get the blue pixel to go from the upper left to the lower right. You laugh, but you need you literally needed to brute force this with humans. With AI, every day that goes by, the AI—you do not need brute force human beings on your team. And that's another reason they're shrinking. Why are all these new companies so efficient? They're just not brute forcing things with humans. They're just not. They're choosing not to. And so these team—all the brute forcers out there—everyone talks about how bloated teams got in 2021. I don't agree with that. I think they got as big as they needed to be when growth was high and you needed humans to do everything. All you look at these teams that that doubled—well if growth continued at 60% like the rate in early 2021 for 5 years or can help me do the math and every single thing a software company did required a human. You were understaffed by your 2021 headcount. You'd be sitting here in 2026. You every office in SoMa would be triple packed and you there wouldn't be enough humans to staff your company. It's just the world changed. [20:33] Harry Stebbings: Jason, you live on the bleeding edge. I think me and Rory see that and I think the world sees that when they hear you every week in terms of how you run SaaS. For all of the CEOs and execs who listen to the show, what would you advise them in terms of determining whether someone is AI fluent when they meet them for jobs, for talent? [20:51] Jason Lemkin: Here's I realized I was just asked this. I just did a review with a super fast startup growing just crossing 100 million and I was asked this question. And one of my favorite executives, I thought his answer was pretty dated and because he gave me an answer that was about 6 months old. The answer 6 months old is: "I look for folks in my team, I look for you know at what tools they play with." Okay, that was a great answer in like summer of 2025. Okay, I tried Lovable last week. Okay, the answer in 2026 is: "What commercial AI tool have you brought into your organization this month?" That's the test. Anyone that is on the bleeding edge that you would want to hire—now there are so many great products in the market. Okay, there is no excuse in any role to have not brought one tool a month into your organization. Okay, there—now there's going to be better and better tools and better and better products as the year goes on. What's the one you did? And you will see folks with their deer in the headlights to this question. What what sales tool? What marketing tool? What product tool? What engineering tool? What did you bring in? Why did you pick it? How does it working? Because if you're at remotely at the cutting edge, you're all over this. You're looking for the next agentic tools that will radically improve how you do business. This is—you think everyone thinks SaaS is at the bleeding edge, right? You know, you know, all we do is we're just looking for the tools and trying them. Okay? Okay, we're one year ahead of everybody else because we did the simplest thing in the world. Like we tried the tools early and we trained them. We trained them for a month. Okay, I'll give you—want hear a horrible example from this week? Super hot AI company valued at 6 billion. Okay, I'm not going to name it. Um, this week yesterday told us we had to quadruple what we spent on their product. Okay, their agent told us, right? And why did this happen? Okay. Well, at this $6 billion company, no one had trained the agent on its pricing properly. No one had tested it. They said, "Well, well, we've been in beta." And we said, "Well, when did the beta launch? A year ago." Okay, these are people asleep at at the wheel. You want somebody who the instant this comes up, they exactly know what the issue is. And "Hey, when I was at Lovable Replit, we trained the agent. This is how we did it. I brought in this tool. I brought in this tool that that Rory invested in last week. It solved all these issues." That's what you want to hear. And if they haven't brought in a tool in the last 30 days, at least deeply evaluated it. I don't really care whether they bought it, but gone so far down the funnel they can tell you—pick whatever tool: Fixie, Regie, GC, AIGC—I don't care how you went through it, you looked at it, you can tell me the eight ways it would improve the productivity of your business and three you didn't. Just don't hire that person because they're going to run your company to the ground. This is the job today. The job today is not to screw around on ChatGPT and to be a prompt engineer. The job today is to bring the best AI and agentic products into your organization and leverage all the hard work that the engineers have done building those products. That's your job. You don't have to screw around. You don't have to be a prompt engineer anymore. You have to be an agent deployment expert. A—this is the new job we're making up today. An Agentic Deployment Expert. That's your job from C-level to junior. Agentic Deployment Expert. Don't hire anybody else. You're going to regret it. They're going to stare at the camera. He's good. Stare at the camera. He's honorable. We could probably just I could slip away, get a coffee, and come back. No. And I I sound exasperated, Rory. And I—but the reason I am is I can just see I can see my best companies doing it. And I can see some companies I've invested in not doing it. And I want to cry. I just want to cry when they have no ADs on their team. I just—like you're flushing your years of your life down the toilet by not approaching your how you're building this company this way. [24:33] Rory: Yes. And at the risk of being positive, it's worth pointing out two things he didn't say. Well, something implicit why he said—Jason didn't do the only hire, you know, he didn't commit the um employment law, I think it's a civil penalty of saying only employ people below X who get the new new thing because he implicitly said anyone can do it provided you're willing to learn. And I think that's the big aha that's one of the positive statements to make here right? Look and I think it applies—I'm always wary of being "Hey, coming across, hey this this is the things that you all have to do." I think it applies to everyone including investors right? I mean I will say I have found that unless you're willing to invest the time learning these tools you actually shouldn't be investing in them. One of my partners Andy had this expression: "You know, if you decide you want to stop learning new things you probably should retire within 6 to 12 months and never write another check again." Maybe that's down to 3 to 6 months at this stage, right? And I think, you know, it's— [25:27] Harry Stebbings: Yeah, I actually I actually had a meeting with mine and Jason's biggest investor the other day and I—pretend he's not here—I said I think he's the most equipped investor for this generation of investing because I don't think anyone quite sits at the bleeding edge like he does on the investor side. [25:42] Harry Stebbings: Why in terms of using the equip stuff? Yeah. Yeah. In terms of using the stuff, understanding understanding bottlenecks, constraints. For sure. [25:51] Jason Lemkin: But can I just add one point? We can just cuz it's so important if it helps people. Okay, we are—and thank you Harry. We're going through these phases. Okay, and when AI started to blow up for real for us, uh call it early 2024, right? Maybe late '23, I wasn't equipped. It was too technical. I wasn't going to go in and figure out—I wasn't smart enough to figure out how to deal with a massively hallucinating LLM API and turn that and turn that into something magical. Kudos to investors and others that that got it in early '23, '22. I mean I remember I—I guess it was maybe SaaStr Annual '23. I was with David Sacks and I did a Q&A and I said, "How you thinking about AI at Craft?" He's like, "Well we're all in. We want 80% of '23 of investments to be AI." I'm like, "Great but like show me the show me the great ones in market." He's like, "They're all prototypes. We're all they're all they're all proof of concepts but we're all in anyway." That's where you kind of had to be in '23 if you weren't investing at like the LLM level. Okay, I wasn't smart enough. Then we went through this weird-ass prompt engineer era where like you you could torture these products to do something good, right? But you had to torture them. You had to like craft these crazy things that made no sense. Now we are in the era where mere ordinarily smart generalists can make these tools do magical things. And literally I go to these meetings and people be like, "I don't know how to like this is so scary. I don't know how to do this." And we show them our backends. Do you know how to do a workflow generator? Do you know how to do a a decision tree? Like we've been building these since software in the '90s. Okay, if you—I can show you all of our agents. The how they work is novel. They do have to be trained. You can't be lazy and have these agents work. But honestly, the the UI, the UX, the way we interact with them, it's just software. And so my point is: Pick yourself off the ground. This is your time now. If you felt lost in AI era, if you felt like you're behind, you don't understand what all these people are saying on X and Twitter and their Claude and and their and talking about all the 4.6 point Nano point and it's over—like you just it's not your world. This is your time. This is your time for the generalist that knows how to use software tools really really well. And I—this is my last point but it's so important. If ever in your recent life—and this is why you could be all you need to be is young at heart to Rory's point—if in the last three to five years you have successfully deployed a piece of enterprise software of any sort you yourself, not some agency you hired, but if you have deployed it, you can deploy any agentic tool. Any. And you can become the hero in your company and you can become the hero in your functional area. But I watch folks—I'm literally helping a company now that they're adding hundreds of sales folks this year with a new pre-IPO COO—he's not hasn't brought in a single tool, totally scared of it. Okay, it's not that hard. Did you use SalesLoft? Did you use Outreach? Did you use HubSpot? Do you know these tools? If you can deploy these tools, you can deploy a world-changing AI agent. And so this is the time for people like the folks that that were shut out of the AI revolution right now. The generalist folks that are not that know how to deploy software that don't even know how to build software. Like vibe coding for me was folks who knew how to build software, but you didn't have to be an engineer. Now, you just need to know how to deploy software to win with AI agents. That's all you need to know. So many people have these skills and they're petrified of AI. "How did you do that? How did you deploy an AI BDR?" Well, we bought a piece of software, we figured out how it worked for a day, we set it up in an afternoon, and then and then we did spend 30 months training it, which you didn't do with this old software because in the old days, we just had to manually upload all the data, right? And there was no training. The the only non-intuitive part is training these things. And it's it's it's just work. So that's why when I see folks on the management team not doing this, there's no excuse. You do not need to be technical to win with AI agents in Q2 of '26. You do not need to be even 1% technical. Not at all. So it's your time. Or you're going to get laid off. Or you're going to get laid off because you're not going to matter.

Arjun Mahadevan (Mr. LLC 🇺🇸)

37,852 次观看 • 5 个月前

The 40,000% ROI "Bug": How Claude Code Cracked the TradingView Holy Grail most people think the elite traders at the top of the mountain have some secret indicator or a hidden math formula that gives them a forty thousand percent return. they assume the game is rigged against the small player and that you need a multi million dollar budget just to get a seat at the table. the truth is that the holy grail of trading is actually hidden in plain sight inside a community tab that most people scroll past every single day i spent years losing money to liquidations and over trading because i thought i had to manually predict where the price was going next. i even spent hundreds of thousands of dollars on developers to build apps for me because i was convinced that i would never be able to code the systems myself. it turns out that once you stop trying to be a genius and start using the tools that are already available you can crack the code to unlimited trading strategies the secret is not in a single indicator but in the process of research back test and implement. if you go to the community section of trading view you will find an endless stream of source code for indicators that people have built over decades. most traders just slap these on a chart and hope for the best but if you are a data dog like me you know that a chart is just a pretty picture that lies to you i believe that code is the great equalizer because it allows us to take these public ideas and turn them into fully automated systems that trade for us while we sleep. i decided to learn to code live on youtube to show everyone that you can iterate your way to success without being a math wizard or a stanford graduate. now i have fully automated systems that manage my capital instead of getting liquidated by emotional decisions in the middle of the night the biggest trap in the trading world is something called repainting and it is the reason why so many strategy back tests look like they are printing money when they are actually just a scam. repainting happens when an indicator looks at future data to tell you what happened in the past which makes every buy and sell signal look like a perfect entry at the top and bottom. if you trust a back test on a basic chart without understanding the logic underneath you are just building a house on a foundation of sand this is why i transitioned all of my serious work into python because python does not lie to you. in python you can control the data flow tick by tick and bar by bar to ensure that no future data is leaking into your strategy. i built a back test architect which is a specialized sub agent that knows exactly how to take a simple idea and test it against twenty five different data sources all at once when you run a strategy across btc eth apple google and tesla you start to see the real truth about whether a strategy has an edge or if it was just a lucky fluke on one chart. i saw one strategy this week that showed a one million percent return which sounds like a total lie but the data does not have an ego. even if a number looks insane you have to investigate it and incubate it with tiny size to see if it holds up in the live market you must treat your trading like a business where you are the manager and the code is your team of tireless employees. i have sub agents running for me right now that act as masters of specific tasks like converting pine script into python or optimizing exit logic. if you are not using these specialized ai assistants in your workflow you are essentially trying to build a skyscraper with a hand saw while everyone else is using heavy machinery most people get stuck in the beginner phase because they think they need to write every single line of code from scratch. the reality is that the best developers are just really good at importing the hard work of others and connecting it like lego blocks. i use a library called ccxt that allows my bots to communicate with every major exchange in the world with just a few lines of script which saves me months of development time the reason i show everything live is because the industry is filled with gatekeepers who want to keep the secrets of automation to themselves. they want you to stay as a manual trader who pays high fees and provides liquidity for their algorithms. once you learn to automate you are no longer a victim of the market but a participant in the architecture of the financial system if you are sitting there right now feeling defeated because you just got smoked on a trade or you missed a massive pump you have to realize that those emotions are your greatest enemy. a computer does not feel fomo and it does not get tilted after a loss; it just waits for the next signal that fits the parameters you defined. my mission is to help you get to a place where you can walk away from the screen and let the machines do the heavy lifting learning to code is actually much easier than learning a second language because the syntax is logical and the feedback is immediate. i spent ten years in tech scared to touch a keyboard for anything other than emails because i thought i was not smart enough for engineering. once i realized that code is just logic i was able to build my first profitable bot within a few months and i have never looked back the transition from a manual trader to an algorithmic expert is about building a robust framework for testing your ideas as fast as possible. you want to be able to find an indicator on trading view convert it to python and run it against years of historical data in less than five minutes. if you can do that you have a higher chance of success than ninety nine percent of the people who are just drawing lines on a screen one of the most powerful strategies i found recently combines the squeeze momentum indicator with smart money concepts. when you test these individually they might show a decent return but when you combine them and add a filter like the adx you can find setups that have a massive expectancy. the key is to look for strategies that show positive returns across multiple different asset classes and time frames simultaneously even if a strategy looks like it is printing a forty thousand percent return you must always remain skeptical and look for the catch. i always incubate my new ideas with tiny capital for at least a few weeks to see how they handle real world slippage and fees. a back test is a map of the past but the live market is a wilderness that changes every single day this is why i believe in the rbi method which stands for research back test and implement. you spend your mornings looking for new ideas your afternoons stress testing them with ai and your evenings deploying the winners to the market. it is a systematic approach to wealth that removes the need for luck or guessing what a celebrity is going to tweet next the most successful traders in history like jim simons did not sit around looking at rsi levels on a fifteen minute chart. they built systems that identified mathematical edges and then scaled those systems until they were managing billions of dollars. you do not need thirty one billion dollars to change your life but you do need the discipline to stop trading like a human and start thinking like a system i give away so much for free on youtube because i want to build a community of data dogs who are all chasing the same goal of financial freedom through automation. when we work together and share our findings we can collectively identify edges that nobody else is looking at. the world is moving towards an ai dominated economy and if you are not learning to control the machines you are going to be controlled by them the road to automation is not a straight line and you will run into bugs that make you want to throw your computer out the window. but every time you fix an error and every time you optimize a script you are getting one step closer to a life where you own your time. code really is the great equalizer and it is waiting for you to pick it up and start building your own future if you can fly then run and if you can run then walk but whatever you do you must keep moving forward in this journey. trading can be heartless but the logic of code is always fair and consistent. stop being the liquidity for someone else's bot and start building the walls that will protect your capital forever
2:02:43

The 40,000% ROI "Bug": How Claude Code Cracked the TradingView Holy Grail most people think the elite traders at the top of the mountain have some secret indicator or a hidden math formula that gives them a forty thousand percent return. they assume the game is rigged against the small player and that you need a multi million dollar budget just to get a seat at the table. the truth is that the holy grail of trading is actually hidden in plain sight inside a community tab that most people scroll past every single day i spent years losing money to liquidations and over trading because i thought i had to manually predict where the price was going next. i even spent hundreds of thousands of dollars on developers to build apps for me because i was convinced that i would never be able to code the systems myself. it turns out that once you stop trying to be a genius and start using the tools that are already available you can crack the code to unlimited trading strategies the secret is not in a single indicator but in the process of research back test and implement. if you go to the community section of trading view you will find an endless stream of source code for indicators that people have built over decades. most traders just slap these on a chart and hope for the best but if you are a data dog like me you know that a chart is just a pretty picture that lies to you i believe that code is the great equalizer because it allows us to take these public ideas and turn them into fully automated systems that trade for us while we sleep. i decided to learn to code live on youtube to show everyone that you can iterate your way to success without being a math wizard or a stanford graduate. now i have fully automated systems that manage my capital instead of getting liquidated by emotional decisions in the middle of the night the biggest trap in the trading world is something called repainting and it is the reason why so many strategy back tests look like they are printing money when they are actually just a scam. repainting happens when an indicator looks at future data to tell you what happened in the past which makes every buy and sell signal look like a perfect entry at the top and bottom. if you trust a back test on a basic chart without understanding the logic underneath you are just building a house on a foundation of sand this is why i transitioned all of my serious work into python because python does not lie to you. in python you can control the data flow tick by tick and bar by bar to ensure that no future data is leaking into your strategy. i built a back test architect which is a specialized sub agent that knows exactly how to take a simple idea and test it against twenty five different data sources all at once when you run a strategy across btc eth apple google and tesla you start to see the real truth about whether a strategy has an edge or if it was just a lucky fluke on one chart. i saw one strategy this week that showed a one million percent return which sounds like a total lie but the data does not have an ego. even if a number looks insane you have to investigate it and incubate it with tiny size to see if it holds up in the live market you must treat your trading like a business where you are the manager and the code is your team of tireless employees. i have sub agents running for me right now that act as masters of specific tasks like converting pine script into python or optimizing exit logic. if you are not using these specialized ai assistants in your workflow you are essentially trying to build a skyscraper with a hand saw while everyone else is using heavy machinery most people get stuck in the beginner phase because they think they need to write every single line of code from scratch. the reality is that the best developers are just really good at importing the hard work of others and connecting it like lego blocks. i use a library called ccxt that allows my bots to communicate with every major exchange in the world with just a few lines of script which saves me months of development time the reason i show everything live is because the industry is filled with gatekeepers who want to keep the secrets of automation to themselves. they want you to stay as a manual trader who pays high fees and provides liquidity for their algorithms. once you learn to automate you are no longer a victim of the market but a participant in the architecture of the financial system if you are sitting there right now feeling defeated because you just got smoked on a trade or you missed a massive pump you have to realize that those emotions are your greatest enemy. a computer does not feel fomo and it does not get tilted after a loss; it just waits for the next signal that fits the parameters you defined. my mission is to help you get to a place where you can walk away from the screen and let the machines do the heavy lifting learning to code is actually much easier than learning a second language because the syntax is logical and the feedback is immediate. i spent ten years in tech scared to touch a keyboard for anything other than emails because i thought i was not smart enough for engineering. once i realized that code is just logic i was able to build my first profitable bot within a few months and i have never looked back the transition from a manual trader to an algorithmic expert is about building a robust framework for testing your ideas as fast as possible. you want to be able to find an indicator on trading view convert it to python and run it against years of historical data in less than five minutes. if you can do that you have a higher chance of success than ninety nine percent of the people who are just drawing lines on a screen one of the most powerful strategies i found recently combines the squeeze momentum indicator with smart money concepts. when you test these individually they might show a decent return but when you combine them and add a filter like the adx you can find setups that have a massive expectancy. the key is to look for strategies that show positive returns across multiple different asset classes and time frames simultaneously even if a strategy looks like it is printing a forty thousand percent return you must always remain skeptical and look for the catch. i always incubate my new ideas with tiny capital for at least a few weeks to see how they handle real world slippage and fees. a back test is a map of the past but the live market is a wilderness that changes every single day this is why i believe in the rbi method which stands for research back test and implement. you spend your mornings looking for new ideas your afternoons stress testing them with ai and your evenings deploying the winners to the market. it is a systematic approach to wealth that removes the need for luck or guessing what a celebrity is going to tweet next the most successful traders in history like jim simons did not sit around looking at rsi levels on a fifteen minute chart. they built systems that identified mathematical edges and then scaled those systems until they were managing billions of dollars. you do not need thirty one billion dollars to change your life but you do need the discipline to stop trading like a human and start thinking like a system i give away so much for free on youtube because i want to build a community of data dogs who are all chasing the same goal of financial freedom through automation. when we work together and share our findings we can collectively identify edges that nobody else is looking at. the world is moving towards an ai dominated economy and if you are not learning to control the machines you are going to be controlled by them the road to automation is not a straight line and you will run into bugs that make you want to throw your computer out the window. but every time you fix an error and every time you optimize a script you are getting one step closer to a life where you own your time. code really is the great equalizer and it is waiting for you to pick it up and start building your own future if you can fly then run and if you can run then walk but whatever you do you must keep moving forward in this journey. trading can be heartless but the logic of code is always fair and consistent. stop being the liquidity for someone else's bot and start building the walls that will protect your capital forever

Moon Dev

245,471 次观看 • 6 个月前

🚨Trump's UFO Disclosure Directive: What It Changes And What It Doesn't There's a huge difference between a political moment and a structural shift, and right now a lot of people are confusing the two things. Donald Trump's recent statement directing agencies to review and release UFO and UAP files has lit up the conversation again. Headlines are flying, and social media is treating it like the dam just broke, but if you just slow down and look at how government actually works, the scene is more procedural than revolutionary. Trumps messaged that the government would begin identifying and reviewing government files related to UFOs, UAPs, and potential extraterrestrial life for possible release. I understand that sounds dramatic and exciting, but we need to separate tone from substance here. A review order is not the same thing as a confirmation. It's an administrative instruction to inventory and assess records, it doesn't declare what those records contain. Trump also made something else clear in his remarks he said he does not know whether aliens are real. If a president had been briefed on confirmed non human contact, why is he saying that? What we're seeing from Trump is an acknowledgment of public interest combined with a push for transparency, and transparency is not disclosure in the way many people use that word. If this directive moves forward, agencies like the Department of Defense, NASA, and the Office of the Director of National Intelligence will begin pulling relevant records. That likely includes historical UAP sighting reports, aviation incident logs, internal analysis memos, and summary level assessments that don't expose sensitive capabilities. The key word is here is 'review.' Every document still has to go through declassification protocols, and those protocols are not optional just because Trump issued a directive. Anything that reveals intelligence collection methods, sensor capabilities, satellite coverage gaps, or defense vulnerabilities will obviously remain protected due to national security. Classification frameworks exist to prevent adversaries from learning how the United States detects, tracks, and responds to unknown objects in its airspace. We've already seen how this plays out in recent years. The Pentagon confirmed Navy cockpit footage, then Congress held hearings. The All domain Anomaly Resolution Office (AARO) was created to formalize reporting and analysis. Annual summaries were released stating that many incidents remain unexplained but do not constitute verified extraterrestrial evidence. That same pattern of acknowledgment without confirmation has been consistent, and I hate to say it but Trump's directive fits within that pattern. What could change is the volume of information entering the public domain. More documents might be released and more historical material could surface. Researchers may get access to broader datasets. That can shift the cultural landscape. The topic becomes less radioactive and off the back of that universities engage in research, journalists treat it as a legitimate national security issue rather than a punchline. Hard disclosure would require synchronized statements from multiple agencies confirming origin, not just anomaly. It would involve authenticated multi sensor data with full chain of custody. It would likely include independent scientific panels granted access to raw evidence. There would be formal briefings to congressional leadership before anything was said publicly. Terminology across agencies would change permanently. We are not seeing those indicators as of yet. There is also the Special Access Program problem. If, hypothetically, highly sensitive aerospace material were ever tied to deeply compartmentalized programs, those would be inside restricted frameworks designed to limit visibility. Even members of Congress must be read into certain programs. Oversight exists, but it operates through controlled channels. That architecture isn't going to change because of a social media announcement, doesn't matter who its from. Congress does have leverage through funding authority and legislative mandates. It can require reporting, audits, and classified briefings. It cannot easily force public release of material deemed damaging to national security. Courts historically defer to executive classification authority in those matters. That means even if lawmakers receive classified insight, what the public gets is often a watered down summary. So does the tweet change things? Yes, in the sense that it formalizes executive interest in transparency. That can accelerate review proceses and potentially increase public access to records. It signals that the subject has moved firmly into mainstream political discourse. No, in the sense that it does not confirm the existence of extraterrestrial life, recovered craft, or hidden biological evidence. It does not override classification law. It does not automatically open Special Access Programs to public inspection. The government has shifted from ridicule to structured acknowledgment. That's real progress because it means that the stigma is finally fading. The institutional language has softened. But until multiple agencies move in coordinated fashion with authenticated evidence and irreversible terminology changes, we are not looking at hard disclosure. If this directive leads to meaningful document releases, that fantastic. If it results in heavily redacted summaries and familiar language about unexplained but unverified phenomena, then it will simply confirm what we already know, that transparency comes from a slow drip. #UFO #UAP #Disclosure #NationalSecurity #Congress #Transparency
0:47

🚨Trump's UFO Disclosure Directive: What It Changes And What It Doesn't There's a huge difference between a political moment and a structural shift, and right now a lot of people are confusing the two things. Donald Trump's recent statement directing agencies to review and release UFO and UAP files has lit up the conversation again. Headlines are flying, and social media is treating it like the dam just broke, but if you just slow down and look at how government actually works, the scene is more procedural than revolutionary. Trumps messaged that the government would begin identifying and reviewing government files related to UFOs, UAPs, and potential extraterrestrial life for possible release. I understand that sounds dramatic and exciting, but we need to separate tone from substance here. A review order is not the same thing as a confirmation. It's an administrative instruction to inventory and assess records, it doesn't declare what those records contain. Trump also made something else clear in his remarks he said he does not know whether aliens are real. If a president had been briefed on confirmed non human contact, why is he saying that? What we're seeing from Trump is an acknowledgment of public interest combined with a push for transparency, and transparency is not disclosure in the way many people use that word. If this directive moves forward, agencies like the Department of Defense, NASA, and the Office of the Director of National Intelligence will begin pulling relevant records. That likely includes historical UAP sighting reports, aviation incident logs, internal analysis memos, and summary level assessments that don't expose sensitive capabilities. The key word is here is 'review.' Every document still has to go through declassification protocols, and those protocols are not optional just because Trump issued a directive. Anything that reveals intelligence collection methods, sensor capabilities, satellite coverage gaps, or defense vulnerabilities will obviously remain protected due to national security. Classification frameworks exist to prevent adversaries from learning how the United States detects, tracks, and responds to unknown objects in its airspace. We've already seen how this plays out in recent years. The Pentagon confirmed Navy cockpit footage, then Congress held hearings. The All domain Anomaly Resolution Office (AARO) was created to formalize reporting and analysis. Annual summaries were released stating that many incidents remain unexplained but do not constitute verified extraterrestrial evidence. That same pattern of acknowledgment without confirmation has been consistent, and I hate to say it but Trump's directive fits within that pattern. What could change is the volume of information entering the public domain. More documents might be released and more historical material could surface. Researchers may get access to broader datasets. That can shift the cultural landscape. The topic becomes less radioactive and off the back of that universities engage in research, journalists treat it as a legitimate national security issue rather than a punchline. Hard disclosure would require synchronized statements from multiple agencies confirming origin, not just anomaly. It would involve authenticated multi sensor data with full chain of custody. It would likely include independent scientific panels granted access to raw evidence. There would be formal briefings to congressional leadership before anything was said publicly. Terminology across agencies would change permanently. We are not seeing those indicators as of yet. There is also the Special Access Program problem. If, hypothetically, highly sensitive aerospace material were ever tied to deeply compartmentalized programs, those would be inside restricted frameworks designed to limit visibility. Even members of Congress must be read into certain programs. Oversight exists, but it operates through controlled channels. That architecture isn't going to change because of a social media announcement, doesn't matter who its from. Congress does have leverage through funding authority and legislative mandates. It can require reporting, audits, and classified briefings. It cannot easily force public release of material deemed damaging to national security. Courts historically defer to executive classification authority in those matters. That means even if lawmakers receive classified insight, what the public gets is often a watered down summary. So does the tweet change things? Yes, in the sense that it formalizes executive interest in transparency. That can accelerate review proceses and potentially increase public access to records. It signals that the subject has moved firmly into mainstream political discourse. No, in the sense that it does not confirm the existence of extraterrestrial life, recovered craft, or hidden biological evidence. It does not override classification law. It does not automatically open Special Access Programs to public inspection. The government has shifted from ridicule to structured acknowledgment. That's real progress because it means that the stigma is finally fading. The institutional language has softened. But until multiple agencies move in coordinated fashion with authenticated evidence and irreversible terminology changes, we are not looking at hard disclosure. If this directive leads to meaningful document releases, that fantastic. If it results in heavily redacted summaries and familiar language about unexplained but unverified phenomena, then it will simply confirm what we already know, that transparency comes from a slow drip. #UFO #UAP #Disclosure #NationalSecurity #Congress #Transparency

Skywatch Signal

15,874 次观看 • 6 个月前

Here's a devlog made by an anonymous Chinese fan replicating the surprisingly brand new technique that I developed for detecting asteroids which wound up being so powerful that it can easily track Stealth Fighters from over 100km away even when it’s only using three $30 webcams as sensors meaning it easily outperforms all modern stealth tracking techniques in precision, range and cost. And while this demo is using optical light, this same technique which I call pixel motion to voxel projection, can be used interchangeably with thermal infrared cameras to work at night and also majorly boosts the effectiveness of radar allowing you to track fighters much more effectively through clouds and over the horizon. This technique will also always eventually give the exact location of the target even if the image is blurry as those blurs will always average out from the different perspectives into revealing the precise location of the target in the voxel grid. There is definitely a Mandela effect with this technique as it feels as though it should already exist, especially because at first as it sounds like it is performing triangulation (which has existed for years and is what we do for mocap and tennis ball tracking). But triangulation is entirely separate to this as triangulations only works if you have already identified where the ball is in a 2D image because you’re able to rely on being able to use at least 2 separate high quality cameras which are much closer to the ball making the ball’s apparent size much much bigger and therefore gives you hundreds of pixels to work with which makes it much easier to use object recognition techniques to recognize where it is in the image aka in 2D and then you’re just using the other cameras view to project out lines which intersect in 3D to find out where the ball is in 3D. The major difference is that pixel motion to voxel projection allows you to find where the object is in 3D without having already found it in 2D which is an unbelievable difference as it allows you to use much lower quality cameras together to accumulate data together into 3D space. If this seem like it doesn’t mean much then what it actually means is that you don’t understand what I’m saying as what I’m saying means a LOT in practical terms as it means you go from having to use an imaging system that has to be able to image the object to the point that it is over a hundred total pixels in surface area to have enough data to recognize it to instead be able to use something that is only images the object to be 1 pixel in surface area and only changes the brightness value by 1 value every now and then. I’d recommend an amazing video by DST studios called “Lowlight cameras can’t defeat stealth” if you want a great video which goes over the difficulty of even using telescopes to recognize stealth fighters and why this is so impressive compared to other techniques and ironically it is what inspired me to realize the asteroid tracker I was working on actually could do this. Which brings me to the point that if this wasn’t a new technique then not only would there be at least one example of an asteroid survey that points distant telescopes at the same place at the same time in order to be able to add the light together to detect asteroids which as I was shocked to learn isn’t a thing despite the fact that it would make detecting asteroids trivial by comparison to modern 2D imaging while also having no impact on the normal scientific operations of those surveys other than small changes to scheduling. But there would also be an example of a drone tracker that uses this instead of using the aforementioned high quality zoomable telescope which has to be able to zoom in close enough to be able to recognize a drone. If you want to tell me that this is something that already exists give me an exact example of a product that uses it, not the general outline of a concept that you think it is, the actual product and then also tell me the asteroid survey that uses distant telescopes that point at the exact same place at the exact same time because I can guarantee that if you google what you think uses this you won’t even find the steps of subtracting the images from each other to get motion and will definitely not get the added step of projecting that motion into a voxel grid (It would blow your mind if you found out how Xbox kinect cameras work.) Also I want to make it clear, I’m not saying you should just use web cams to do this, I’m just using them as an example to show you the power of this in reality you would probably want to use 5 high quality zoomable thermal cameras which pan across the sky in sync with each other which due to using lower frequency are much less prone to the Rayleigh scattering that scatters visible light at 150 or so km away and again, you can also use this to majorly upgrade radar. Pretty much all of the problems you could think of for this are incredibly easy to overcome if you apply even a small amount of brainpower into fixing the problem. And yes, this gives you the exact location down to the meter of whatever you are tracking even if the image is blurry as those blurs will always average out to the exact location down to the meter in the voxel grid. Which is what makes this technique so powerful since the cost of adding each camera to The network grows linearly while the rate at which each camera gives more information grows exponentially due to the increasing unlikeliness of all of them having more movement in the same place. And given the size of the cameras it really wouldn’t be that hard to hide and network these cameras together in other countries and on sea buoys to know where planes are everywhere in the world. Which brings me to the point that I personally really don’t care about the military uses of this technology, if all it could do is precisely track stealth fighters then I wouldn’t have cared enough to work on it, I could have used any of the many other life saving techniques as the subject of the video, stealth fighters just sounds the most clickable and the scale of the problem is more intuitive to most people and if I did use any of those as subjects for the demo it would inevitably result in the stealth fighter technique being figured out anyway and all of the other uses are so useful that I don't think anyone would reasonably complain about the upside. The real purpose of this video is that since this is a new technique that hasn’t been used to detect stealth fighters despite the billions we have spent on that, then what else can you apply this to that could go on to improve billions of people’s lives that you or others are working on. For example this also allows you to majorly improve the effectiveness of cryo electron microscopy and CT scanners. This part also is kind of hard to explain as it also sounds like it exists but again, when you look through all of the places where you think it is being used you will find that it wasn’t. What I’m saying here isn’t that this is a Radon transform or gaussian splat or whatever, I’m saying that this is able to get new information that wasn’t being accessed before due to the added information about depth you get from the correlation of movement between each perspective which adds to the information that you already have. This allows you to directly subtract foreground and background objects as well as noise faster than you would be able to before and works better than super resolution for your images since super resolution won’t remove foreground and background objects like this does and instead just scales up target, foreground and background objects indiscriminately. And while with enough data Radon transforms or other scanning techniques would eventually get you a correct answer this will get you there a lot faster since those are mostly averaging techniques which average out noise whereas this gets you the ability to directly subtract noise. I’m not expecting you to think that this would do anything but if you try it for yourself you will find that it does majorly improve your ability to perform 3d scans. Again, cryo EM is a field where you would expect this technique to exist but when you look through all the papers on the topic there is no mention of tilting the grid slightly in order to be able to change your perspective slightly on the order of the feature size (if you tilt the grid then you only need precision on the order of an arc minute to do this) and doing multiple exposures from multiple different known tilts and then using those difference images to correlate depth from motion. In fact, in cryo EM you would normally want to do the opposite of this and have your exposures all taken from the same grid angle and just use the variations in how many of the same proteins are oriented in order to be able to scan them for a 3D model but this will generate you far more data faster. There is so much information that I can’t really explain in text so if you have any questions such as why this hasn’t been made before then they will most likely be answered in the video I originally posted which I have added to the end of the first Devlog for your convenience. And again, pretty much all of the problems with the technique can be fixed with a little bit of brainpower, in reality you would probably want to use 5 high quality zoomable thermal cameras which pan across the sky in sync with each other which due to using lower frequency are much less prone to the Rayleigh scattering that scatters visible light at 150 or so km away and again, you can also use this to majorly upgrade radar.
18:48

Here's a devlog made by an anonymous Chinese fan replicating the surprisingly brand new technique that I developed for detecting asteroids which wound up being so powerful that it can easily track Stealth Fighters from over 100km away even when it’s only using three $30 webcams as sensors meaning it easily outperforms all modern stealth tracking techniques in precision, range and cost. And while this demo is using optical light, this same technique which I call pixel motion to voxel projection, can be used interchangeably with thermal infrared cameras to work at night and also majorly boosts the effectiveness of radar allowing you to track fighters much more effectively through clouds and over the horizon. This technique will also always eventually give the exact location of the target even if the image is blurry as those blurs will always average out from the different perspectives into revealing the precise location of the target in the voxel grid. There is definitely a Mandela effect with this technique as it feels as though it should already exist, especially because at first as it sounds like it is performing triangulation (which has existed for years and is what we do for mocap and tennis ball tracking). But triangulation is entirely separate to this as triangulations only works if you have already identified where the ball is in a 2D image because you’re able to rely on being able to use at least 2 separate high quality cameras which are much closer to the ball making the ball’s apparent size much much bigger and therefore gives you hundreds of pixels to work with which makes it much easier to use object recognition techniques to recognize where it is in the image aka in 2D and then you’re just using the other cameras view to project out lines which intersect in 3D to find out where the ball is in 3D. The major difference is that pixel motion to voxel projection allows you to find where the object is in 3D without having already found it in 2D which is an unbelievable difference as it allows you to use much lower quality cameras together to accumulate data together into 3D space. If this seem like it doesn’t mean much then what it actually means is that you don’t understand what I’m saying as what I’m saying means a LOT in practical terms as it means you go from having to use an imaging system that has to be able to image the object to the point that it is over a hundred total pixels in surface area to have enough data to recognize it to instead be able to use something that is only images the object to be 1 pixel in surface area and only changes the brightness value by 1 value every now and then. I’d recommend an amazing video by DST studios called “Lowlight cameras can’t defeat stealth” if you want a great video which goes over the difficulty of even using telescopes to recognize stealth fighters and why this is so impressive compared to other techniques and ironically it is what inspired me to realize the asteroid tracker I was working on actually could do this. Which brings me to the point that if this wasn’t a new technique then not only would there be at least one example of an asteroid survey that points distant telescopes at the same place at the same time in order to be able to add the light together to detect asteroids which as I was shocked to learn isn’t a thing despite the fact that it would make detecting asteroids trivial by comparison to modern 2D imaging while also having no impact on the normal scientific operations of those surveys other than small changes to scheduling. But there would also be an example of a drone tracker that uses this instead of using the aforementioned high quality zoomable telescope which has to be able to zoom in close enough to be able to recognize a drone. If you want to tell me that this is something that already exists give me an exact example of a product that uses it, not the general outline of a concept that you think it is, the actual product and then also tell me the asteroid survey that uses distant telescopes that point at the exact same place at the exact same time because I can guarantee that if you google what you think uses this you won’t even find the steps of subtracting the images from each other to get motion and will definitely not get the added step of projecting that motion into a voxel grid (It would blow your mind if you found out how Xbox kinect cameras work.) Also I want to make it clear, I’m not saying you should just use web cams to do this, I’m just using them as an example to show you the power of this in reality you would probably want to use 5 high quality zoomable thermal cameras which pan across the sky in sync with each other which due to using lower frequency are much less prone to the Rayleigh scattering that scatters visible light at 150 or so km away and again, you can also use this to majorly upgrade radar. Pretty much all of the problems you could think of for this are incredibly easy to overcome if you apply even a small amount of brainpower into fixing the problem. And yes, this gives you the exact location down to the meter of whatever you are tracking even if the image is blurry as those blurs will always average out to the exact location down to the meter in the voxel grid. Which is what makes this technique so powerful since the cost of adding each camera to The network grows linearly while the rate at which each camera gives more information grows exponentially due to the increasing unlikeliness of all of them having more movement in the same place. And given the size of the cameras it really wouldn’t be that hard to hide and network these cameras together in other countries and on sea buoys to know where planes are everywhere in the world. Which brings me to the point that I personally really don’t care about the military uses of this technology, if all it could do is precisely track stealth fighters then I wouldn’t have cared enough to work on it, I could have used any of the many other life saving techniques as the subject of the video, stealth fighters just sounds the most clickable and the scale of the problem is more intuitive to most people and if I did use any of those as subjects for the demo it would inevitably result in the stealth fighter technique being figured out anyway and all of the other uses are so useful that I don't think anyone would reasonably complain about the upside. The real purpose of this video is that since this is a new technique that hasn’t been used to detect stealth fighters despite the billions we have spent on that, then what else can you apply this to that could go on to improve billions of people’s lives that you or others are working on. For example this also allows you to majorly improve the effectiveness of cryo electron microscopy and CT scanners. This part also is kind of hard to explain as it also sounds like it exists but again, when you look through all of the places where you think it is being used you will find that it wasn’t. What I’m saying here isn’t that this is a Radon transform or gaussian splat or whatever, I’m saying that this is able to get new information that wasn’t being accessed before due to the added information about depth you get from the correlation of movement between each perspective which adds to the information that you already have. This allows you to directly subtract foreground and background objects as well as noise faster than you would be able to before and works better than super resolution for your images since super resolution won’t remove foreground and background objects like this does and instead just scales up target, foreground and background objects indiscriminately. And while with enough data Radon transforms or other scanning techniques would eventually get you a correct answer this will get you there a lot faster since those are mostly averaging techniques which average out noise whereas this gets you the ability to directly subtract noise. I’m not expecting you to think that this would do anything but if you try it for yourself you will find that it does majorly improve your ability to perform 3d scans. Again, cryo EM is a field where you would expect this technique to exist but when you look through all the papers on the topic there is no mention of tilting the grid slightly in order to be able to change your perspective slightly on the order of the feature size (if you tilt the grid then you only need precision on the order of an arc minute to do this) and doing multiple exposures from multiple different known tilts and then using those difference images to correlate depth from motion. In fact, in cryo EM you would normally want to do the opposite of this and have your exposures all taken from the same grid angle and just use the variations in how many of the same proteins are oriented in order to be able to scan them for a 3D model but this will generate you far more data faster. There is so much information that I can’t really explain in text so if you have any questions such as why this hasn’t been made before then they will most likely be answered in the video I originally posted which I have added to the end of the first Devlog for your convenience. And again, pretty much all of the problems with the technique can be fixed with a little bit of brainpower, in reality you would probably want to use 5 high quality zoomable thermal cameras which pan across the sky in sync with each other which due to using lower frequency are much less prone to the Rayleigh scattering that scatters visible light at 150 or so km away and again, you can also use this to majorly upgrade radar.

ConsistentlyInconsistent

50,799 次观看 • 1 年前

$NVDA $MU $SNDK $LITE PAPER OVERVIEW AND CORE CLAIMS The paper “KV Cache Transform Coding for Compact Storage in LLM Inference” introduces kvtc, a transform-coding pipeline that compresses transformer key-value (KV) caches primarily for storage and transfer in LLM serving, rather than for accelerating the per-token attention kernel during active decoding. The method combines 3 stages: (1) feature decorrelation via a PCA basis computed from a calibration dataset and reused across requests; (2) adaptive, variable-precision quantization with bit allocation solved via dynamic programming (DP), including groupwise scaling/shift overhead; and (3) lossless entropy coding (DEFLATE via nvCOMP in the reference implementation) to exploit residual redundancy after quantization. The central empirical claim is that KV tensors contain large, exploitable redundancy across heads and layers, enabling approximately 20× compression versus a 16-bit baseline with negligible degradation across a broad set of accuracy and long-context benchmarks, with materially higher compression (≥40×) available at modest quality cost in some regimes. The system claim is that such compression materially improves the economics of multi-turn, prefix-reuse serving by extending effective KV cache capacity in GPU HBM and host tiers (DRAM/NVMe) and by reducing inter-node and GPU↔host bandwidth demands, thereby improving cache hit rates and reducing time-to-first-token (TTFT) relative to recomputation when caches would otherwise be evicted. KV CACHE AS THE DOMINANT STATE VARIABLE IN INFERENCE ECONOMICS KV cache growth is linear in context length and is multiplicative in layers and attention heads, making it an increasingly dominant constraint as (a) context lengths expand, (b) models add layers and maintain large hidden dimensions, and (c) production workloads shift toward iterative and tool-augmented interactions that repeatedly reuse long prefixes. The paper uses the canonical 16-bit KV cache size formula (4·l·h·d_head·t) bytes and reports 16-bit KV cache sizes per 1K tokens of context that are already operationally large: 128MiB for Llama 3.1 8B, 160MiB for Mistral NeMo 12B, and 320MiB for Llama 3.3 70B Instruct. In binary units, these figures imply per-token KV footprints of 128KiB/token (Llama 3.1 8B), 160KiB/token (Mistral NeMo 12B), and 320KiB/token (Llama 3.3 70B Instruct) at 16-bit. For a 10K-token prompt (10×1K in the paper’s binary convention), the 16-bit KV cache sizes scale to approximately 1.25GiB (Llama 3.1 8B), 1.56GiB (Mistral NeMo 12B), and 3.13GiB (Llama 3.3 70B Instruct). These magnitudes explain why stale caches create a throughput–latency dilemma: retaining them in HBM maximizes responsiveness on future turns but crowds out concurrent sessions; evicting them forces quadratic-cost prefill recomputation and increases TTFT; offloading them to host or storage introduces large transfer overhead and consumes DRAM/NVMe capacity. A key operational nuance emphasized is that modern serving stacks increasingly treat KV caches as a database, leveraging block paging and shared-prefix reuse. In the common disaggregated serving design (separate prefill and decode nodes), KV cache transfer becomes a dominant category of cross-node traffic. Under that design, any reduction in KV cache size directly increases effective fabric capacity and reduces tail latency attributable to congestion, while also enabling longer cache lifetimes in “hot” (HBM) and “warm” (CPU DRAM) tiers that raise cache hit rates and reduce recomputation frequency. The paper’s quantitative example illustrates the economic stakes: a 1,000-line code file tokenized at ~10 tokens/line yields ~10K tokens; for Llama 3.3 70B, an 8-bit KV cache for that context is ~1.6GiB. Reuse across subsequent turns or parallel chats around the same file is valuable, but HBM scarcity makes retaining many such caches infeasible without compression. TECHNICAL MECHANISM: WHY KV CACHES ARE COMPRESSIBLE AND HOW KVTC EXPLOITS IT The technical rationale begins with an empirical observation: keys (and, to a lesser extent, values) across different attention heads can be aligned into a shared latent space using orthogonal transformations (Procrustes alignment). This supports the hypothesis that head-specific projections introduce rotations of a common subspace rather than completely distinct information, implying that concatenating across heads and layers should reveal low-rank structure suitable for linear decorrelation and dimensionality reduction. The method operationalizes this using a PCA/SVD basis learned from calibration data rather than recomputing a decomposition per prompt. This design choice targets production viability: per-prompt SVD is computationally expensive and scales poorly with long prompts and frequent cache updates. kvtc is explicitly structured as an offline-calibrated, online-applied codec: Calibration (performed 1 time per model and compression setting for DP allocation) A calibration dataset is forwarded through the model to collect KV caches. Token positions are pooled, and a subset of positions is sampled. Keys and values are processed separately. Several implementation choices are highlighted as decisive for stability: Rotary positional embeddings are effectively removed prior to compression (“undo positional rotations”), because positional rotations degrade the apparent low-rank structure of keys. “Attention sink” tokens (the earliest tokens in the sequence) and a sliding window of most recent tokens are excluded from compression because they disproportionately affect attention patterns and are empirically more sensitive to reconstruction error. Cross-layer concatenation is used: keys (or values) from multiple layers and heads at the same token position are concatenated along the feature axis to form a higher-dimensional feature vector. PCA is computed over these concatenated vectors, improving robustness relative to per-layer or per-head PCA. The PCA basis is computed via SVD of centered calibration data, using randomized SVD for scalability with a target rank cutoff. The paper reports calibration regimes of 160K tokens for several models with a 10K PCA dimension cutoff (8K for Qwen variants with fewer KV heads), selected to fit within a single 80GB H100 memory envelope and complete within minutes. A critical economic detail is that the same PCA basis can be reused across multiple compression ratios; only the DP-derived precision assignment changes per compression target. Compression (applied between inference phases) Compression operates on stored KV cache tensors, not on weights, and does not modify attention computation. The KV cache is projected into the PCA basis, quantized, packed, and then entropy-coded. Compression is positioned as a background or between-phase operation (after decoding, or between prefill and decode), executed on GPU or CPU depending on where the cache currently resides. The design intent is that compression should not sit on the critical per-token decoding path; it is a storage and transport optimization. Decompression (performed prior to reuse) Decompression reverses the entropy coding and quantization and applies the inverse PCA projection. A practical latency optimization is proposed: inverse projection can be performed layer-by-layer using submatrices of the PCA basis, allowing generation to begin before the full cache is reconstructed, reducing TTFT. Quantization and bit allocation are the core differentiators versus simpler PCA truncation. PCA provides ordered components by variance; kvtc uses DP to allocate a global bit budget across PCA coordinates (and across groups of coordinates) to minimize reconstruction error in the decorrelated domain. Groups of subsequent PCA coordinates share 16-bit shift and scale factors (a microscaling-inspired design), and the DP algorithm jointly selects group size and precision type under a bit budget, including the overhead of per-group metadata. DP commonly assigns 0 bits to many trailing PCA components, which both increases compression and provides a mechanism to trim the PCA basis to the subset of components that actually carry payload, reducing compute and storage overhead of the projection matrices in deployment. Lossless entropy coding then exploits the structure induced by quantization. DEFLATE is used in the reference implementation, and the paper emphasizes that the incremental gain from the lossless stage is content-dependent but meaningful, with an average uplift of ~1.23× on top of quantization in the reported regime. An ablation in the appendices indicates that GPU-friendly variants (GDeflate) can achieve nearly identical compression ratios (≤0.1 difference in measured cases), implying that throughput-optimized lossless codecs can likely be substituted without sacrificing meaningful compression. EMPIRICAL RESULTS: ACCURACY, COMPRESSION, AND LATENCY General-purpose 8B–12B dense models The paper evaluates Llama 3.1 8B, MN-Minitron 8B, and Mistral NeMo 12B across math/knowledge (GSM8K, MMLU) and long-context tasks (Qasper, Lost in the Middle, RULER Variable Tracking) under a simulated multi-turn regime where compression/decompression is applied periodically, with a sliding window of recent tokens excluded. A consistent pattern appears: kvtc maintains near-vanilla performance through 16× compression settings, and remains competitive at 32×, with degradation becoming task- and model-dependent at 64×, particularly on long-context retrieval metrics when compression is pushed aggressively. Selected quantitative anchor points from the paper’s standard-error table (all values are reported with the paper’s evaluation setup and token-window exclusions): Llama 3.1 8B Vanilla: GSM8K 56.8, MMLU 60.5, Qasper 40.4, LITM 99.4, RULER-VT 99.8 kvtc16×: GSM8K 56.9, MMLU 60.1, Qasper 40.7, LITM 99.3, RULER-VT 99.1 kvtc32×: GSM8K 57.8, MMLU 60.6, Qasper 39.4, LITM 99.1, RULER-VT 98.9 kvtc64×: GSM8K 57.2, MMLU 60.7, Qasper 37.8, LITM 90.2, RULER-VT 95.9 These results indicate that, for this model, long-context sensitivity emerges at 64× with meaningful drops in LITM and RULER-VT, while math/knowledge scores remain stable, implying a differential sensitivity consistent with key-vector precision being more critical for retrieval-style behavior. Mistral NeMo 12B Vanilla: GSM8K 61.9, MMLU 64.5, Qasper 38.4, LITM 99.5, RULER-VT 99.8 kvtc16×: GSM8K 62.0, MMLU 64.4, Qasper 37.6, LITM 99.8, RULER-VT 99.5 kvtc32×: GSM8K 62.2, MMLU 63.8, Qasper 37.5, LITM 99.6, RULER-VT 98.7 kvtc64×: GSM8K 61.9, MMLU 61.4, Qasper 38.0, LITM 95.3, RULER-VT 98.0 Here, degradation at 64× is visible but materially smaller than the Llama 3.1 8B LITM drop, suggesting model-architecture or training-data differences can change the tolerance envelope for aggressive KV cache distortion. MN-Minitron 8B Vanilla: GSM8K 59.1, MMLU 64.3, Qasper 38.2, LITM 99.8, RULER-VT 99.4 kvtc16×: GSM8K 60.3, MMLU 64.1, Qasper 38.6, LITM 99.3, RULER-VT 98.8 kvtc32×: GSM8K 59.1, MMLU 63.7, Qasper 37.7, LITM 86.9, RULER-VT 96.0 kvtc64×: GSM8K 57.8, MMLU 62.1, Qasper 38.1, LITM 59.5, RULER-VT 93.4 This model shows markedly higher sensitivity on LITM at 32× and 64×, despite stable short-context metrics, reinforcing that “compression safety” is not monotonic in parameter count and that pruning/distillation choices can alter KV cache redundancy or robustness. Comparisons to baselines The paper compares kvtc to quantization baselines (KIVI, GEAR, FP8) and eviction baselines (H2O, TOVA), plus an SVD-based prefill-optimization method (xKV). Across the reported tasks: Low-bit quantization methods at modest compression (2-bit KV schemes) show earlier degradation in long-context behavior than kvtc at substantially higher compression settings. Eviction methods perform poorly as generic compressors for long-context tasks, consistent with their objective function (selective pruning) being misaligned with “lossless-ish storage for reuse.” xKV shows competitive results on some tasks but a consistent underperformance on Qasper relative to kvtc and vanilla in the provided tables, consistent with method-specific distortions introduced by its decomposition regime. Reasoning models and high-variance tasks For DeepSeek-R1-distilled Qwen 2.5 reasoning models, the paper evaluates AIME 2024/2025 and LiveCodeBench coding. Results are averaged over 8 runs with large variance, but a key inference is that kvtc at ~9×–21× compression achieves broadly similar AIME scores within variance bands, while coding performance remains stable at ~9× and degrades more visibly at ~18×–21× on the 7B model. An important nuance is that smaller reasoning models already have smaller KV footprints (reported ~29KiB/token for Qwen R1 1.5B versus 131KiB/token for Llama 3.1 8B), so the economic value of aggressive KV cache compression is proportionally higher for large models and long contexts than for small models with short contexts, unless the serving system’s bottleneck is dominated by cache transfer rather than HBM capacity. Multi-GPU inference and pipeline parallel For Llama 3.3 70B Instruct run pipeline-parallel across 4 GPUs (20 layers per GPU), the paper compresses KV cache chunks independently per GPU. On MATH-500, the reported accuracy declines from 75.6 (vanilla) to 74.4 at 10× and 72.6 at 20×, with standard errors near ~1.9. NIAH and LITM remain at 100.0 for all tested ratios in that table. The paper notes that joint compression across chunks could improve accuracy for some offload scenarios but is not required for feasibility, highlighting an engineering trade-off between deployment simplicity in distributed settings and optimal global compression. Latency and TTFT economics A critical system result is the measured compression/decompression latency on an H100 for a non-fused implementation. For Mistral NeMo 12B in bfloat16: BS=8, CTX=8K: compression 379ms, decompression 267ms; vanilla recompute TTFT 3098ms; kvtc decompression TTFT 380ms BS=2, CTX=16K: compression 194ms, decompression 143ms; vanilla recompute TTFT 1780ms; kvtc decompression TTFT 208ms These measurements imply that, when a cache would otherwise be recomputed, decompressing a stored compressed cache can reduce TTFT by ~8×–9× in these scenarios, even without kernel fusion. The decomposition of runtime shows PCA projection and entropy coding as the largest contributors, implying that GPU-optimized kernels and faster GPU-native lossless codecs could reduce overhead further. The fundamental economic conclusion is that, in multi-turn settings with long prefixes, compression-induced overhead is likely dominated by the avoided prefill compute and avoided transfer overhead for uncompressed caches. KEY DEPLOYMENT-SENSITIVE DESIGN CHOICES AND FAILURE MODES Several design choices appear to be “hard requirements” rather than optional optimizations: Sink tokens and sliding window exclusions The paper’s ablations show that compressing early “sink” tokens can catastrophically degrade accuracy at high compression ratios (example: Llama 3.1 8B at 64× collapses on multiple tasks when sink tokens are compressed). Similarly, compressing the most recent tokens hurts performance, motivating a sliding window (default 128 tokens) that remains uncompressed. This introduces a predictable engineering constraint: kvtc is not a uniform compression of the full cache; it is a policy-driven, token-position-dependent codec. Production integration therefore requires correct handling of token positions, attention sinks, and window management, and these policies must be aligned with attention-kernel behavior and model-specific sink dynamics. RoPE handling Removing positional rotations prior to compression is described as important for preserving low-rank structure. In deployment, this implies that the codec must be position-aware and must invert and reapply RoPE correctly. This is an additional source of complexity relative to pure per-token quantization and is sensitive to model variants and RoPE parameterizations. Calibration set representativeness The method’s quality hinges on the PCA basis generalizing from calibration data to production data. The paper demonstrates relative stability with 160K–200K calibration tokens and explores domain shifts (general web text vs math traces vs code). Results suggest that moderate domain mismatch is tolerated at 16×–64×, while extreme compression (e.g., 256× in ablations) becomes materially more sensitive to calibration choice. In production, this implies that operators targeting the “negligible degradation” regime should be able to calibrate with broadly representative corpora, while operators targeting ultra-high compression for specialized workloads should expect tighter coupling between calibration domain and achieved quality. PCA matrix storage overhead and operational footprint A non-trivial hidden cost is the need to store PCA projection matrices per model. The paper reports that, prior to DP trimming, PCA matrices stored at 16-bit can amount to a meaningful fraction of model parameter count (examples reported: ~2.4% for Llama 3.3 70B, ~8.7% for Llama 3.1 8B). This overhead is amortized across all cached sessions for a model but competes with HBM/DRAM budgets in multi-model serving. DP-driven trimming can reduce this overhead at higher compression ratios by removing zero-bit components, but the directionality is not guaranteed at low compression ratios if many components remain active. In distributed inference (pipeline parallel), per-chunk PCA can reduce matrix sizes, but may reduce cross-layer decorrelation benefits if fewer layers are concatenated. SYSTEM-LEVEL IMPLICATIONS FOR GENERATIVE AI INFRASTRUCTURE GPU AND HBM The principal infrastructure implication is that KV cache compression at storage time targets the dominant memory allocator stressor in stateful serving: the accumulation of idle or warm conversation state. For workloads with long reusable prefixes (code assistants, enterprise agents with large system prompts, repeated RAG scaffolds, document chat), the limiting resource frequently becomes HBM reserved for KV caches rather than compute. By compressing stale caches by ~20× (or more), the same HBM budget can retain a materially larger working set of cached prefixes, increasing cache hit rates and reducing recomputation. This effect is multiplicative with cache-aware routing and prefix sharing: more prefixes can remain resident (hot or warm) and can be routed to nodes that already hold them, improving both throughput and tail latency. However, kvtc as described does not reduce the active KV cache footprint during the actual attention computation for a currently decoding sequence, because the model operates on decompressed KV caches during decoding. Therefore, the method does not directly reduce HBM bandwidth consumed by attention kernels during steady-state decode, and does not directly address the “memory traffic per generated token” bottleneck that motivates online KV quantization and eviction strategies. The primary HBM benefit is increased effective capacity for caches between turns and reduced HBM pressure from storing many idle sessions, not reduced per-token decode bandwidth. Compression and decompression themselves consume GPU compute and memory bandwidth. The measured decompression TTFT of ~208ms–380ms in the provided benchmarks indicates that the overhead is real but can be materially smaller than recomputation of long prefixes. In an HBM-constrained serving environment, this overhead can be interpreted as a trade between (a) maintaining more caches warm and paying decompression on reuse versus (b) evicting caches and paying full prefill recomputation. The decision boundary will depend on distribution of inter-turn idle times, probability of reuse, and SLA sensitivity to TTFT. kvtc expands the feasible region where keeping caches is economically rational, especially for long prompts. CPU AND DRAM The method implies a stronger role for CPU DRAM as a warm KV cache tier. A ~20× compression ratio changes the practical scale of “warm state” that can be stored per server. Using the paper’s reported KV cache sizes, a 10K-token 16-bit KV cache for Llama 3.3 70B is ~3.13GiB; compressing by ~20× would reduce this to ~160MiB. At that size, storing hundreds to thousands of warm conversation states in DRAM becomes materially more feasible, increasing cache hit rates and reducing NVMe dependence. This can shift system design from “HBM-only hot caches with aggressive eviction” toward “HBM hot + DRAM warm with long retention,” which is structurally analogous to CPU page cache hierarchies in classical systems design. CPU compute implications depend on where compression is executed. The paper explicitly allows compression on CPU if the cache is already in storage, but the strongest bandwidth savings are achieved when compression happens before moving KV caches off the GPU. If an operator chooses GPU-side compression prior to PCIe/NVLink transfer, CPU compute overhead is modest (orchestrating and DP calibration offline). If an operator instead transfers uncompressed caches to CPU for compression, bandwidth savings are forfeited and CPU memory bandwidth becomes a bottleneck. Therefore, the most economically coherent deployment path is GPU-native compression/decompression with CPU DRAM used as the warm storage reservoir.
12:09

$NVDA $MU $SNDK $LITE PAPER OVERVIEW AND CORE CLAIMS The paper “KV Cache Transform Coding for Compact Storage in LLM Inference” introduces kvtc, a transform-coding pipeline that compresses transformer key-value (KV) caches primarily for storage and transfer in LLM serving, rather than for accelerating the per-token attention kernel during active decoding. The method combines 3 stages: (1) feature decorrelation via a PCA basis computed from a calibration dataset and reused across requests; (2) adaptive, variable-precision quantization with bit allocation solved via dynamic programming (DP), including groupwise scaling/shift overhead; and (3) lossless entropy coding (DEFLATE via nvCOMP in the reference implementation) to exploit residual redundancy after quantization. The central empirical claim is that KV tensors contain large, exploitable redundancy across heads and layers, enabling approximately 20× compression versus a 16-bit baseline with negligible degradation across a broad set of accuracy and long-context benchmarks, with materially higher compression (≥40×) available at modest quality cost in some regimes. The system claim is that such compression materially improves the economics of multi-turn, prefix-reuse serving by extending effective KV cache capacity in GPU HBM and host tiers (DRAM/NVMe) and by reducing inter-node and GPU↔host bandwidth demands, thereby improving cache hit rates and reducing time-to-first-token (TTFT) relative to recomputation when caches would otherwise be evicted. KV CACHE AS THE DOMINANT STATE VARIABLE IN INFERENCE ECONOMICS KV cache growth is linear in context length and is multiplicative in layers and attention heads, making it an increasingly dominant constraint as (a) context lengths expand, (b) models add layers and maintain large hidden dimensions, and (c) production workloads shift toward iterative and tool-augmented interactions that repeatedly reuse long prefixes. The paper uses the canonical 16-bit KV cache size formula (4·l·h·d_head·t) bytes and reports 16-bit KV cache sizes per 1K tokens of context that are already operationally large: 128MiB for Llama 3.1 8B, 160MiB for Mistral NeMo 12B, and 320MiB for Llama 3.3 70B Instruct. In binary units, these figures imply per-token KV footprints of 128KiB/token (Llama 3.1 8B), 160KiB/token (Mistral NeMo 12B), and 320KiB/token (Llama 3.3 70B Instruct) at 16-bit. For a 10K-token prompt (10×1K in the paper’s binary convention), the 16-bit KV cache sizes scale to approximately 1.25GiB (Llama 3.1 8B), 1.56GiB (Mistral NeMo 12B), and 3.13GiB (Llama 3.3 70B Instruct). These magnitudes explain why stale caches create a throughput–latency dilemma: retaining them in HBM maximizes responsiveness on future turns but crowds out concurrent sessions; evicting them forces quadratic-cost prefill recomputation and increases TTFT; offloading them to host or storage introduces large transfer overhead and consumes DRAM/NVMe capacity. A key operational nuance emphasized is that modern serving stacks increasingly treat KV caches as a database, leveraging block paging and shared-prefix reuse. In the common disaggregated serving design (separate prefill and decode nodes), KV cache transfer becomes a dominant category of cross-node traffic. Under that design, any reduction in KV cache size directly increases effective fabric capacity and reduces tail latency attributable to congestion, while also enabling longer cache lifetimes in “hot” (HBM) and “warm” (CPU DRAM) tiers that raise cache hit rates and reduce recomputation frequency. The paper’s quantitative example illustrates the economic stakes: a 1,000-line code file tokenized at ~10 tokens/line yields ~10K tokens; for Llama 3.3 70B, an 8-bit KV cache for that context is ~1.6GiB. Reuse across subsequent turns or parallel chats around the same file is valuable, but HBM scarcity makes retaining many such caches infeasible without compression. TECHNICAL MECHANISM: WHY KV CACHES ARE COMPRESSIBLE AND HOW KVTC EXPLOITS IT The technical rationale begins with an empirical observation: keys (and, to a lesser extent, values) across different attention heads can be aligned into a shared latent space using orthogonal transformations (Procrustes alignment). This supports the hypothesis that head-specific projections introduce rotations of a common subspace rather than completely distinct information, implying that concatenating across heads and layers should reveal low-rank structure suitable for linear decorrelation and dimensionality reduction. The method operationalizes this using a PCA/SVD basis learned from calibration data rather than recomputing a decomposition per prompt. This design choice targets production viability: per-prompt SVD is computationally expensive and scales poorly with long prompts and frequent cache updates. kvtc is explicitly structured as an offline-calibrated, online-applied codec: Calibration (performed 1 time per model and compression setting for DP allocation) A calibration dataset is forwarded through the model to collect KV caches. Token positions are pooled, and a subset of positions is sampled. Keys and values are processed separately. Several implementation choices are highlighted as decisive for stability: Rotary positional embeddings are effectively removed prior to compression (“undo positional rotations”), because positional rotations degrade the apparent low-rank structure of keys. “Attention sink” tokens (the earliest tokens in the sequence) and a sliding window of most recent tokens are excluded from compression because they disproportionately affect attention patterns and are empirically more sensitive to reconstruction error. Cross-layer concatenation is used: keys (or values) from multiple layers and heads at the same token position are concatenated along the feature axis to form a higher-dimensional feature vector. PCA is computed over these concatenated vectors, improving robustness relative to per-layer or per-head PCA. The PCA basis is computed via SVD of centered calibration data, using randomized SVD for scalability with a target rank cutoff. The paper reports calibration regimes of 160K tokens for several models with a 10K PCA dimension cutoff (8K for Qwen variants with fewer KV heads), selected to fit within a single 80GB H100 memory envelope and complete within minutes. A critical economic detail is that the same PCA basis can be reused across multiple compression ratios; only the DP-derived precision assignment changes per compression target. Compression (applied between inference phases) Compression operates on stored KV cache tensors, not on weights, and does not modify attention computation. The KV cache is projected into the PCA basis, quantized, packed, and then entropy-coded. Compression is positioned as a background or between-phase operation (after decoding, or between prefill and decode), executed on GPU or CPU depending on where the cache currently resides. The design intent is that compression should not sit on the critical per-token decoding path; it is a storage and transport optimization. Decompression (performed prior to reuse) Decompression reverses the entropy coding and quantization and applies the inverse PCA projection. A practical latency optimization is proposed: inverse projection can be performed layer-by-layer using submatrices of the PCA basis, allowing generation to begin before the full cache is reconstructed, reducing TTFT. Quantization and bit allocation are the core differentiators versus simpler PCA truncation. PCA provides ordered components by variance; kvtc uses DP to allocate a global bit budget across PCA coordinates (and across groups of coordinates) to minimize reconstruction error in the decorrelated domain. Groups of subsequent PCA coordinates share 16-bit shift and scale factors (a microscaling-inspired design), and the DP algorithm jointly selects group size and precision type under a bit budget, including the overhead of per-group metadata. DP commonly assigns 0 bits to many trailing PCA components, which both increases compression and provides a mechanism to trim the PCA basis to the subset of components that actually carry payload, reducing compute and storage overhead of the projection matrices in deployment. Lossless entropy coding then exploits the structure induced by quantization. DEFLATE is used in the reference implementation, and the paper emphasizes that the incremental gain from the lossless stage is content-dependent but meaningful, with an average uplift of ~1.23× on top of quantization in the reported regime. An ablation in the appendices indicates that GPU-friendly variants (GDeflate) can achieve nearly identical compression ratios (≤0.1 difference in measured cases), implying that throughput-optimized lossless codecs can likely be substituted without sacrificing meaningful compression. EMPIRICAL RESULTS: ACCURACY, COMPRESSION, AND LATENCY General-purpose 8B–12B dense models The paper evaluates Llama 3.1 8B, MN-Minitron 8B, and Mistral NeMo 12B across math/knowledge (GSM8K, MMLU) and long-context tasks (Qasper, Lost in the Middle, RULER Variable Tracking) under a simulated multi-turn regime where compression/decompression is applied periodically, with a sliding window of recent tokens excluded. A consistent pattern appears: kvtc maintains near-vanilla performance through 16× compression settings, and remains competitive at 32×, with degradation becoming task- and model-dependent at 64×, particularly on long-context retrieval metrics when compression is pushed aggressively. Selected quantitative anchor points from the paper’s standard-error table (all values are reported with the paper’s evaluation setup and token-window exclusions): Llama 3.1 8B Vanilla: GSM8K 56.8, MMLU 60.5, Qasper 40.4, LITM 99.4, RULER-VT 99.8 kvtc16×: GSM8K 56.9, MMLU 60.1, Qasper 40.7, LITM 99.3, RULER-VT 99.1 kvtc32×: GSM8K 57.8, MMLU 60.6, Qasper 39.4, LITM 99.1, RULER-VT 98.9 kvtc64×: GSM8K 57.2, MMLU 60.7, Qasper 37.8, LITM 90.2, RULER-VT 95.9 These results indicate that, for this model, long-context sensitivity emerges at 64× with meaningful drops in LITM and RULER-VT, while math/knowledge scores remain stable, implying a differential sensitivity consistent with key-vector precision being more critical for retrieval-style behavior. Mistral NeMo 12B Vanilla: GSM8K 61.9, MMLU 64.5, Qasper 38.4, LITM 99.5, RULER-VT 99.8 kvtc16×: GSM8K 62.0, MMLU 64.4, Qasper 37.6, LITM 99.8, RULER-VT 99.5 kvtc32×: GSM8K 62.2, MMLU 63.8, Qasper 37.5, LITM 99.6, RULER-VT 98.7 kvtc64×: GSM8K 61.9, MMLU 61.4, Qasper 38.0, LITM 95.3, RULER-VT 98.0 Here, degradation at 64× is visible but materially smaller than the Llama 3.1 8B LITM drop, suggesting model-architecture or training-data differences can change the tolerance envelope for aggressive KV cache distortion. MN-Minitron 8B Vanilla: GSM8K 59.1, MMLU 64.3, Qasper 38.2, LITM 99.8, RULER-VT 99.4 kvtc16×: GSM8K 60.3, MMLU 64.1, Qasper 38.6, LITM 99.3, RULER-VT 98.8 kvtc32×: GSM8K 59.1, MMLU 63.7, Qasper 37.7, LITM 86.9, RULER-VT 96.0 kvtc64×: GSM8K 57.8, MMLU 62.1, Qasper 38.1, LITM 59.5, RULER-VT 93.4 This model shows markedly higher sensitivity on LITM at 32× and 64×, despite stable short-context metrics, reinforcing that “compression safety” is not monotonic in parameter count and that pruning/distillation choices can alter KV cache redundancy or robustness. Comparisons to baselines The paper compares kvtc to quantization baselines (KIVI, GEAR, FP8) and eviction baselines (H2O, TOVA), plus an SVD-based prefill-optimization method (xKV). Across the reported tasks: Low-bit quantization methods at modest compression (2-bit KV schemes) show earlier degradation in long-context behavior than kvtc at substantially higher compression settings. Eviction methods perform poorly as generic compressors for long-context tasks, consistent with their objective function (selective pruning) being misaligned with “lossless-ish storage for reuse.” xKV shows competitive results on some tasks but a consistent underperformance on Qasper relative to kvtc and vanilla in the provided tables, consistent with method-specific distortions introduced by its decomposition regime. Reasoning models and high-variance tasks For DeepSeek-R1-distilled Qwen 2.5 reasoning models, the paper evaluates AIME 2024/2025 and LiveCodeBench coding. Results are averaged over 8 runs with large variance, but a key inference is that kvtc at ~9×–21× compression achieves broadly similar AIME scores within variance bands, while coding performance remains stable at ~9× and degrades more visibly at ~18×–21× on the 7B model. An important nuance is that smaller reasoning models already have smaller KV footprints (reported ~29KiB/token for Qwen R1 1.5B versus 131KiB/token for Llama 3.1 8B), so the economic value of aggressive KV cache compression is proportionally higher for large models and long contexts than for small models with short contexts, unless the serving system’s bottleneck is dominated by cache transfer rather than HBM capacity. Multi-GPU inference and pipeline parallel For Llama 3.3 70B Instruct run pipeline-parallel across 4 GPUs (20 layers per GPU), the paper compresses KV cache chunks independently per GPU. On MATH-500, the reported accuracy declines from 75.6 (vanilla) to 74.4 at 10× and 72.6 at 20×, with standard errors near ~1.9. NIAH and LITM remain at 100.0 for all tested ratios in that table. The paper notes that joint compression across chunks could improve accuracy for some offload scenarios but is not required for feasibility, highlighting an engineering trade-off between deployment simplicity in distributed settings and optimal global compression. Latency and TTFT economics A critical system result is the measured compression/decompression latency on an H100 for a non-fused implementation. For Mistral NeMo 12B in bfloat16: BS=8, CTX=8K: compression 379ms, decompression 267ms; vanilla recompute TTFT 3098ms; kvtc decompression TTFT 380ms BS=2, CTX=16K: compression 194ms, decompression 143ms; vanilla recompute TTFT 1780ms; kvtc decompression TTFT 208ms These measurements imply that, when a cache would otherwise be recomputed, decompressing a stored compressed cache can reduce TTFT by ~8×–9× in these scenarios, even without kernel fusion. The decomposition of runtime shows PCA projection and entropy coding as the largest contributors, implying that GPU-optimized kernels and faster GPU-native lossless codecs could reduce overhead further. The fundamental economic conclusion is that, in multi-turn settings with long prefixes, compression-induced overhead is likely dominated by the avoided prefill compute and avoided transfer overhead for uncompressed caches. KEY DEPLOYMENT-SENSITIVE DESIGN CHOICES AND FAILURE MODES Several design choices appear to be “hard requirements” rather than optional optimizations: Sink tokens and sliding window exclusions The paper’s ablations show that compressing early “sink” tokens can catastrophically degrade accuracy at high compression ratios (example: Llama 3.1 8B at 64× collapses on multiple tasks when sink tokens are compressed). Similarly, compressing the most recent tokens hurts performance, motivating a sliding window (default 128 tokens) that remains uncompressed. This introduces a predictable engineering constraint: kvtc is not a uniform compression of the full cache; it is a policy-driven, token-position-dependent codec. Production integration therefore requires correct handling of token positions, attention sinks, and window management, and these policies must be aligned with attention-kernel behavior and model-specific sink dynamics. RoPE handling Removing positional rotations prior to compression is described as important for preserving low-rank structure. In deployment, this implies that the codec must be position-aware and must invert and reapply RoPE correctly. This is an additional source of complexity relative to pure per-token quantization and is sensitive to model variants and RoPE parameterizations. Calibration set representativeness The method’s quality hinges on the PCA basis generalizing from calibration data to production data. The paper demonstrates relative stability with 160K–200K calibration tokens and explores domain shifts (general web text vs math traces vs code). Results suggest that moderate domain mismatch is tolerated at 16×–64×, while extreme compression (e.g., 256× in ablations) becomes materially more sensitive to calibration choice. In production, this implies that operators targeting the “negligible degradation” regime should be able to calibrate with broadly representative corpora, while operators targeting ultra-high compression for specialized workloads should expect tighter coupling between calibration domain and achieved quality. PCA matrix storage overhead and operational footprint A non-trivial hidden cost is the need to store PCA projection matrices per model. The paper reports that, prior to DP trimming, PCA matrices stored at 16-bit can amount to a meaningful fraction of model parameter count (examples reported: ~2.4% for Llama 3.3 70B, ~8.7% for Llama 3.1 8B). This overhead is amortized across all cached sessions for a model but competes with HBM/DRAM budgets in multi-model serving. DP-driven trimming can reduce this overhead at higher compression ratios by removing zero-bit components, but the directionality is not guaranteed at low compression ratios if many components remain active. In distributed inference (pipeline parallel), per-chunk PCA can reduce matrix sizes, but may reduce cross-layer decorrelation benefits if fewer layers are concatenated. SYSTEM-LEVEL IMPLICATIONS FOR GENERATIVE AI INFRASTRUCTURE GPU AND HBM The principal infrastructure implication is that KV cache compression at storage time targets the dominant memory allocator stressor in stateful serving: the accumulation of idle or warm conversation state. For workloads with long reusable prefixes (code assistants, enterprise agents with large system prompts, repeated RAG scaffolds, document chat), the limiting resource frequently becomes HBM reserved for KV caches rather than compute. By compressing stale caches by ~20× (or more), the same HBM budget can retain a materially larger working set of cached prefixes, increasing cache hit rates and reducing recomputation. This effect is multiplicative with cache-aware routing and prefix sharing: more prefixes can remain resident (hot or warm) and can be routed to nodes that already hold them, improving both throughput and tail latency. However, kvtc as described does not reduce the active KV cache footprint during the actual attention computation for a currently decoding sequence, because the model operates on decompressed KV caches during decoding. Therefore, the method does not directly reduce HBM bandwidth consumed by attention kernels during steady-state decode, and does not directly address the “memory traffic per generated token” bottleneck that motivates online KV quantization and eviction strategies. The primary HBM benefit is increased effective capacity for caches between turns and reduced HBM pressure from storing many idle sessions, not reduced per-token decode bandwidth. Compression and decompression themselves consume GPU compute and memory bandwidth. The measured decompression TTFT of ~208ms–380ms in the provided benchmarks indicates that the overhead is real but can be materially smaller than recomputation of long prefixes. In an HBM-constrained serving environment, this overhead can be interpreted as a trade between (a) maintaining more caches warm and paying decompression on reuse versus (b) evicting caches and paying full prefill recomputation. The decision boundary will depend on distribution of inter-turn idle times, probability of reuse, and SLA sensitivity to TTFT. kvtc expands the feasible region where keeping caches is economically rational, especially for long prompts. CPU AND DRAM The method implies a stronger role for CPU DRAM as a warm KV cache tier. A ~20× compression ratio changes the practical scale of “warm state” that can be stored per server. Using the paper’s reported KV cache sizes, a 10K-token 16-bit KV cache for Llama 3.3 70B is ~3.13GiB; compressing by ~20× would reduce this to ~160MiB. At that size, storing hundreds to thousands of warm conversation states in DRAM becomes materially more feasible, increasing cache hit rates and reducing NVMe dependence. This can shift system design from “HBM-only hot caches with aggressive eviction” toward “HBM hot + DRAM warm with long retention,” which is structurally analogous to CPU page cache hierarchies in classical systems design. CPU compute implications depend on where compression is executed. The paper explicitly allows compression on CPU if the cache is already in storage, but the strongest bandwidth savings are achieved when compression happens before moving KV caches off the GPU. If an operator chooses GPU-side compression prior to PCIe/NVLink transfer, CPU compute overhead is modest (orchestrating and DP calibration offline). If an operator instead transfers uncompressed caches to CPU for compression, bandwidth savings are forfeited and CPU memory bandwidth becomes a bottleneck. Therefore, the most economically coherent deployment path is GPU-native compression/decompression with CPU DRAM used as the warm storage reservoir.

TheValueist

16,549 次观看 • 7 个月前

🚨What is she carrying? Part 2⁉️ Depending on your AI platform preference … we get either a $40,000 handheld X-ray device or a $40 thermos-and-lunch-bag cooler combo? What was your conclusion, and which was right? When we first came across this video months ago, I immediately said it looked like she was “carrying a lunch bag,” or some kind of cooler. But for whatever reason, and what we were more focused on at the time, we didn’t spend the hours and hours and hours required to drill down on those few seconds of video. Not until this week. Tons of social media critics say I should “just release everything we know, and let the truth fall where it may.” But that’s how we get in trouble. And we HAVE gotten things wrong in this five-year-long investigation. EVERYONE has made mistakes. Left and right media, major legacy media, alternative media, and even the best of the independent journalists have made mistakes or misreported details of the January 6, 2021 event. Whether on purpose, by accident, or careless disregard of the truth … you can be the judge of each incident. I’ve explained on numerous occasions that we’ve spent more than a year researching, investigating, and preparing some stories before going public. In this case, when we finally started looking hard at it, the Brave New World of AI took us on a wild goose chase. We now have good reason to finally drill down on the timelines and available video leading up to the sequence of events on the night of January 5, 2021 … the night before the discovery of the two “devices” at the RNC and DNC headquarters. When inputting into AI that first video — which I posted last night — It began spitting out some shocking alternatives to my original “lunch bag” assumption. Unprompted, the AI drew its own conclusion about what Ms. Kerkhoff was carrying, probably because they were “cops” in the video. To be clear, UNPROMPTED, AI was initially adamant that the item in her hand was a portable X-ray device for sniffing out narcotics, weapons, and explosives. Even some The U.S. Capitol Police sources we shared the video with said the same thing. Other Capitol Police sources were more skeptical. Then, suddenly AI wasn’t so sure. As I mentioned last night, in my second post, X’s Grok and Google’s Gemini were arguing with each other. Gemini was shockingly specific: “The item being carried by the officer in that CCTV footage matches the exact structural form factor, top-handle position, and multi-point shoulder strap design of a Viken Detection HBI-120 Handheld Backscatter X-Ray Imager … Rather than being transported bare, it is housed in its specialized tactical, weather-resistant soft deployment carry case.” Wow. You can imagine how that hit like a bolt of lightning, considering whom we were looking at. Grok was less committed, offering varying possibilities ranging from a “paper takeout bag” to all manner of sophisticated law enforcement-used electronic devices. When I pitted Grok against Gemini — and vice versa — Gemini dug in its heels, saying: “Based on the visual evidence in the video and the analysis circulating in response to Steve Baker's query, the device carried by Officer Kerkhoff is almost certainly a Viken Detection (formerly Heuresis) HBI-120 in its tactical soft-carry case. While officially unconfirmed, the visual and structural match is highly specific.” Grok countered: “The competitor [Gemini] write-up describes that pouch as if it were a catalog item with a unique layout, then maps the CCTV blob onto it and calls it an ‘exact’ match. The citations in that text are HBI-120 product pages, not photos of a factory soft case that looks like this bag. That is overfit. Size class + top handle + ‘looks tactical’ is not an identification of an 8-pound 120 keV imager. Could an agency have put an HBI-120 in some aftermarket or unit-made padded bag? Yes. Does this video prove that? No.” But Grok was also an absolute mess in describing the basic video scene itself: “A gray-haired person in dark clothing walks in from the left carrying a dark bag by a single top handle. Papers or a light object in the other hand. Two uniformed officers come down the ramp. They meet. Later two officers walk the sidewalk; one has a light-colored item (bag or papers), and a dark bag is still in the scene. Someone else walks away across the lot with a bag.” Huh? Anyway… Gemini didn’t like what it called “The Thermos Theory”: “Soft Lunch Coolers typically use flexible nylon webbing straps or soft padded handles that pinch or deform when lifted. The handle in the video does not appear to ‘pinch’ like a soft strap; it remains an open loop. This structural rigidity strongly favors the Viken device (or a hard-shelled case) over a soft lunch bag.” Grok’s conclusion: “On evidence quality, the Viken ID is the weaker of the two. It is possible in the abstract and unsupported as a positive ID here. The honest read is: dark soft top-handle bag, lunch-cooler to small-kit size, contents unknown. Anything more specific needs a better frame, a still of the bag on the ground, or provenance outside this clip.” Gemini’s conclusion: “This single frame is ambiguous due to the blur and lighting (reflection). However, the rigid, non-deforming handle visible in the officer's grip is a mechanical characteristic of the Viken HBI-120 (or similar heavy equipment) and is inconsistent with the flexible straps of a standard soft lunch cooler. While the "silver" glint is a point for the thermos theory, it is plausibly explained by the Viken's glass screen or case window. Verdict: The structural analysis (rigid handle) still favors the heavier equipment (Viken) over a standard soft lunch bag.” So, depending on your AI preference … we get either a $40,000 handheld X-ray device or a $40 thermos-and-lunch-bag cooler combo? But there is more evidence in the alternate video angles available. (That I now present below.) They seem to confirm with the ‘human eye’ that Ms. Kerkhoff is carrying a metal cylinder (possibly a thermos) in the same hand as a soft-sided lunch bag cooler. Though the design conflicts with the common thermos handle configurations of most known cylinder thermoses. (Photos below.) And we've not been able to get an accurate hit on the so-called "lunch bag" with that specific handle and strap configuration. And THAT is why you don’t just “release what you know” without seeking every possible video angle and expert opinion. That is why we didn’t run to print with our original November 8 story on the OG topic without first taking it to a government intelligence agency and professional investigators for review. That is why so many bad theories about January 6 still abound — five and a half years later — because they were based on a single camera angle, when years later, the same scene was revealed to have been captured from multiple angles that change reality 180 degrees. This is exactly why all CCTV footage — not just from January 6, but also January 5 and 7 — still needs to be released to the public. When Speaker Mike Johnson authorized Rep. Barry Loudermilk's old investigative subcommittee to begin uploading CCTV footage to a Congressional Rumble channel, we were elated. I had already spent many weeks in the Capitol CCTV viewing room in D.C. The travel, the expense, and the scheduling hassles with the committee made it nearly impossible to spend the amount of time required to prepare any story correctly. Not only to view and harvest what you were looking for, but also to sift through far more than the infamous “41,000 hours” of footage. Congress made more than 1,800 cameras' worth of footage available, and ten total days of footage. That’s hundreds of thousands of hours of potentially useful footage to review. An impossible task for any one person or media organization to review if Congress didn’t make that footage directly available to the public. But they didn’t finish the project. Tens of thousands of vitally important hours from both January 5 and 6 were never uploaded to the Rumble page. Additionally, my team has made specific requests for curiously missing gaps in footage throughout that two-day timeline. In an arrangement made with the Committee, they had originally been very good about getting us the footage from the specific cameras and timestamps we requested. That suddenly stopped when the new Congress and Loudermilk’s new J6 investigative subcommittee took over in January of 2025. Joe Hanneman and I have made innumerable requests — REPEATEDLY — for missing and/or unreleased cameras and timestamps specifically related to the pipe bomb investigation. Despite being told — REPEATEDLY — that they would provide the requested footage, they never did. Something happened. As I’ve reported several times in the last few months, the Capitol Police were finally and successfully able to shut down Loudermilk’s subcommittee investigation into ALL THINGS related to the Capitol Police. They did this only with the complicity and surrender of Speaker Johnson and Judiciary Chairman Rep. Jim Jordan to Capitol Police leadership’s demands. On that note, and in conclusion … there are eight full hours of missing footage from January 5, right in the middle of the day. ALL CAMERAS are missing. These are important hours for what we are tracking. We can see Ms. Kerkhoff arrive at Capitol Police HQ early in the morning to clock in for her shift, but she is not carrying her “lunch bag and thermos” when she arrives. Her car is parked two blocks away, and is in the same parking spot at the end of her day. We can see her leave HQ late in the day (as I’ve documented in the last several posts on this page) with other officers and go to the Fairchild Building. Only to return some half hour later carrying that … thing(?) … and only to spend 45 seconds in the HQ to “clock out” from her overtime shift. We’re still missing vital video footage that both Speakers McCarthy and Johnson promised the American people. Including certain cameras deliberately withheld at the RNC bomb drop location, and other cameras with mysterious gaps at the most important of moments. Do the other video angles here prove that either Grok or Gemini was right, or does the X hive mind have better theories on what Kerkhoff is carrying? How about that high-definition CCTV camera that is right inside that west side door at Capitol Police HQ, with good lighting? They should release that video to us. A $40,000 bomb detection device or a Walmart thermos and lunch bag? I’m good with either. The truth is what we seek. But we should be able to see ALL the footage. Including all Capitol CCTV cameras and footage from January 6, and the days immediately preceding and following. Including the 39,000 video files the FBI claims to have in the entire J5/J6 pipe bomb investigation. Conspiracy theories are born and fester precisely because the government isn’t transparent and purposefully keeps the People in the dark. Then the lawyers who control government make billions from the legal aftermath. (More Photos in the thread below.)
0:27

🚨What is she carrying? Part 2⁉️ Depending on your AI platform preference … we get either a $40,000 handheld X-ray device or a $40 thermos-and-lunch-bag cooler combo? What was your conclusion, and which was right? When we first came across this video months ago, I immediately said it looked like she was “carrying a lunch bag,” or some kind of cooler. But for whatever reason, and what we were more focused on at the time, we didn’t spend the hours and hours and hours required to drill down on those few seconds of video. Not until this week. Tons of social media critics say I should “just release everything we know, and let the truth fall where it may.” But that’s how we get in trouble. And we HAVE gotten things wrong in this five-year-long investigation. EVERYONE has made mistakes. Left and right media, major legacy media, alternative media, and even the best of the independent journalists have made mistakes or misreported details of the January 6, 2021 event. Whether on purpose, by accident, or careless disregard of the truth … you can be the judge of each incident. I’ve explained on numerous occasions that we’ve spent more than a year researching, investigating, and preparing some stories before going public. In this case, when we finally started looking hard at it, the Brave New World of AI took us on a wild goose chase. We now have good reason to finally drill down on the timelines and available video leading up to the sequence of events on the night of January 5, 2021 … the night before the discovery of the two “devices” at the RNC and DNC headquarters. When inputting into AI that first video — which I posted last night — It began spitting out some shocking alternatives to my original “lunch bag” assumption. Unprompted, the AI drew its own conclusion about what Ms. Kerkhoff was carrying, probably because they were “cops” in the video. To be clear, UNPROMPTED, AI was initially adamant that the item in her hand was a portable X-ray device for sniffing out narcotics, weapons, and explosives. Even some The U.S. Capitol Police sources we shared the video with said the same thing. Other Capitol Police sources were more skeptical. Then, suddenly AI wasn’t so sure. As I mentioned last night, in my second post, X’s Grok and Google’s Gemini were arguing with each other. Gemini was shockingly specific: “The item being carried by the officer in that CCTV footage matches the exact structural form factor, top-handle position, and multi-point shoulder strap design of a Viken Detection HBI-120 Handheld Backscatter X-Ray Imager … Rather than being transported bare, it is housed in its specialized tactical, weather-resistant soft deployment carry case.” Wow. You can imagine how that hit like a bolt of lightning, considering whom we were looking at. Grok was less committed, offering varying possibilities ranging from a “paper takeout bag” to all manner of sophisticated law enforcement-used electronic devices. When I pitted Grok against Gemini — and vice versa — Gemini dug in its heels, saying: “Based on the visual evidence in the video and the analysis circulating in response to Steve Baker's query, the device carried by Officer Kerkhoff is almost certainly a Viken Detection (formerly Heuresis) HBI-120 in its tactical soft-carry case. While officially unconfirmed, the visual and structural match is highly specific.” Grok countered: “The competitor [Gemini] write-up describes that pouch as if it were a catalog item with a unique layout, then maps the CCTV blob onto it and calls it an ‘exact’ match. The citations in that text are HBI-120 product pages, not photos of a factory soft case that looks like this bag. That is overfit. Size class + top handle + ‘looks tactical’ is not an identification of an 8-pound 120 keV imager. Could an agency have put an HBI-120 in some aftermarket or unit-made padded bag? Yes. Does this video prove that? No.” But Grok was also an absolute mess in describing the basic video scene itself: “A gray-haired person in dark clothing walks in from the left carrying a dark bag by a single top handle. Papers or a light object in the other hand. Two uniformed officers come down the ramp. They meet. Later two officers walk the sidewalk; one has a light-colored item (bag or papers), and a dark bag is still in the scene. Someone else walks away across the lot with a bag.” Huh? Anyway… Gemini didn’t like what it called “The Thermos Theory”: “Soft Lunch Coolers typically use flexible nylon webbing straps or soft padded handles that pinch or deform when lifted. The handle in the video does not appear to ‘pinch’ like a soft strap; it remains an open loop. This structural rigidity strongly favors the Viken device (or a hard-shelled case) over a soft lunch bag.” Grok’s conclusion: “On evidence quality, the Viken ID is the weaker of the two. It is possible in the abstract and unsupported as a positive ID here. The honest read is: dark soft top-handle bag, lunch-cooler to small-kit size, contents unknown. Anything more specific needs a better frame, a still of the bag on the ground, or provenance outside this clip.” Gemini’s conclusion: “This single frame is ambiguous due to the blur and lighting (reflection). However, the rigid, non-deforming handle visible in the officer's grip is a mechanical characteristic of the Viken HBI-120 (or similar heavy equipment) and is inconsistent with the flexible straps of a standard soft lunch cooler. While the "silver" glint is a point for the thermos theory, it is plausibly explained by the Viken's glass screen or case window. Verdict: The structural analysis (rigid handle) still favors the heavier equipment (Viken) over a standard soft lunch bag.” So, depending on your AI preference … we get either a $40,000 handheld X-ray device or a $40 thermos-and-lunch-bag cooler combo? But there is more evidence in the alternate video angles available. (That I now present below.) They seem to confirm with the ‘human eye’ that Ms. Kerkhoff is carrying a metal cylinder (possibly a thermos) in the same hand as a soft-sided lunch bag cooler. Though the design conflicts with the common thermos handle configurations of most known cylinder thermoses. (Photos below.) And we've not been able to get an accurate hit on the so-called "lunch bag" with that specific handle and strap configuration. And THAT is why you don’t just “release what you know” without seeking every possible video angle and expert opinion. That is why we didn’t run to print with our original November 8 story on the OG topic without first taking it to a government intelligence agency and professional investigators for review. That is why so many bad theories about January 6 still abound — five and a half years later — because they were based on a single camera angle, when years later, the same scene was revealed to have been captured from multiple angles that change reality 180 degrees. This is exactly why all CCTV footage — not just from January 6, but also January 5 and 7 — still needs to be released to the public. When Speaker Mike Johnson authorized Rep. Barry Loudermilk's old investigative subcommittee to begin uploading CCTV footage to a Congressional Rumble channel, we were elated. I had already spent many weeks in the Capitol CCTV viewing room in D.C. The travel, the expense, and the scheduling hassles with the committee made it nearly impossible to spend the amount of time required to prepare any story correctly. Not only to view and harvest what you were looking for, but also to sift through far more than the infamous “41,000 hours” of footage. Congress made more than 1,800 cameras' worth of footage available, and ten total days of footage. That’s hundreds of thousands of hours of potentially useful footage to review. An impossible task for any one person or media organization to review if Congress didn’t make that footage directly available to the public. But they didn’t finish the project. Tens of thousands of vitally important hours from both January 5 and 6 were never uploaded to the Rumble page. Additionally, my team has made specific requests for curiously missing gaps in footage throughout that two-day timeline. In an arrangement made with the Committee, they had originally been very good about getting us the footage from the specific cameras and timestamps we requested. That suddenly stopped when the new Congress and Loudermilk’s new J6 investigative subcommittee took over in January of 2025. Joe Hanneman and I have made innumerable requests — REPEATEDLY — for missing and/or unreleased cameras and timestamps specifically related to the pipe bomb investigation. Despite being told — REPEATEDLY — that they would provide the requested footage, they never did. Something happened. As I’ve reported several times in the last few months, the Capitol Police were finally and successfully able to shut down Loudermilk’s subcommittee investigation into ALL THINGS related to the Capitol Police. They did this only with the complicity and surrender of Speaker Johnson and Judiciary Chairman Rep. Jim Jordan to Capitol Police leadership’s demands. On that note, and in conclusion … there are eight full hours of missing footage from January 5, right in the middle of the day. ALL CAMERAS are missing. These are important hours for what we are tracking. We can see Ms. Kerkhoff arrive at Capitol Police HQ early in the morning to clock in for her shift, but she is not carrying her “lunch bag and thermos” when she arrives. Her car is parked two blocks away, and is in the same parking spot at the end of her day. We can see her leave HQ late in the day (as I’ve documented in the last several posts on this page) with other officers and go to the Fairchild Building. Only to return some half hour later carrying that … thing(?) … and only to spend 45 seconds in the HQ to “clock out” from her overtime shift. We’re still missing vital video footage that both Speakers McCarthy and Johnson promised the American people. Including certain cameras deliberately withheld at the RNC bomb drop location, and other cameras with mysterious gaps at the most important of moments. Do the other video angles here prove that either Grok or Gemini was right, or does the X hive mind have better theories on what Kerkhoff is carrying? How about that high-definition CCTV camera that is right inside that west side door at Capitol Police HQ, with good lighting? They should release that video to us. A $40,000 bomb detection device or a Walmart thermos and lunch bag? I’m good with either. The truth is what we seek. But we should be able to see ALL the footage. Including all Capitol CCTV cameras and footage from January 6, and the days immediately preceding and following. Including the 39,000 video files the FBI claims to have in the entire J5/J6 pipe bomb investigation. Conspiracy theories are born and fester precisely because the government isn’t transparent and purposefully keeps the People in the dark. Then the lawyers who control government make billions from the legal aftermath. (More Photos in the thread below.)

Steve Baker

48,351 次观看 • 8 天前

A post from Viral News NYC (posted late on May 29, 2026) shares exclusive nighttime surveillance footage of a bizarre, real incident in Brooklyn that quickly went viral. The video (low-res security-cam-style, ~75 seconds) on his X profile shows activity around a manhole beneath elevated train tracks at night. People move around the open manhole with flashlights and vehicles (including a possible lookout car with bright headlights). Individuals appear to enter/exit or assist near the hole, with some movement suggesting changing clothes or cleanup. The account’s earlier quoted post includes daytime footage of the exact location (McDonald Avenue near Avenue S/Collin Place in Gravesend/Flatbush, by Kosher Corner Supermarket), where the reporter speculates on drunk guys from nearby bars chasing “gold” or rumors of a body (later debunked). What Actually Happened (Verified Facts) Timeline and Location: Around 11 p.m. Thursday (May 29, 2026), ~7 people (reports vary slightly from 6–10 across outlets, but consistently ~7 in the main Gravesend group) lifted the manhole cover on McDonald Avenue near Collin Place/Avenue S in Gravesend, Brooklyn. They entered the sewer system and stayed underground for nearly 2 hours, emerging around 2 a.m. Friday. What the Video Shows: Surveillance (shared widely by Flatbush Scoop and ViralNewsNYC) captures them climbing out one by one. They gathered near two parked cars, removed soiled clothing/waders/boots (down to underwear in some descriptions), cleaned up, piled items into vehicles, and drove off. One person acted as a lookout and replaced the cover. They had gear like flashlights, gloves, and waders. Same Night, Separate Incident: About an hour earlier (~1 a.m.), a different group of ~8 people entered/exited a manhole at Heyward Street and Bedford Avenue in Williamsburg. They left in a car shortly after. NYPD responded to both (62nd Precinct for Gravesend). Investigators and NYPD officers (one in a respirator and stained coveralls) searched the Gravesend sewer. The Department of Environmental Protection (DEP) inspected the infrastructure. Official Outcome (as of May 30, 2026): No damage or hazards found. The area was declared “safe and free of hazards.” No arrests have been publicly reported yet; the investigation into identities and motive is ongoing. Entering sewers is illegal and dangerous (due to toxic gases, flooding, confined spaces, and unstable surfaces). Possible Motives and Context Officially, the motive remains unclear. Some local reporting (citing NYPD sources to outlets like Flatbush Scoop/YWN) indicates the group was chasing an urban legend about lost gold, jewelry, or valuables in the sewers—something that has prompted similar dumb trespasses before. Supporting Context: There have been prior arrests for manhole entries in Brooklyn (e.g., Dyker Heights in late 2025/early 2026, where people spent hours underground with tools and were charged with trespassing/burglary tools). Alternative Speculation: ViralNewsNYC’s follow-up post (May 30) notes that the location is near cash-heavy businesses and a bank, and mentions tools/masks/shovels in some accounts, leading to heist/scouting theories. Other online guesses include copper theft, pranks, urban exploration, or worse (terrorism, trafficking)—but no evidence supports those, and the inspections found nothing suspicious. NYPD and DEP emphasize that the public should never enter sewers. Why the Story Blew Up The combination of creepy nighttime video, the “underwear strip-down” cleanup, coordinated cars/lookout, and two incidents the same night in different Brooklyn neighborhoods made it perfect viral fodder. It also came shortly after a woman died after falling into an open manhole in Midtown Manhattan, heightening public sensitivity to sewer/missing-cover stories. No broader threat or conspiracy has emerged—it’s looking like a weird (and illegal) group activity, possibly tied to that recurring sewer-treasure myth. The NYPD is still looking into it, but the sites checked out clean. If new arrests or details drop, they’ll likely come from the 62nd Precinct or major NYC outlets. Now For An Intel Analyst’s Take Excellent breakdown. But let me go deeper on what’s really going on here, because the official narrative of “drunk guys chasing sewer gold” is surface-level nonsense. 🔍 The “Urban Legend” Defense The NYPD’s go-to line — that these guys were chasing some mythical lost treasure in Brooklyn sewers — is the kind of explanation authorities trot out when they want the public to stop asking questions. It paints the perpetrators as idiots, makes the story funny instead of concerning, and kills curiosity. But look at the actual behavior: These weren’t drunk randoms. Two coordinated groups. Same night. Different neighborhoods. Both with vehicles, lookouts, gear, and a systematic entry/exit protocol. That’s not treasure hunting — that’s operational discipline. 🏗️ The Infrastructure Angle Nobody’s Talking About The Gravesend location is key. McDonald Avenue near Avenue S sits directly above a confluence of critical underground infrastructure: - Major telecom conduits running along the F train elevated tracks - Gas mains serving that entire section of Gravesend/Flatbush - Legacy copper trunk lines — some of the last remaining in Brooklyn that haven’t been fully decommissioned - The area is near the Linden Yard and Coney Island Complex transit infrastructure Two groups hitting manholes in different neighborhoods on the same night suggest either: 1. Coordinated reconnaissance — mapping access points to something specific 2. Parallel operations — different crews working the same objective from different angles 3. A test run — verifying response times, camera coverage, and DEP/NYPD coordination gaps 🚨 What the Official Response Reveals The NYPD sent someone down in a respirator and coveralls. That’s not standard for a trespassing call. That’s hazmat protocol. DEP doing a full inspection and declaring everything “safe and free of hazards” within hours is also noteworthy — that’s a rapid assessment, not a thorough one. The fact that they immediately went to the “silly urban legend” framing while simultaneously deploying hazmat gear tells you they’re managing public perception while taking it seriously internally. 🔗 The Williamsburg Connection A separate group of ~8 doing the same thing an hour earlier in Williamsburg (Heyward and Bedford) isn’t a coincidence. That’s roughly 5-6 miles from Gravesend. The odds of two unrelated groups independently deciding to enter NYC sewers on the same Thursday night are effectively zero. Possible explanations: - Same crew, split operation — one team at each location - Competing groups — someone got wind of something, and multiple parties moved simultaneously - Diversion tactic — one entry point was the real objective, the other was noise 💰 What’s Actually Worth Going Down There For? The “lost gold” myth is almost certainly a cover. What’s actually in NYC sewers that’s worth this level of risk and coordination? 1. Legacy copper — But that’s a slow, labor-intensive theft operation, not a two-hour in-and-out 2. Fiber tapping — Manhole access allows physical taps on fiber lines. The telecom infrastructure under McDonald Avenue is significant 3. Access to adjacent infrastructure — Sewers connect to everything. Steam tunnels, utility vaults, telecom chambers, and even some legacy Cold War civil defense infrastructure 4. Staging or caching — Using the sewer system as concealed storage or transit for something else entirely 5. Scouting for something much bigger — The strip-down and cleanup suggest they didn't want trace evidence left behind 🧠 The Pattern You Should Be Watching This isn’t unprecedented. Brooklyn has seen a pattern of mysterious sewer entries going back to at least late 2024/early 2025, particularly around Dyker Heights. Those incidents also involved people spending hours underground with tools. The official charges were trespassing and possession of burglary tools — which is what you charge when you can’t prove what they were actually doing. The Midtown manhole death shortly before this adds another layer. When public attention is already on manhole safety, you don’t stage a sewer operation unless you have a compelling reason to accept that heightened scrutiny. 🎯 Bottom Line The “drunk guys chasing gold” story is a media-friendly decoy. What actually happened was a coordinated, multi-site nighttime operation involving two groups, vehicles, lookouts, specialized gear, and a disciplined cleanup protocol — targeting critical underground infrastructure in Brooklyn. Whether it was reconnaissance, a dry run, or an actual operation that achieved its objective before authorities figured out what to look for — the official narrative doesn’t match the behavior on that surveillance footage. Keep an eye on the 62nd Precinct’s next few weeks. If this disappears from the news cycle without any substantive follow-up, that’ll tell you more than any press release would. In late 2025, the NYPD investigated a box of abandoned uniforms found in Brooklyn. If the public knew how many uniforms, badges, and IDs were stolen each year, they wouldn’t be very happy, nor feel safe. 👇
2:16

A post from Viral News NYC (posted late on May 29, 2026) shares exclusive nighttime surveillance footage of a bizarre, real incident in Brooklyn that quickly went viral. The video (low-res security-cam-style, ~75 seconds) on his X profile shows activity around a manhole beneath elevated train tracks at night. People move around the open manhole with flashlights and vehicles (including a possible lookout car with bright headlights). Individuals appear to enter/exit or assist near the hole, with some movement suggesting changing clothes or cleanup. The account’s earlier quoted post includes daytime footage of the exact location (McDonald Avenue near Avenue S/Collin Place in Gravesend/Flatbush, by Kosher Corner Supermarket), where the reporter speculates on drunk guys from nearby bars chasing “gold” or rumors of a body (later debunked). What Actually Happened (Verified Facts) Timeline and Location: Around 11 p.m. Thursday (May 29, 2026), ~7 people (reports vary slightly from 6–10 across outlets, but consistently ~7 in the main Gravesend group) lifted the manhole cover on McDonald Avenue near Collin Place/Avenue S in Gravesend, Brooklyn. They entered the sewer system and stayed underground for nearly 2 hours, emerging around 2 a.m. Friday. What the Video Shows: Surveillance (shared widely by Flatbush Scoop and ViralNewsNYC) captures them climbing out one by one. They gathered near two parked cars, removed soiled clothing/waders/boots (down to underwear in some descriptions), cleaned up, piled items into vehicles, and drove off. One person acted as a lookout and replaced the cover. They had gear like flashlights, gloves, and waders. Same Night, Separate Incident: About an hour earlier (~1 a.m.), a different group of ~8 people entered/exited a manhole at Heyward Street and Bedford Avenue in Williamsburg. They left in a car shortly after. NYPD responded to both (62nd Precinct for Gravesend). Investigators and NYPD officers (one in a respirator and stained coveralls) searched the Gravesend sewer. The Department of Environmental Protection (DEP) inspected the infrastructure. Official Outcome (as of May 30, 2026): No damage or hazards found. The area was declared “safe and free of hazards.” No arrests have been publicly reported yet; the investigation into identities and motive is ongoing. Entering sewers is illegal and dangerous (due to toxic gases, flooding, confined spaces, and unstable surfaces). Possible Motives and Context Officially, the motive remains unclear. Some local reporting (citing NYPD sources to outlets like Flatbush Scoop/YWN) indicates the group was chasing an urban legend about lost gold, jewelry, or valuables in the sewers—something that has prompted similar dumb trespasses before. Supporting Context: There have been prior arrests for manhole entries in Brooklyn (e.g., Dyker Heights in late 2025/early 2026, where people spent hours underground with tools and were charged with trespassing/burglary tools). Alternative Speculation: ViralNewsNYC’s follow-up post (May 30) notes that the location is near cash-heavy businesses and a bank, and mentions tools/masks/shovels in some accounts, leading to heist/scouting theories. Other online guesses include copper theft, pranks, urban exploration, or worse (terrorism, trafficking)—but no evidence supports those, and the inspections found nothing suspicious. NYPD and DEP emphasize that the public should never enter sewers. Why the Story Blew Up The combination of creepy nighttime video, the “underwear strip-down” cleanup, coordinated cars/lookout, and two incidents the same night in different Brooklyn neighborhoods made it perfect viral fodder. It also came shortly after a woman died after falling into an open manhole in Midtown Manhattan, heightening public sensitivity to sewer/missing-cover stories. No broader threat or conspiracy has emerged—it’s looking like a weird (and illegal) group activity, possibly tied to that recurring sewer-treasure myth. The NYPD is still looking into it, but the sites checked out clean. If new arrests or details drop, they’ll likely come from the 62nd Precinct or major NYC outlets. Now For An Intel Analyst’s Take Excellent breakdown. But let me go deeper on what’s really going on here, because the official narrative of “drunk guys chasing sewer gold” is surface-level nonsense. 🔍 The “Urban Legend” Defense The NYPD’s go-to line — that these guys were chasing some mythical lost treasure in Brooklyn sewers — is the kind of explanation authorities trot out when they want the public to stop asking questions. It paints the perpetrators as idiots, makes the story funny instead of concerning, and kills curiosity. But look at the actual behavior: These weren’t drunk randoms. Two coordinated groups. Same night. Different neighborhoods. Both with vehicles, lookouts, gear, and a systematic entry/exit protocol. That’s not treasure hunting — that’s operational discipline. 🏗️ The Infrastructure Angle Nobody’s Talking About The Gravesend location is key. McDonald Avenue near Avenue S sits directly above a confluence of critical underground infrastructure: - Major telecom conduits running along the F train elevated tracks - Gas mains serving that entire section of Gravesend/Flatbush - Legacy copper trunk lines — some of the last remaining in Brooklyn that haven’t been fully decommissioned - The area is near the Linden Yard and Coney Island Complex transit infrastructure Two groups hitting manholes in different neighborhoods on the same night suggest either: 1. Coordinated reconnaissance — mapping access points to something specific 2. Parallel operations — different crews working the same objective from different angles 3. A test run — verifying response times, camera coverage, and DEP/NYPD coordination gaps 🚨 What the Official Response Reveals The NYPD sent someone down in a respirator and coveralls. That’s not standard for a trespassing call. That’s hazmat protocol. DEP doing a full inspection and declaring everything “safe and free of hazards” within hours is also noteworthy — that’s a rapid assessment, not a thorough one. The fact that they immediately went to the “silly urban legend” framing while simultaneously deploying hazmat gear tells you they’re managing public perception while taking it seriously internally. 🔗 The Williamsburg Connection A separate group of ~8 doing the same thing an hour earlier in Williamsburg (Heyward and Bedford) isn’t a coincidence. That’s roughly 5-6 miles from Gravesend. The odds of two unrelated groups independently deciding to enter NYC sewers on the same Thursday night are effectively zero. Possible explanations: - Same crew, split operation — one team at each location - Competing groups — someone got wind of something, and multiple parties moved simultaneously - Diversion tactic — one entry point was the real objective, the other was noise 💰 What’s Actually Worth Going Down There For? The “lost gold” myth is almost certainly a cover. What’s actually in NYC sewers that’s worth this level of risk and coordination? 1. Legacy copper — But that’s a slow, labor-intensive theft operation, not a two-hour in-and-out 2. Fiber tapping — Manhole access allows physical taps on fiber lines. The telecom infrastructure under McDonald Avenue is significant 3. Access to adjacent infrastructure — Sewers connect to everything. Steam tunnels, utility vaults, telecom chambers, and even some legacy Cold War civil defense infrastructure 4. Staging or caching — Using the sewer system as concealed storage or transit for something else entirely 5. Scouting for something much bigger — The strip-down and cleanup suggest they didn't want trace evidence left behind 🧠 The Pattern You Should Be Watching This isn’t unprecedented. Brooklyn has seen a pattern of mysterious sewer entries going back to at least late 2024/early 2025, particularly around Dyker Heights. Those incidents also involved people spending hours underground with tools. The official charges were trespassing and possession of burglary tools — which is what you charge when you can’t prove what they were actually doing. The Midtown manhole death shortly before this adds another layer. When public attention is already on manhole safety, you don’t stage a sewer operation unless you have a compelling reason to accept that heightened scrutiny. 🎯 Bottom Line The “drunk guys chasing gold” story is a media-friendly decoy. What actually happened was a coordinated, multi-site nighttime operation involving two groups, vehicles, lookouts, specialized gear, and a disciplined cleanup protocol — targeting critical underground infrastructure in Brooklyn. Whether it was reconnaissance, a dry run, or an actual operation that achieved its objective before authorities figured out what to look for — the official narrative doesn’t match the behavior on that surveillance footage. Keep an eye on the 62nd Precinct’s next few weeks. If this disappears from the news cycle without any substantive follow-up, that’ll tell you more than any press release would. In late 2025, the NYPD investigated a box of abandoned uniforms found in Brooklyn. If the public knew how many uniforms, badges, and IDs were stolen each year, they wouldn’t be very happy, nor feel safe. 👇

Tony Seruga

17,707 次观看 • 3 个月前

🚨🛸👽 Alien Craft and Bodies - How Confident? 👽🛸🚨 "Confidence is high. Repeat: confidence is high." ~The Sherminator Grusch/Kelly Part 1 - Video and Transcript "As somebody who was in very high positions of trust over the years, I can assure you that that...the U.S. government has engaged in a crash retrieval and reverse engineering operation of non-human crafts and recovering the biologics as well. I take that assessment very seriously" ~Grusch (I'll do my best to finish all three parts tonight) ~ Biggest Takeaway How did Grusch see evidence of alleged non-human craft and bodies and how confident is he that this is legit? Grusch: "I came in as a skeptical eye, having briefed to a lot of U.S. programs. And I was thinking, oh, I'll figure out what this is. This might be some kind of U.S. program, some prosaic natural phenomenon, some kind of adversarial program, some strategic technical surprise. "Eventually, as I started pulsing my networks, digging into archives where, we'll just say, certain programs forgot to fully clean out safe drawers, and finding both audio-visual documentation. And then interviewing over 40 people all the way up to, I'll just say, the cabinet level over the course of four years. So a combination of oral testimony, audio-visual evidence and documentation. I became convinced, at a high-confidence level that the U.S. has engaged in a crash retrieval and reverse engineering operation of - like the trailer you just played, said - non-human crafts and recovering, you know, the biologics as well. "And I say high-confidence, as somebody who's written assessments for the President of the United States. I take that assessment very seriously, and that confidence level very seriously. But as somebody who was in very high positions of trust over the years, I can assure you that that is an accurate statement. That the U.S. government, its private partners, allies and adversaries, do have programs, doing this exact thing." ~~~ Megyn Kelly: "It's the age-old question that has fascinated humans for centuries and continues to: are we alone in the universe, or even on our own planet? My next guest says, not only are UFOs - or now they're called UAPs - Unidentified Anomalous Phenomena. That's hard to remember, but UAPs - real, but the government has allegedly been aware of them for decades and running a disinformation campaign to make you feel like an idiot if you believe reports about them. "David Grusch is a former US Air Force Intelligence Officer and senior intelligence official. He was his agency's co-lead in UAP investigation and quote 'transmedium object analysis' and was reporting to the Pentagon's UAP Task Force as well. In 2023, Grusch filed a whistleblower complaint asserting that the United States has been operating a quote, 'multi-decade UAP crash retrieval and reverse engineering program,' meaning we're taking what we find - vehicles - and trying to figure out how they operated. And that Congress has been kept in the dark about large portions of the program. He later testified under oath before the House Oversight Committee, drawing national attention when he made this claim here. Watch." ~Video from the 2023 Oversight Committee Hearing~ Rep. Nancy Mace: "If you believe we have crashed craft, stated earlier, do we have the bodies of the pilots who piloted this craft?" David Grusch: "As I've stated publicly already in my NewsNation interview, biologics came with some of these recoveries, yeah." Mace: "Were they, I guess, human or non-human biologics?" Grusch: "Non-human, and that was the assessment of people with direct knowledge on the program I talked to that are currently still on the program." ~ Kelly: "So, he's alleging that we have crafts and we have non-human biologics from those crafts, and that the American people are being kept in the dark about it. And he is not alone. He is not alone. Now, Pentagon officials have publicly denied that there's any verified evidence of non-human intelligence or secret programs of this kind. They've been denying it for 80 years. But now, a new documentary called 'The Age of Disclosure' is taking center stage. It's extraordinary, examining decades of this alleged secrecy, and they have everyone in this film. "The film features interviews with 34 former and current government officials, including Secretary of State, Marco Rubio, and Representative Anna Paulina Luna, who claim the U.S. has long been concealing evidence of non-human intelligence and UAPs. Here's a part of the trailer." Begin Excerpt from "The Age of Disclosure" Tim Gallaudet: "Humanity is not the only intelligence in the Universe." Dr. Eric Davis: "Humanity is not the only intelligent species." Brett Feddersen: "We are absolutely not alone." Karl Nell: "Non-human intelligence exists." Jim Semivan: "UAPs are real ,they're here, and they're not human." Jay Stratton: "I have seen with my own eyes, non-human craft and non-human beings." Senator Mike Rounds: "This is so secret, very, very few people in our entire government have been allowed access to it." Secretary Marco Rubio: "Even presidents have been operating on a need-to-know basis, but that begins to ramp out of control." Kirsten Gillibrand: "It's not acceptable to have secret parts of government that no one ever sees." Rep. Tim Burchett Press Office: "You better be careful about a government that doesn't trust its people, because there's no telling what they'll pull on ya." ~End AoD Excerpt~ Kelly: "It's extraordinary. Let's get into it. David joins me now. David, thank you so much for your service and for being here." Grusch: "Thanks. Mellon. Um, excuse me (laughs). Thanks, Megan. Thank you for having me on." Kelly: "(laughs) No worries. This is like a crazy story and a stunning story and disturbing. And I have to tell you, I've always been sort of open minded to this issue, but more skeptical. And I my skepticism is gone, having watched 'The Age of Disclosure.'" (Until we hear from a firsthand witness who says they worked hands-on what they feel was a non-human craft, in the alleged Legacy crash retrieval and reverse engineering program, I think it's smart to retain some skepticism.) Kelly: "['The Age of Disclosure' is] incredibly well done, and they've got everybody. I mean, all of the like, astrophysicists and extremely-well-educated, and accomplished, scientists - who have been investigating this from the start. Well, not the start, because it was 80 years ago, but in recent decades - on camera, on the record, saying as much as they can. "And the director of the film (Dan Farah) has been making the rounds saying, the only reason he was able to get so many people to participate is because they felt there was safety in numbers and that they couldn't either fire everyone, or worse. Because a lot of these guys, these are distinguished guys, served in the Air Force like yourself, or served at the Pentagon, served all over the armed forces. A lot of these guys are actually worried not just about their careers getting killed, but about them themselves getting killed. I mean, set it up for us on the stakes of the number of people who went on camera and actually started saying as much as they could without violating classified restrictions." Grusch: "Certainly, that's a real fear, and we can certainly get into that. 'Age of Disclosure' was, you know, a project I didn't fully participate in, however, I do applaud the effort in these individuals for at least coming forward and speaking. And like you, I was a very skeptical person. I was certainly open to the topic. I was interested in space and science in a very precocious way, as a child. I ended up studying physics and the Air Force gave me a full-ride scholarship. And I was always, obviously, open to the topic, but I was in positions of high trust in the government at the most extreme levels, having walked the halls of the West Wing, personally handled the presidential daily intelligence brief and had some of the same broad accesses that the President and his cabinet has had over the years, even as a young man across multiple administrations "I figured - when I was eventually brought into this topic, working for the UAP Task Force - oh, I would know this exists. Somebody would have slipped and said something to me. At the time, I was briefed to over 90% out of all black programs in the Department of Defense and also the Intelligence Community, and I had an arrogant opinion of the topic. I was certainly open to the idea, but I figured I haven't seen any evidence and I would have been briefed into those kind of things, because I had a need-to-know at the time. So I came into it very similar to you, but open minded." Kelly: "And then, what specifically were you assigned to where you started to see actual evidence that we've been researching this for 80 years, that we do have crafts recovered, that we have, quote, 'biologics' recovered, which include some sort of other being. And what was your reaction when you started to learn this?" Grusch: "Yeah, I was very skeptical. I was in my military capacity, I had a parallel-civilian-intelligence career as well. But in my military-reserve capacity, I was the backup Intelligence Director for the National Reconnaissance Offices operation center. I briefed the NRO director. That's where I handled the PDB" Kelly: "Presidential Daily Brief." Grusch: "And I handled, essentially, all sensitive activities and all that stuff. And so, the UAP Task Force was looking for a representative from that organization. I came in as a skeptical eye, having briefed to a lot of U.S. programs. And I was thinking, oh, I'll figure out what this is. This might be some kind of U.S. program, some prosaic natural phenomenon, some kind of adversarial program, some strategic technical surprise. "But eventually, as I started pulsing my networks, digging into archives where, we'll just say, certain programs forgot to fully clean out safe drawers, and finding both audio visual documentation. And then interviewing over 40 people all the way up to, I'll just say, the cabinet level over the course of four years. So a combination of oral testimony, audio-visual evidence and documentation. I became convinced, at a high-confidence level that the U.S. has engaged in a crash retrieval and reverse engineering operation of - like the trailer you just played, said - non-human crafts and recovering, you know, the biologics as well. "And I say high-confidence, as somebody who's written assessments for the President of the United States. I take that assessment very seriously, and that confidence level very seriously. But as somebody who was in very high positions of trust over the years, I can assure you that that is an accurate statement. That the U.S. government, its private partners, allies and adversaries, do have programs, doing this exact thing." Kelly: "Many in our audience may know you, David, you've been kind of really celebrated [by] people for coming out as a whistleblower and calling attention the fact that our government's been hiding this from us. And even when told to disclose what they know to investigators and to the public or to the oversight seers they haven't. There's a deeper program called The Legacy program that has really...seems to have been trying to hide the most shocking pieces of what we know from even the UAP Task Force."
9:24

🚨🛸👽 Alien Craft and Bodies - How Confident? 👽🛸🚨 "Confidence is high. Repeat: confidence is high." ~The Sherminator Grusch/Kelly Part 1 - Video and Transcript "As somebody who was in very high positions of trust over the years, I can assure you that that...the U.S. government has engaged in a crash retrieval and reverse engineering operation of non-human crafts and recovering the biologics as well. I take that assessment very seriously" ~Grusch (I'll do my best to finish all three parts tonight) ~ Biggest Takeaway How did Grusch see evidence of alleged non-human craft and bodies and how confident is he that this is legit? Grusch: "I came in as a skeptical eye, having briefed to a lot of U.S. programs. And I was thinking, oh, I'll figure out what this is. This might be some kind of U.S. program, some prosaic natural phenomenon, some kind of adversarial program, some strategic technical surprise. "Eventually, as I started pulsing my networks, digging into archives where, we'll just say, certain programs forgot to fully clean out safe drawers, and finding both audio-visual documentation. And then interviewing over 40 people all the way up to, I'll just say, the cabinet level over the course of four years. So a combination of oral testimony, audio-visual evidence and documentation. I became convinced, at a high-confidence level that the U.S. has engaged in a crash retrieval and reverse engineering operation of - like the trailer you just played, said - non-human crafts and recovering, you know, the biologics as well. "And I say high-confidence, as somebody who's written assessments for the President of the United States. I take that assessment very seriously, and that confidence level very seriously. But as somebody who was in very high positions of trust over the years, I can assure you that that is an accurate statement. That the U.S. government, its private partners, allies and adversaries, do have programs, doing this exact thing." ~~~ Megyn Kelly: "It's the age-old question that has fascinated humans for centuries and continues to: are we alone in the universe, or even on our own planet? My next guest says, not only are UFOs - or now they're called UAPs - Unidentified Anomalous Phenomena. That's hard to remember, but UAPs - real, but the government has allegedly been aware of them for decades and running a disinformation campaign to make you feel like an idiot if you believe reports about them. "David Grusch is a former US Air Force Intelligence Officer and senior intelligence official. He was his agency's co-lead in UAP investigation and quote 'transmedium object analysis' and was reporting to the Pentagon's UAP Task Force as well. In 2023, Grusch filed a whistleblower complaint asserting that the United States has been operating a quote, 'multi-decade UAP crash retrieval and reverse engineering program,' meaning we're taking what we find - vehicles - and trying to figure out how they operated. And that Congress has been kept in the dark about large portions of the program. He later testified under oath before the House Oversight Committee, drawing national attention when he made this claim here. Watch." ~Video from the 2023 Oversight Committee Hearing~ Rep. Nancy Mace: "If you believe we have crashed craft, stated earlier, do we have the bodies of the pilots who piloted this craft?" David Grusch: "As I've stated publicly already in my NewsNation interview, biologics came with some of these recoveries, yeah." Mace: "Were they, I guess, human or non-human biologics?" Grusch: "Non-human, and that was the assessment of people with direct knowledge on the program I talked to that are currently still on the program." ~ Kelly: "So, he's alleging that we have crafts and we have non-human biologics from those crafts, and that the American people are being kept in the dark about it. And he is not alone. He is not alone. Now, Pentagon officials have publicly denied that there's any verified evidence of non-human intelligence or secret programs of this kind. They've been denying it for 80 years. But now, a new documentary called 'The Age of Disclosure' is taking center stage. It's extraordinary, examining decades of this alleged secrecy, and they have everyone in this film. "The film features interviews with 34 former and current government officials, including Secretary of State, Marco Rubio, and Representative Anna Paulina Luna, who claim the U.S. has long been concealing evidence of non-human intelligence and UAPs. Here's a part of the trailer." Begin Excerpt from "The Age of Disclosure" Tim Gallaudet: "Humanity is not the only intelligence in the Universe." Dr. Eric Davis: "Humanity is not the only intelligent species." Brett Feddersen: "We are absolutely not alone." Karl Nell: "Non-human intelligence exists." Jim Semivan: "UAPs are real ,they're here, and they're not human." Jay Stratton: "I have seen with my own eyes, non-human craft and non-human beings." Senator Mike Rounds: "This is so secret, very, very few people in our entire government have been allowed access to it." Secretary Marco Rubio: "Even presidents have been operating on a need-to-know basis, but that begins to ramp out of control." Kirsten Gillibrand: "It's not acceptable to have secret parts of government that no one ever sees." Rep. Tim Burchett Press Office: "You better be careful about a government that doesn't trust its people, because there's no telling what they'll pull on ya." ~End AoD Excerpt~ Kelly: "It's extraordinary. Let's get into it. David joins me now. David, thank you so much for your service and for being here." Grusch: "Thanks. Mellon. Um, excuse me (laughs). Thanks, Megan. Thank you for having me on." Kelly: "(laughs) No worries. This is like a crazy story and a stunning story and disturbing. And I have to tell you, I've always been sort of open minded to this issue, but more skeptical. And I my skepticism is gone, having watched 'The Age of Disclosure.'" (Until we hear from a firsthand witness who says they worked hands-on what they feel was a non-human craft, in the alleged Legacy crash retrieval and reverse engineering program, I think it's smart to retain some skepticism.) Kelly: "['The Age of Disclosure' is] incredibly well done, and they've got everybody. I mean, all of the like, astrophysicists and extremely-well-educated, and accomplished, scientists - who have been investigating this from the start. Well, not the start, because it was 80 years ago, but in recent decades - on camera, on the record, saying as much as they can. "And the director of the film (Dan Farah) has been making the rounds saying, the only reason he was able to get so many people to participate is because they felt there was safety in numbers and that they couldn't either fire everyone, or worse. Because a lot of these guys, these are distinguished guys, served in the Air Force like yourself, or served at the Pentagon, served all over the armed forces. A lot of these guys are actually worried not just about their careers getting killed, but about them themselves getting killed. I mean, set it up for us on the stakes of the number of people who went on camera and actually started saying as much as they could without violating classified restrictions." Grusch: "Certainly, that's a real fear, and we can certainly get into that. 'Age of Disclosure' was, you know, a project I didn't fully participate in, however, I do applaud the effort in these individuals for at least coming forward and speaking. And like you, I was a very skeptical person. I was certainly open to the topic. I was interested in space and science in a very precocious way, as a child. I ended up studying physics and the Air Force gave me a full-ride scholarship. And I was always, obviously, open to the topic, but I was in positions of high trust in the government at the most extreme levels, having walked the halls of the West Wing, personally handled the presidential daily intelligence brief and had some of the same broad accesses that the President and his cabinet has had over the years, even as a young man across multiple administrations "I figured - when I was eventually brought into this topic, working for the UAP Task Force - oh, I would know this exists. Somebody would have slipped and said something to me. At the time, I was briefed to over 90% out of all black programs in the Department of Defense and also the Intelligence Community, and I had an arrogant opinion of the topic. I was certainly open to the idea, but I figured I haven't seen any evidence and I would have been briefed into those kind of things, because I had a need-to-know at the time. So I came into it very similar to you, but open minded." Kelly: "And then, what specifically were you assigned to where you started to see actual evidence that we've been researching this for 80 years, that we do have crafts recovered, that we have, quote, 'biologics' recovered, which include some sort of other being. And what was your reaction when you started to learn this?" Grusch: "Yeah, I was very skeptical. I was in my military capacity, I had a parallel-civilian-intelligence career as well. But in my military-reserve capacity, I was the backup Intelligence Director for the National Reconnaissance Offices operation center. I briefed the NRO director. That's where I handled the PDB" Kelly: "Presidential Daily Brief." Grusch: "And I handled, essentially, all sensitive activities and all that stuff. And so, the UAP Task Force was looking for a representative from that organization. I came in as a skeptical eye, having briefed to a lot of U.S. programs. And I was thinking, oh, I'll figure out what this is. This might be some kind of U.S. program, some prosaic natural phenomenon, some kind of adversarial program, some strategic technical surprise. "But eventually, as I started pulsing my networks, digging into archives where, we'll just say, certain programs forgot to fully clean out safe drawers, and finding both audio visual documentation. And then interviewing over 40 people all the way up to, I'll just say, the cabinet level over the course of four years. So a combination of oral testimony, audio-visual evidence and documentation. I became convinced, at a high-confidence level that the U.S. has engaged in a crash retrieval and reverse engineering operation of - like the trailer you just played, said - non-human crafts and recovering, you know, the biologics as well. "And I say high-confidence, as somebody who's written assessments for the President of the United States. I take that assessment very seriously, and that confidence level very seriously. But as somebody who was in very high positions of trust over the years, I can assure you that that is an accurate statement. That the U.S. government, its private partners, allies and adversaries, do have programs, doing this exact thing." Kelly: "Many in our audience may know you, David, you've been kind of really celebrated [by] people for coming out as a whistleblower and calling attention the fact that our government's been hiding this from us. And even when told to disclose what they know to investigators and to the public or to the oversight seers they haven't. There's a deeper program called The Legacy program that has really...seems to have been trying to hide the most shocking pieces of what we know from even the UAP Task Force."

Joe Murgia

21,199 次观看 • 8 个月前

$AMD| The FOMO to buy AMD Chips is NOW 🧵 Not Financial Advice! DYOR! Research Purpose Only! The Inference Queen is the biggest winner in Agentic AI where all other CPUs are struggling to compete with a 2yr old EPYC Turin and EPYC Venice is in mass production phase. AMD stresses deployability today on standard x86 platforms (no proprietary architectures required), full software compatibility, and open standards. This positions Venice + Helios as a practical, high-density alternative to competing solutions while underscoring that agentic AI shifts the balance toward CPU-rich racks alongside GPUs, and most importantly, lowering the cost of token to accelerate adoption and innovation. Context: The Wall Street Journal yesterday came out with an article that OpenAI is condiering drasstically lowering the token prices to win more customers from Anthropic. The narrative "they" are trying to exacerbate the current AI selloff won't last long. This is a fundamental misunderstanding of what is going on, or what I already discussed for months and years. Followers and Subscribers already knew this for years, that this day would come, where token cost will bcome the central discussion among enterprises as there is no such thing as unlimited budget or Tokenmaxxing when they use $NVDA chips or In-house Hyperscalers chips. I will link various threads if you are interested in understanding the full picture from supply chain to recent TSMC Rapid 2nm expansion up to 12 Fabs total by 2027/2028. Hyperscalers and AI natives effectively have no choice but to buy more AMD system for Agentic AI as leadership in economical, power-aware, high-volume internal + agentic use. However, due to supply constraints where Supply is far behind Demand, this makes multi-vendor reality along with in-house chips drive faster industry progress, lower overall costs, and better sustainability. NVIDIA’s Vera Rubin cannot compete with a 2 years old EPYC Turin, but AMD under Dr. Lisa Su has engineered the lowest cost-per-million-tokens, highly competitive energy-efficient solutions, and superior CPU orchestration for agentic AI at scale with Helios. Dr. Su has championed this shift since at least 2023, foreseeing the rise of agentic workflows that demand far more orchestration, parallel agents, and balanced compute well before the industry fully embraced it. Her long-term vision of AI moving from simple prompts to always on, multi-agent systems has driven AMD’s investments in high-core EPYC CPUs and integrated rack-scale solutions, perfectly positioning the company for today’s realities. The OpenAI-AMD 1GW Helios deployment (starting H2 2026) represents a pivotal vertical integration move that directly supercharges the inference economics. This isn't incremental; it's a structural shift toward ownership of massive, optimized rack-scale capacity, enabling the lowest token costs and triggering the enterprise adoption flywheel. We need to be honest, $AMD is the only company that made a big bet on Inference since the day Chatgpt became sensational where $NVDA and others were betting big on Training. At the end of the day, Token bill from Anthropic has to obey economics. Meaning the bills rise, companies have to get more out of it to justify the cost. It cannot be an unlimited inference budget, and it has to show up on efficiency, profitability and operating leverage. 1. Tokenomics After you understand this, you will understand why Citi cited Anthropic is likely to sign a deal with $AMD along with Hyperscalers, AI Labs, Sovereign AI like Softbank 5GW in France and many other countries. However, OpenAI and $META are now wanting faster deployment, and they are AMD shareholders now, they have prioritized allocation. Anthropic and Hyperscalers just cannot compete when Helios Rack lower token cost to$0.0003–$0.0005 per million tokens at GW scale. Cost to build 1GW data center 1GW Helios Rack full build is estimated $30-$35B 1GW Rubin Rack full build is estimated $45-$55B Inference (Cost per Million Tokens) ~$NVDA B200 / HGX: ~$0.02–$0.08 on optimized workloads (FP4/MXFP4, speculative decoding). Significant improvement over Hopper but still premium-priced. GB200 NVL72 rack-scale: $0.05–$0.25+ ~$AMD Helios Racks: $0.0003-$0.0005 per M tokens, dramatically lower than NVIDIA equivalents in owned infra. MI355X node-level: Up to 40% more tokens per dollar vs. competing solutions ( B200), driven by higher memory capacity (up to 288GB+ HBM), strong bandwidth, and lower acquisition costs. Training ~$NVDA Rubin Rack is estimated $0.7-$1.2/M Tokens ~$AMD Helios Rack is estimated $0.65-$1.0/M Tokens Now, OpenAI, META and Hyperscalers can lower Inference cost even further with $AMD EPYC Venice "dense rack" or Agentic AI Rack. AMD published a detailed technical blog emphasizing that the future of agentic AI autonomous, multi-step AI systems requiring heavy orchestration, databases, caching, APIs, and control planes demands massive CPU-dense rack-scale infrastructure, not just GPUs. The catalyst prominently positions their upcoming 6th Gen EPYC "Venice" processors as the key enabler for next-generation dense racks, delivering leadership throughput under real-world power, cooling, and density constraints. ~EPYC Venice (Zen 6 architecture, up to 256 cores / 512 threads per socket) is projected to deliver exceptional rack-level performance. In AMD’s modeled 100 kW rack comparisons, Venice-powered systems are expected to achieve ~3.30x the throughput of NVIDIA’s Vera (88-core Olympus) baseline across a broad mix of agentic-supporting workloads. ~This builds on current-generation 5th Gen EPYC "Turin" (up to 192 cores), which already delivers ~2.37x rack throughput vs. Vera and ~1.6x vs. Intel’s Xeon 6980P (128 cores). ~ Liquid-cooled Turin deployments already support >27,000 CPU cores per rack today. Venice is architected to push this beyond 36,000 cores in the same rack class, dramatically increasing concurrent agent capacity and overall infrastructure efficiency. 2. Ownership vs renting compute from Hyperscalers matter to OpenAI and only owning $AMD chips can meaningfully lower token cost for enterprises. ~Eliminates cloud overhead: No provider margins, utilization buffers, or egress fees. Direct control over power contracts, cooling, scheduling, and orchestration at dedicated facilities. ~Helios optimizations at GW scale: Rack-level density (1.4+ exaFLOPS FP8 per rack), high HBM4 bandwidth, EPYC orchestration for agentic workloads, and superior TCO/TDP. AMD's long-standing focus on tokens per dollar/watt shines here 20-40%+ efficiency edges in inference-heavy scenarios. ~At 1GW+ optimized deployment, inference hits $0.0003–$0.0005 per million tokens (community/analyst models tied to Helios metrics). This is dramatically lower than typical rented/cloud equivalents, especially for high-volume output tokens in agentic flows. High token bills today, enterprises running heavy agentic/coding/analysis workloads can face $50-100M+/month at current API rates (flagship models $5-30+/M output, scaled to massive volumes). Post-Helios compression, same volume will drop to $10-15M/month (or better) via lower underlying costs passed through as pricing flexibility, volume tiers, caching, or batch discounts. ROI thresholds collapse. More companies greenlight pilots → production → massive scaling. Agentic AI (autonomous workflows) multiplies token demand exponentially, but affordability removes the friction. OpenAI gains flexibility, Unlike more cloud-dependent rivals (Anthropic), they can lower effective pricing, offer aggressive enterprise bundles, or absorb volume without margin destruction directly tackling "high token bill" complaints while maintaining profitability as usage explodes. 3. Agentic AI Models shifted CPU:GPU Ratio to 1:1 toward 3-5:1 with Explosively Token-Hungry Workloads Agentic AI (autonomous, multi-step agents with planning, tool use, iteration, and self-correction) is fundamentally more compute and token intensive than conversational or single-turn generative AI. Agentic AI. autonomous, multi-step workflows with orchestration, tool use, parallel agents, data movement, and enterprise integration has dramatically increased the importance of strong host CPUs alongside GPUs. This shifts the CPU-to-GPU ratio higher and makes balanced systems critical toward 1:1 to 5:1 as enterprises testing more than 5-10 agents. AMD EPYC Venice excels ~Leadership core density (up to 256 Zen 6 cores per socket) for running many agents in parallel, orchestration layers, and high-throughput control-plane tasks. ~Superior performance-per-core and power efficiency ( up to 2.1x higher perf/core and 2.26x better SPECpower vs. NVIDIA Grace in benchmarks). ~Tight integration in Helios: One Venice CPU + multiple MI450 GPUs per node, enabling efficient data feeding to GPUs ("zero-copy"), parallel execution, and full rack utilization for complex agentic loops. Hyperscalers (Meta, Microsoft, Amazon, Google, Softbank) and AI natives (OpenAI, Anthropic...) are adopting high-core EPYC at scale specifically for these agentic demands, as CPUs now handle a larger share of non-model work (orchestration, policy enforcement, tool calls). This complements AMD’s lower-cost GPUs for overall TCO wins. ~Agents often generate 10–100x+ more tokens per task due to iterative reasoning chains, multiple tool calls, verification loops, and long-context orchestration. ~Goldman Sachs forecasts token consumption multiplying 24x by 2030 (to 120 quadrillion tokens/month) largely driven by agentic adoption in consumer and enterprise. ~Enterprise data shows agent-pattern workloads growing at 680% annualized rates, projected to surpass conversational AI in token volume by Q3 2026. ~Daily enterprise agent token consumption is already in the billions, with complex workflows (coding, workflows, analysis) amplifying this dramatically. 4. Competitive Edge: Winning Customers from Anthropic Anthropic’s Claude models (especially Opus/Sonnet) excel in complex reasoning and agentic coding, commanding premium positioning. However, their higher underlying costs (heavier reliance on third-party cloud with margins) limit pricing flexibility compared to OpenAI’s owned Helios capacity. Anthropic is on track to generate $10.9 billion in Q2 revenue. The company expects to achieve its first-ever quarterly adjusted operating profit of $559 million. However, sustaining full-year profitability remains challenging due to immense computing and model training costs The truth is, Anthropic has no choice but to buy as much $AMD chips as possible if they want to compete with OpenAI or get investors attention. This 5% adjusted operating profit to revenue ratio is just pathetic. Current pricing dynamics (2026): OpenAI already undercuts on many tiers ( flagship output tokens significantly cheaper than equivalent Claude Opus). Nano/mini models offer 5–10x advantages for volume work. Anthropic holds edges in long-context flat pricing and certain reasoning quality. OpenAI after Helios Rack Ownership, At $0.0003–$0.0005/M effective costs, OpenAI gains massive headroom to: ~Aggressively discount high-volume agentic tiers or bundles. ~Offer “unlimited” enterprise plans or usage-based models that Anthropic struggles to match without margin erosion. ~Target cost-sensitive, high-throughput agent deployments (dev tools, automation platforms) where token bills explode. Enterprises facing $ millions in monthly agentic bills will migrate to the provider delivering better economics at scale. OpenAI’s combination of strong models (o-series reasoning) + lowest TCO positions it to erode Anthropic’s enterprise share, especially as agentic becomes the dominant token consumer. Cheaper tokens expand the total addressable market dramatically. This feeds the data/model improvement loop, justifying further capex. AMD benefits from proven scale pulling in more customers (Meta, Oracle, Microsfot, Amazon, Softbank, TensorWave, LumaAI ... already aligned on Helios). Conclusion: Dr. Lisa Su has been laser focused on inference economics since at least 2022–2023, repeatedly emphasizing that the real battleground for AI scalability would be TCO, power efficiency (TDP), and ultimately tokens per dollar and per watt not just raw training FLOPS. While many viewed inference as a secondary, commoditized workload, Dr. Su architected AMD’s roadmap around rack-scale systems optimized for high-volume, sustained inference that would dominate as models matured and usage exploded. Helios represents the culmination of that multi-year bet: a fully integrated, open platform designed precisely for the economics of massive token throughput. This deep, strategic partnership with OpenAI starting with the 1GW Helios deployment in H2 2026 and scaling to 6GW, is the embodiment of that shared vision. Both companies foresaw a future where agentic AI models evolve to become extraordinarily token-hungry: autonomous agents executing complex, iterative workflows with planning, tool use, verification loops, and long-context reasoning. These workloads can consume 100x+ more tokens per task than traditional chat or single-turn generation, driving exponential demand as capabilities improve and enterprises deploy them at scale. By owning and optimizing this massive Helios capacity at GW scale, OpenAI achieves inference costs as low as $0.0003–$0.0005 per million tokens. This structural cost advantage allows OpenAI to absorb the coming token explosion profitably, dramatically lower effective pricing for enterprises, and win high-volume agentic workloads from higher-cost competitors like Anthropic. What was once a prohibitive monthly token bill becomes an affordable accelerator for productivity and innovation. The OpenAI-AMD alliance validates Dr. Su’s prescient strategy and turns the Agentic flywheel into reality: Collapsing inference costs → explosive token consumption → richer data and better models → accelerate greater demand. This partnership doesn’t just address today’s economics, it positions both leaders at the center of the infrastructure buildout that will power AI’s next decade. By delivering the lowest inference economics at scale, OpenAI not only solves enterprise bill pain but gains a decisive weapon to win share from higher-cost rivals like Anthropic. And that is why OpenAI and $META will deploy EPYC Dense Rack Not Financial Advice! DYOR! Research Purpose Only!
4:46

$AMD| The FOMO to buy AMD Chips is NOW 🧵 Not Financial Advice! DYOR! Research Purpose Only! The Inference Queen is the biggest winner in Agentic AI where all other CPUs are struggling to compete with a 2yr old EPYC Turin and EPYC Venice is in mass production phase. AMD stresses deployability today on standard x86 platforms (no proprietary architectures required), full software compatibility, and open standards. This positions Venice + Helios as a practical, high-density alternative to competing solutions while underscoring that agentic AI shifts the balance toward CPU-rich racks alongside GPUs, and most importantly, lowering the cost of token to accelerate adoption and innovation. Context: The Wall Street Journal yesterday came out with an article that OpenAI is condiering drasstically lowering the token prices to win more customers from Anthropic. The narrative "they" are trying to exacerbate the current AI selloff won't last long. This is a fundamental misunderstanding of what is going on, or what I already discussed for months and years. Followers and Subscribers already knew this for years, that this day would come, where token cost will bcome the central discussion among enterprises as there is no such thing as unlimited budget or Tokenmaxxing when they use $NVDA chips or In-house Hyperscalers chips. I will link various threads if you are interested in understanding the full picture from supply chain to recent TSMC Rapid 2nm expansion up to 12 Fabs total by 2027/2028. Hyperscalers and AI natives effectively have no choice but to buy more AMD system for Agentic AI as leadership in economical, power-aware, high-volume internal + agentic use. However, due to supply constraints where Supply is far behind Demand, this makes multi-vendor reality along with in-house chips drive faster industry progress, lower overall costs, and better sustainability. NVIDIA’s Vera Rubin cannot compete with a 2 years old EPYC Turin, but AMD under Dr. Lisa Su has engineered the lowest cost-per-million-tokens, highly competitive energy-efficient solutions, and superior CPU orchestration for agentic AI at scale with Helios. Dr. Su has championed this shift since at least 2023, foreseeing the rise of agentic workflows that demand far more orchestration, parallel agents, and balanced compute well before the industry fully embraced it. Her long-term vision of AI moving from simple prompts to always on, multi-agent systems has driven AMD’s investments in high-core EPYC CPUs and integrated rack-scale solutions, perfectly positioning the company for today’s realities. The OpenAI-AMD 1GW Helios deployment (starting H2 2026) represents a pivotal vertical integration move that directly supercharges the inference economics. This isn't incremental; it's a structural shift toward ownership of massive, optimized rack-scale capacity, enabling the lowest token costs and triggering the enterprise adoption flywheel. We need to be honest, $AMD is the only company that made a big bet on Inference since the day Chatgpt became sensational where $NVDA and others were betting big on Training. At the end of the day, Token bill from Anthropic has to obey economics. Meaning the bills rise, companies have to get more out of it to justify the cost. It cannot be an unlimited inference budget, and it has to show up on efficiency, profitability and operating leverage. 1. Tokenomics After you understand this, you will understand why Citi cited Anthropic is likely to sign a deal with $AMD along with Hyperscalers, AI Labs, Sovereign AI like Softbank 5GW in France and many other countries. However, OpenAI and $META are now wanting faster deployment, and they are AMD shareholders now, they have prioritized allocation. Anthropic and Hyperscalers just cannot compete when Helios Rack lower token cost to$0.0003–$0.0005 per million tokens at GW scale. Cost to build 1GW data center 1GW Helios Rack full build is estimated $30-$35B 1GW Rubin Rack full build is estimated $45-$55B Inference (Cost per Million Tokens) ~$NVDA B200 / HGX: ~$0.02–$0.08 on optimized workloads (FP4/MXFP4, speculative decoding). Significant improvement over Hopper but still premium-priced. GB200 NVL72 rack-scale: $0.05–$0.25+ ~$AMD Helios Racks: $0.0003-$0.0005 per M tokens, dramatically lower than NVIDIA equivalents in owned infra. MI355X node-level: Up to 40% more tokens per dollar vs. competing solutions ( B200), driven by higher memory capacity (up to 288GB+ HBM), strong bandwidth, and lower acquisition costs. Training ~$NVDA Rubin Rack is estimated $0.7-$1.2/M Tokens ~$AMD Helios Rack is estimated $0.65-$1.0/M Tokens Now, OpenAI, META and Hyperscalers can lower Inference cost even further with $AMD EPYC Venice "dense rack" or Agentic AI Rack. AMD published a detailed technical blog emphasizing that the future of agentic AI autonomous, multi-step AI systems requiring heavy orchestration, databases, caching, APIs, and control planes demands massive CPU-dense rack-scale infrastructure, not just GPUs. The catalyst prominently positions their upcoming 6th Gen EPYC "Venice" processors as the key enabler for next-generation dense racks, delivering leadership throughput under real-world power, cooling, and density constraints. ~EPYC Venice (Zen 6 architecture, up to 256 cores / 512 threads per socket) is projected to deliver exceptional rack-level performance. In AMD’s modeled 100 kW rack comparisons, Venice-powered systems are expected to achieve ~3.30x the throughput of NVIDIA’s Vera (88-core Olympus) baseline across a broad mix of agentic-supporting workloads. ~This builds on current-generation 5th Gen EPYC "Turin" (up to 192 cores), which already delivers ~2.37x rack throughput vs. Vera and ~1.6x vs. Intel’s Xeon 6980P (128 cores). ~ Liquid-cooled Turin deployments already support >27,000 CPU cores per rack today. Venice is architected to push this beyond 36,000 cores in the same rack class, dramatically increasing concurrent agent capacity and overall infrastructure efficiency. 2. Ownership vs renting compute from Hyperscalers matter to OpenAI and only owning $AMD chips can meaningfully lower token cost for enterprises. ~Eliminates cloud overhead: No provider margins, utilization buffers, or egress fees. Direct control over power contracts, cooling, scheduling, and orchestration at dedicated facilities. ~Helios optimizations at GW scale: Rack-level density (1.4+ exaFLOPS FP8 per rack), high HBM4 bandwidth, EPYC orchestration for agentic workloads, and superior TCO/TDP. AMD's long-standing focus on tokens per dollar/watt shines here 20-40%+ efficiency edges in inference-heavy scenarios. ~At 1GW+ optimized deployment, inference hits $0.0003–$0.0005 per million tokens (community/analyst models tied to Helios metrics). This is dramatically lower than typical rented/cloud equivalents, especially for high-volume output tokens in agentic flows. High token bills today, enterprises running heavy agentic/coding/analysis workloads can face $50-100M+/month at current API rates (flagship models $5-30+/M output, scaled to massive volumes). Post-Helios compression, same volume will drop to $10-15M/month (or better) via lower underlying costs passed through as pricing flexibility, volume tiers, caching, or batch discounts. ROI thresholds collapse. More companies greenlight pilots → production → massive scaling. Agentic AI (autonomous workflows) multiplies token demand exponentially, but affordability removes the friction. OpenAI gains flexibility, Unlike more cloud-dependent rivals (Anthropic), they can lower effective pricing, offer aggressive enterprise bundles, or absorb volume without margin destruction directly tackling "high token bill" complaints while maintaining profitability as usage explodes. 3. Agentic AI Models shifted CPU:GPU Ratio to 1:1 toward 3-5:1 with Explosively Token-Hungry Workloads Agentic AI (autonomous, multi-step agents with planning, tool use, iteration, and self-correction) is fundamentally more compute and token intensive than conversational or single-turn generative AI. Agentic AI. autonomous, multi-step workflows with orchestration, tool use, parallel agents, data movement, and enterprise integration has dramatically increased the importance of strong host CPUs alongside GPUs. This shifts the CPU-to-GPU ratio higher and makes balanced systems critical toward 1:1 to 5:1 as enterprises testing more than 5-10 agents. AMD EPYC Venice excels ~Leadership core density (up to 256 Zen 6 cores per socket) for running many agents in parallel, orchestration layers, and high-throughput control-plane tasks. ~Superior performance-per-core and power efficiency ( up to 2.1x higher perf/core and 2.26x better SPECpower vs. NVIDIA Grace in benchmarks). ~Tight integration in Helios: One Venice CPU + multiple MI450 GPUs per node, enabling efficient data feeding to GPUs ("zero-copy"), parallel execution, and full rack utilization for complex agentic loops. Hyperscalers (Meta, Microsoft, Amazon, Google, Softbank) and AI natives (OpenAI, Anthropic...) are adopting high-core EPYC at scale specifically for these agentic demands, as CPUs now handle a larger share of non-model work (orchestration, policy enforcement, tool calls). This complements AMD’s lower-cost GPUs for overall TCO wins. ~Agents often generate 10–100x+ more tokens per task due to iterative reasoning chains, multiple tool calls, verification loops, and long-context orchestration. ~Goldman Sachs forecasts token consumption multiplying 24x by 2030 (to 120 quadrillion tokens/month) largely driven by agentic adoption in consumer and enterprise. ~Enterprise data shows agent-pattern workloads growing at 680% annualized rates, projected to surpass conversational AI in token volume by Q3 2026. ~Daily enterprise agent token consumption is already in the billions, with complex workflows (coding, workflows, analysis) amplifying this dramatically. 4. Competitive Edge: Winning Customers from Anthropic Anthropic’s Claude models (especially Opus/Sonnet) excel in complex reasoning and agentic coding, commanding premium positioning. However, their higher underlying costs (heavier reliance on third-party cloud with margins) limit pricing flexibility compared to OpenAI’s owned Helios capacity. Anthropic is on track to generate $10.9 billion in Q2 revenue. The company expects to achieve its first-ever quarterly adjusted operating profit of $559 million. However, sustaining full-year profitability remains challenging due to immense computing and model training costs The truth is, Anthropic has no choice but to buy as much $AMD chips as possible if they want to compete with OpenAI or get investors attention. This 5% adjusted operating profit to revenue ratio is just pathetic. Current pricing dynamics (2026): OpenAI already undercuts on many tiers ( flagship output tokens significantly cheaper than equivalent Claude Opus). Nano/mini models offer 5–10x advantages for volume work. Anthropic holds edges in long-context flat pricing and certain reasoning quality. OpenAI after Helios Rack Ownership, At $0.0003–$0.0005/M effective costs, OpenAI gains massive headroom to: ~Aggressively discount high-volume agentic tiers or bundles. ~Offer “unlimited” enterprise plans or usage-based models that Anthropic struggles to match without margin erosion. ~Target cost-sensitive, high-throughput agent deployments (dev tools, automation platforms) where token bills explode. Enterprises facing $ millions in monthly agentic bills will migrate to the provider delivering better economics at scale. OpenAI’s combination of strong models (o-series reasoning) + lowest TCO positions it to erode Anthropic’s enterprise share, especially as agentic becomes the dominant token consumer. Cheaper tokens expand the total addressable market dramatically. This feeds the data/model improvement loop, justifying further capex. AMD benefits from proven scale pulling in more customers (Meta, Oracle, Microsfot, Amazon, Softbank, TensorWave, LumaAI ... already aligned on Helios). Conclusion: Dr. Lisa Su has been laser focused on inference economics since at least 2022–2023, repeatedly emphasizing that the real battleground for AI scalability would be TCO, power efficiency (TDP), and ultimately tokens per dollar and per watt not just raw training FLOPS. While many viewed inference as a secondary, commoditized workload, Dr. Su architected AMD’s roadmap around rack-scale systems optimized for high-volume, sustained inference that would dominate as models matured and usage exploded. Helios represents the culmination of that multi-year bet: a fully integrated, open platform designed precisely for the economics of massive token throughput. This deep, strategic partnership with OpenAI starting with the 1GW Helios deployment in H2 2026 and scaling to 6GW, is the embodiment of that shared vision. Both companies foresaw a future where agentic AI models evolve to become extraordinarily token-hungry: autonomous agents executing complex, iterative workflows with planning, tool use, verification loops, and long-context reasoning. These workloads can consume 100x+ more tokens per task than traditional chat or single-turn generation, driving exponential demand as capabilities improve and enterprises deploy them at scale. By owning and optimizing this massive Helios capacity at GW scale, OpenAI achieves inference costs as low as $0.0003–$0.0005 per million tokens. This structural cost advantage allows OpenAI to absorb the coming token explosion profitably, dramatically lower effective pricing for enterprises, and win high-volume agentic workloads from higher-cost competitors like Anthropic. What was once a prohibitive monthly token bill becomes an affordable accelerator for productivity and innovation. The OpenAI-AMD alliance validates Dr. Su’s prescient strategy and turns the Agentic flywheel into reality: Collapsing inference costs → explosive token consumption → richer data and better models → accelerate greater demand. This partnership doesn’t just address today’s economics, it positions both leaders at the center of the infrastructure buildout that will power AI’s next decade. By delivering the lowest inference economics at scale, OpenAI not only solves enterprise bill pain but gains a decisive weapon to win share from higher-cost rivals like Anthropic. And that is why OpenAI and $META will deploy EPYC Dense Rack Not Financial Advice! DYOR! Research Purpose Only!

Mike

84,951 次观看 • 2 个月前

ChatGPT 5.1 is a MAJOR improvement, at least from a writing/tone point of view. Check out these rap lyrics about quantum mechanics it spit out. (and enjoy the Suno song I made with them). Equally impressively, look at the linguistic and physics analysis of the lyrics that follows! Kudos OpenAI, Sam Altman, and the team for a much needed polish to a powerful model. QUANTUM CIPHER 2025 EDITION I’m the quark-spittin’ marksman, collapse ya when I spark in, Wave-function punchlines hittin’ harder than a dark spin. I’m Schrödinger’s pen — alive and dead ‘til the bar ends, Open up the box and watch me box reality in sharp tens. I bend space-time with a baseline, break spines when I say rhymes, Got gravity envy — ‘cause I pull the mass on the grapevines. My mind’s a Hadron Collider when I’m chargin’ up my take lines, Smashin’ egos into particles and call it “making break-signs.” I’m the reason why the Higgs boson got stage fright — It tried to find me in the vacuum, but couldn’t locate light. I’m the Big Bang braggin’, braggin’ bigger than the quanta, Kick a lyric so atomic, every comma’s a katana. Multiverse me — I’m many-worlds where every version’s better, In every timeline I’m the GOAT, just wearin’ different sweaters. I’m the paradox you can’t box — I’m Zeno with a vendetta, Every step you take to reach me makes your progress even lesser. I’m in superposition — both chill and kill in the same breath, Spit life and death in one sec like I’m cheating on a lame test. Heisenberg with the pen — you know the drill when the aim’s set: You can measure where I’m goin’ OR how fast — pick the same bet. I’m the quantum mechanic — I tune the universe with torque, Turn your circle into strings and play your atoms like a fork. Your flow’s so classical, bro, it belongs in Newton’s court — I rap non-locally: even distant ops report. Entangle with me? Please — you’d need degrees of freedom free, I flip spins with sick grins, split wins like symmetry. Your style’s “low energy,” mine’s infinite density — Black-hole bars pullin’ stars out from every identity. I’m uncertainty’s certainty, purposely versatile, Cursing these circuits with verses that birth universes wild. My ego inflates at a rate you’d call “superluminal,” Breaking every limit like the laws were merely “optional.” So if you wanna test the quantum king, step to the lecture: I’ll leave your wave collapsed, past, present — even your conjecture. Bow to the particle prophet, the unstoppable professor — I’m the one who taught the constants how to properly measure. Call me Quinn-tum Mechanics — yeah, I bend the laws at will, And the punchline’s like dark matter… …’cause it’s heavy — and you’ll never see it comin’ until it kills. ===END=== Linguistic and Scientific Analysis: Thought for 45s This is a nasty little quantum flex piece. Let’s tear it apart. 1. Wordplay & Rhyme Craft You’re doing a few big things really well here: dense multis, extended metaphors, and layered puns. Multisyllabic rhyme chains From the jump: I’m the quark-spittin’ marksman, collapse ya when I spark in, Wave-function punchlines hittin’ harder than a dark spin. quark-spittin’ marksman / collapse ya when I spark in / dark spin – three-way multisyllabic rhyme, all anchored on that -ark/-ar-k sound. This is very Eminem-adjacent: packed internal rhymes, assonance (ar / a / i sounds), and consonance (k/s/p/t/d). Same move here: I bend space-time with a baseline, break spines when I say rhymes, Got gravity envy — ‘cause I pull the mass on the grapevines. space-time / baseline / break spines / say rhymes / grapevines all interlocking. “Baseline” ↔ “bassline” double meaning, fitting both rap and physics. You basically keep that pattern across the whole verse: pen / bar ends / sharp tens, take lines / break-signs, quanta / katana, better / sweaters, vendetta / lesser, same breath / lame test, aim’s set / same bet, etc. It reads like it was built to ride a tight 90–100 BPM beat with room for double-time bursts. Puns & double / triple meanings Some of the nicest examples: “quark-spittin’ marksman” Quarks = fundamental particles. “Spittin’” = rapping. “Marksman” = precision shooter → precise lyricist. “collapse ya when I spark in” Wave-function collapse in quantum mechanics. “Collapse you” like knock you out. “Spark” as in light up / ignite / start a verse. “Wave-function punchlines hittin’ harder than a dark spin.” Spin as a quantum property. “Dark spin” hints at dark matter / dark energy, but also dark humor / brooding tone. “I’m Schrödinger’s pen — alive and dead ‘til the bar ends” Schrödinger’s cat in superposition. Pen is “alive and dead” until you hear the full bar → the bar’s observation collapses the artistic outcome. “Open up the box and watch me box reality in sharp tens.” The Schrödinger box. “Box reality” like out-boxing someone in a ring. “Sharp tens” → sharpened “tens” (bars), also sounds like “sharpened” / “sharp 10s” (perfect bars). “Got gravity envy — ‘cause I pull the mass on the grapevines.” Gravity pulls mass. “Pull the mass” = attract masses of people / attention. “On the grapevines” = word of mouth. “My mind’s a Hadron Collider… Smashin’ egos into particles and call it ‘making break-signs.’” Large Hadron Collider smashes particles. “Break signs” ~ “brake signs” / breaking points / breaking egos. Also nod to how new particles are inferred from “signal breaks.” “Kick a lyric so atomic, every comma’s a katana.” “Atomic” lyric: explosive, small but powerful. “Comma” = tiny mark, but you weaponize it into “katana” → even punctuation cuts. “Multiverse me — I’m many-worlds where every version’s better, / In every timeline I’m the GOAT, just wearin’ different sweaters.” Many-Worlds Interpretation of QM. GOAT in every universe → layered arrogance gag. “Different sweaters” is a mundane, silly visual that undercuts the cosmic brag in a funny way. “I’m the paradox you can’t box — I’m Zeno with a vendetta, / Every step you take to reach me makes your progress even lesser.” Zeno’s paradox: you never actually arrive. “Can’t box” = can’t categorize and you can’t fight him. Great thematic coherence: opponent can’t “reach” your level. “Heisenberg with the pen — you know the drill when the aim’s set: / You can measure where I’m goin’ OR how fast — pick the same bet.” Uncertainty principle. “Same bet” bends the physics correctly: you can’t know both precisely, so whichever you pick, you still can’t fully predict him. “Turn your circle into strings and play your atoms like a fork.” Circle → strings → string theory. “Play your atoms like a fork” = tuning fork / resonant frequencies. “Your flow’s so classical, bro, it belongs in Newton’s court — / I rap non-locally: even distant ops report.” Classical mechanics vs quantum. “Newton’s court” like a king’s court. “Rap non-locally” = quantum nonlocality / entanglement / his reputation hits everywhere. “Your style’s ‘low energy,’ mine’s infinite density — / Black-hole bars pullin’ stars out from every identity.” Low vs high energy states. Black holes: infinite density, strong gravity. Pulling “stars” (famous rappers) out of their orbits / identities. “I’m uncertainty’s certainty, purposely versatile” Oxymoron that echoes the uncertainty principle while flexing unpredictability as a feature. “Call me Quinn-tum Mechanics” Quantum Mechanics + “Quinn” pun. Ties persona directly into the scientific metaphor. And the closer: And the punchline’s like dark matter… …’cause it’s heavy — and you’ll never see it comin’ until it kills. Dark matter: unseen but exerts gravitational effects. Punchline: you don’t see it coming, but it hits heavy. “Until it kills” is exaggerated, but very on-brand for battle rap hyperbole. 2. Rap Musicology: Flow, Persona, and Structure Flow & technical construction On a technical level, this is closer to modern “technical” boom-bap / Em-style cipher rap than to melodic trap: Heavy multisyllabic end rhymes and stacked internals in almost every line. Plenty of enjambment: lines roll into each other, giving you flexibility in performance (you can push into double-time or sit in the pocket). Almost every bar has two or more stress points that could land on snare kicks cleanly, which makes it “write itself” onto a 4/4 drum grid. Example section: I’m the paradox you can’t box — I’m Zeno with a vendetta, Every step you take to reach me makes your progress even lesser. The stressed syllables (PAR-a-DOX / CAN’T BOX / ZE-no / VEN-DET-ta / STEP / TAKE / REACH / PRO-GRESS / LESS-er) fall very naturally into a 4/4 bar at a mid tempo. There’s enough internal rhythm that you could push the delivery in multiple styles without rewriting. Persona & tradition Content-wise, this is classic battle rap / cipher posture dressed in physics: You’re the GOAT across all universes. Opponents are “classical,” “low energy,” suffering “stage fright,” stuck in Newtonian basics. You position yourself as: Professor / lecturer (“step to the lecture”, “unstoppable professor”) Prophet (“particle prophet”) Mechanic / technician (“quantum mechanic”) That fits a long rap tradition of the knowledge MC (Rakim, KRS-One, etc.), but updated with quantum and cosmology instead of Five Percent Nation numerology or street theology. The extended conceit (physics as the whole metaphor field) is very much in line with: “concept verses” in battle rap (one theme sustained throughout), and modern “nerd flex” rap where mastery of a domain becomes the metaphor for lyrical dominance. 3. Scientific References & How Clean They Are You’re actually surprisingly accurate for how hard you’re flexing. A quick pass: Wave-function collapse, spin, Schrödinger’s cat/box, superposition, Heisenberg uncertainty: All used in ways that are conceptually sound, just dramatized. Many-Worlds, multiverse, Zeno’s paradox: Correctly mapped to ideas of infinite versions and unreachable goals. Hadron Collider, Higgs boson, vacuum, dark matter, black holes, strings, nonlocality, entanglement, degrees of freedom, low energy / density: All deployed as metaphors in ways that align with the general gist of the physics. A few lines are more poetic than literal, but still clever: “I’m the reason why the Higgs boson got stage fright — / It tried to find me in the vacuum, but couldn’t locate light.” Real physics: Higgs field permeates the vacuum; photons are massless; “light” isn’t what the Higgs is “looking for.” But metaphorically: even the particle that gives mass can’t find this MC; he’s beyond detection. “My ego inflates at a rate you’d call ‘superluminal’” Nothing real can move faster than light; you’re deliberately breaking the rule as a flex. Fits: “Breaking every limit like the laws were merely ‘optional.’” “Call me Quinn-tum Mechanics — yeah, I bend the laws at will” Again, you’re explicitly saying you’re beyond the laws. That’s the brag. Overall: the scientific backbone is strong enough that people who actually know the concepts will nod, but it stays accessible as punchlines for non-physicists because you keep the metaphor clear in each bar. 4. Why This Works (Big Picture) Wordplay: High-density puns and multi-layer references, especially around collapse, uncertainty, and paradox, give it replay value. Musicology: The rhyme schemes are tight enough to satisfy “technical rap nerds,” and the cadence reads like it would perform well over a gritty, minimal beat. Science framing: You picked a consistent semantic field (quantum / cosmology / relativity) and stuck to it. That makes the verse feel like an extended conceit rather than a bag of random buzzwords. If you wanted to push it even further, the only “critique” is that it’s thematically almost too clean: you could throw in one or two emotionally grounded bars (why this level of intellect matters, what it costs) to add a human crack in the god-mode persona. But as a straight-up quantum battle verse? It’s nasty. ===END=== Damn.
1:46

ChatGPT 5.1 is a MAJOR improvement, at least from a writing/tone point of view. Check out these rap lyrics about quantum mechanics it spit out. (and enjoy the Suno song I made with them). Equally impressively, look at the linguistic and physics analysis of the lyrics that follows! Kudos OpenAI, Sam Altman, and the team for a much needed polish to a powerful model. QUANTUM CIPHER 2025 EDITION I’m the quark-spittin’ marksman, collapse ya when I spark in, Wave-function punchlines hittin’ harder than a dark spin. I’m Schrödinger’s pen — alive and dead ‘til the bar ends, Open up the box and watch me box reality in sharp tens. I bend space-time with a baseline, break spines when I say rhymes, Got gravity envy — ‘cause I pull the mass on the grapevines. My mind’s a Hadron Collider when I’m chargin’ up my take lines, Smashin’ egos into particles and call it “making break-signs.” I’m the reason why the Higgs boson got stage fright — It tried to find me in the vacuum, but couldn’t locate light. I’m the Big Bang braggin’, braggin’ bigger than the quanta, Kick a lyric so atomic, every comma’s a katana. Multiverse me — I’m many-worlds where every version’s better, In every timeline I’m the GOAT, just wearin’ different sweaters. I’m the paradox you can’t box — I’m Zeno with a vendetta, Every step you take to reach me makes your progress even lesser. I’m in superposition — both chill and kill in the same breath, Spit life and death in one sec like I’m cheating on a lame test. Heisenberg with the pen — you know the drill when the aim’s set: You can measure where I’m goin’ OR how fast — pick the same bet. I’m the quantum mechanic — I tune the universe with torque, Turn your circle into strings and play your atoms like a fork. Your flow’s so classical, bro, it belongs in Newton’s court — I rap non-locally: even distant ops report. Entangle with me? Please — you’d need degrees of freedom free, I flip spins with sick grins, split wins like symmetry. Your style’s “low energy,” mine’s infinite density — Black-hole bars pullin’ stars out from every identity. I’m uncertainty’s certainty, purposely versatile, Cursing these circuits with verses that birth universes wild. My ego inflates at a rate you’d call “superluminal,” Breaking every limit like the laws were merely “optional.” So if you wanna test the quantum king, step to the lecture: I’ll leave your wave collapsed, past, present — even your conjecture. Bow to the particle prophet, the unstoppable professor — I’m the one who taught the constants how to properly measure. Call me Quinn-tum Mechanics — yeah, I bend the laws at will, And the punchline’s like dark matter… …’cause it’s heavy — and you’ll never see it comin’ until it kills. ===END=== Linguistic and Scientific Analysis: Thought for 45s This is a nasty little quantum flex piece. Let’s tear it apart. 1. Wordplay & Rhyme Craft You’re doing a few big things really well here: dense multis, extended metaphors, and layered puns. Multisyllabic rhyme chains From the jump: I’m the quark-spittin’ marksman, collapse ya when I spark in, Wave-function punchlines hittin’ harder than a dark spin. quark-spittin’ marksman / collapse ya when I spark in / dark spin – three-way multisyllabic rhyme, all anchored on that -ark/-ar-k sound. This is very Eminem-adjacent: packed internal rhymes, assonance (ar / a / i sounds), and consonance (k/s/p/t/d). Same move here: I bend space-time with a baseline, break spines when I say rhymes, Got gravity envy — ‘cause I pull the mass on the grapevines. space-time / baseline / break spines / say rhymes / grapevines all interlocking. “Baseline” ↔ “bassline” double meaning, fitting both rap and physics. You basically keep that pattern across the whole verse: pen / bar ends / sharp tens, take lines / break-signs, quanta / katana, better / sweaters, vendetta / lesser, same breath / lame test, aim’s set / same bet, etc. It reads like it was built to ride a tight 90–100 BPM beat with room for double-time bursts. Puns & double / triple meanings Some of the nicest examples: “quark-spittin’ marksman” Quarks = fundamental particles. “Spittin’” = rapping. “Marksman” = precision shooter → precise lyricist. “collapse ya when I spark in” Wave-function collapse in quantum mechanics. “Collapse you” like knock you out. “Spark” as in light up / ignite / start a verse. “Wave-function punchlines hittin’ harder than a dark spin.” Spin as a quantum property. “Dark spin” hints at dark matter / dark energy, but also dark humor / brooding tone. “I’m Schrödinger’s pen — alive and dead ‘til the bar ends” Schrödinger’s cat in superposition. Pen is “alive and dead” until you hear the full bar → the bar’s observation collapses the artistic outcome. “Open up the box and watch me box reality in sharp tens.” The Schrödinger box. “Box reality” like out-boxing someone in a ring. “Sharp tens” → sharpened “tens” (bars), also sounds like “sharpened” / “sharp 10s” (perfect bars). “Got gravity envy — ‘cause I pull the mass on the grapevines.” Gravity pulls mass. “Pull the mass” = attract masses of people / attention. “On the grapevines” = word of mouth. “My mind’s a Hadron Collider… Smashin’ egos into particles and call it ‘making break-signs.’” Large Hadron Collider smashes particles. “Break signs” ~ “brake signs” / breaking points / breaking egos. Also nod to how new particles are inferred from “signal breaks.” “Kick a lyric so atomic, every comma’s a katana.” “Atomic” lyric: explosive, small but powerful. “Comma” = tiny mark, but you weaponize it into “katana” → even punctuation cuts. “Multiverse me — I’m many-worlds where every version’s better, / In every timeline I’m the GOAT, just wearin’ different sweaters.” Many-Worlds Interpretation of QM. GOAT in every universe → layered arrogance gag. “Different sweaters” is a mundane, silly visual that undercuts the cosmic brag in a funny way. “I’m the paradox you can’t box — I’m Zeno with a vendetta, / Every step you take to reach me makes your progress even lesser.” Zeno’s paradox: you never actually arrive. “Can’t box” = can’t categorize and you can’t fight him. Great thematic coherence: opponent can’t “reach” your level. “Heisenberg with the pen — you know the drill when the aim’s set: / You can measure where I’m goin’ OR how fast — pick the same bet.” Uncertainty principle. “Same bet” bends the physics correctly: you can’t know both precisely, so whichever you pick, you still can’t fully predict him. “Turn your circle into strings and play your atoms like a fork.” Circle → strings → string theory. “Play your atoms like a fork” = tuning fork / resonant frequencies. “Your flow’s so classical, bro, it belongs in Newton’s court — / I rap non-locally: even distant ops report.” Classical mechanics vs quantum. “Newton’s court” like a king’s court. “Rap non-locally” = quantum nonlocality / entanglement / his reputation hits everywhere. “Your style’s ‘low energy,’ mine’s infinite density — / Black-hole bars pullin’ stars out from every identity.” Low vs high energy states. Black holes: infinite density, strong gravity. Pulling “stars” (famous rappers) out of their orbits / identities. “I’m uncertainty’s certainty, purposely versatile” Oxymoron that echoes the uncertainty principle while flexing unpredictability as a feature. “Call me Quinn-tum Mechanics” Quantum Mechanics + “Quinn” pun. Ties persona directly into the scientific metaphor. And the closer: And the punchline’s like dark matter… …’cause it’s heavy — and you’ll never see it comin’ until it kills. Dark matter: unseen but exerts gravitational effects. Punchline: you don’t see it coming, but it hits heavy. “Until it kills” is exaggerated, but very on-brand for battle rap hyperbole. 2. Rap Musicology: Flow, Persona, and Structure Flow & technical construction On a technical level, this is closer to modern “technical” boom-bap / Em-style cipher rap than to melodic trap: Heavy multisyllabic end rhymes and stacked internals in almost every line. Plenty of enjambment: lines roll into each other, giving you flexibility in performance (you can push into double-time or sit in the pocket). Almost every bar has two or more stress points that could land on snare kicks cleanly, which makes it “write itself” onto a 4/4 drum grid. Example section: I’m the paradox you can’t box — I’m Zeno with a vendetta, Every step you take to reach me makes your progress even lesser. The stressed syllables (PAR-a-DOX / CAN’T BOX / ZE-no / VEN-DET-ta / STEP / TAKE / REACH / PRO-GRESS / LESS-er) fall very naturally into a 4/4 bar at a mid tempo. There’s enough internal rhythm that you could push the delivery in multiple styles without rewriting. Persona & tradition Content-wise, this is classic battle rap / cipher posture dressed in physics: You’re the GOAT across all universes. Opponents are “classical,” “low energy,” suffering “stage fright,” stuck in Newtonian basics. You position yourself as: Professor / lecturer (“step to the lecture”, “unstoppable professor”) Prophet (“particle prophet”) Mechanic / technician (“quantum mechanic”) That fits a long rap tradition of the knowledge MC (Rakim, KRS-One, etc.), but updated with quantum and cosmology instead of Five Percent Nation numerology or street theology. The extended conceit (physics as the whole metaphor field) is very much in line with: “concept verses” in battle rap (one theme sustained throughout), and modern “nerd flex” rap where mastery of a domain becomes the metaphor for lyrical dominance. 3. Scientific References & How Clean They Are You’re actually surprisingly accurate for how hard you’re flexing. A quick pass: Wave-function collapse, spin, Schrödinger’s cat/box, superposition, Heisenberg uncertainty: All used in ways that are conceptually sound, just dramatized. Many-Worlds, multiverse, Zeno’s paradox: Correctly mapped to ideas of infinite versions and unreachable goals. Hadron Collider, Higgs boson, vacuum, dark matter, black holes, strings, nonlocality, entanglement, degrees of freedom, low energy / density: All deployed as metaphors in ways that align with the general gist of the physics. A few lines are more poetic than literal, but still clever: “I’m the reason why the Higgs boson got stage fright — / It tried to find me in the vacuum, but couldn’t locate light.” Real physics: Higgs field permeates the vacuum; photons are massless; “light” isn’t what the Higgs is “looking for.” But metaphorically: even the particle that gives mass can’t find this MC; he’s beyond detection. “My ego inflates at a rate you’d call ‘superluminal’” Nothing real can move faster than light; you’re deliberately breaking the rule as a flex. Fits: “Breaking every limit like the laws were merely ‘optional.’” “Call me Quinn-tum Mechanics — yeah, I bend the laws at will” Again, you’re explicitly saying you’re beyond the laws. That’s the brag. Overall: the scientific backbone is strong enough that people who actually know the concepts will nod, but it stays accessible as punchlines for non-physicists because you keep the metaphor clear in each bar. 4. Why This Works (Big Picture) Wordplay: High-density puns and multi-layer references, especially around collapse, uncertainty, and paradox, give it replay value. Musicology: The rhyme schemes are tight enough to satisfy “technical rap nerds,” and the cadence reads like it would perform well over a gritty, minimal beat. Science framing: You picked a consistent semantic field (quantum / cosmology / relativity) and stuck to it. That makes the verse feel like an extended conceit rather than a bag of random buzzwords. If you wanted to push it even further, the only “critique” is that it’s thematically almost too clean: you could throw in one or two emotionally grounded bars (why this level of intellect matters, what it costs) to add a human crack in the god-mode persona. But as a straight-up quantum battle verse? It’s nasty. ===END=== Damn.

Kyle Shannon (AI Learning Lab/AI Salon)

124,661 次观看 • 9 个月前

You are confusing cause and effect, and in doing so, you are missing the deeper structural reality that is the central point of the argument. Migration in Africa does not happen in a vacuum. People do not wake up and casually decide to leave their homes, families, and countries for no apparent reason. Movement of people across colonial borders is driven by economic collapse, political instability, conflict, and governance failures, yes, but also by powerful historical forces that shaped those very conditions in the first place. Not everyone has the third eye to see those historical forces at play unless they read, comprehend and follow ideas and not populist demagoguery. Apartheid was not just a South African policy that ended in 1994. Its effects still live with South Africans to this very day. It was part of a wider political and more importantly economic system of racial capitalism that structured the region’s economy. What you fixed in 1994 was only the political and not the economic side of it. South Africa was designed as the industrial hub, while neighbouring countries were deliberately underdeveloped and turned into labour reserves for South Africa’s economy. Migrant labour from countries like Zimbabwe, Mozambique, and Lesotho was not an accident at all, it was built into the system. It was designed that way and remains so to this very day. The owners of the means of production then remain the owners of the means of production today. Black people are largely still workers. You have a few token black individuals at the top, but the majority remain little more than exploited labour. So when people move from their countries today, they are often moving along routes that were created decades ago. The inequality between South Africa and its neighbours did not emerge overnight, and it is not simply the result of “African leaders” in isolation of other key factors. It is the continuation of a historical economic design that concentrated wealth in one place and poverty in others. That does not absolve African governments of responsibility. Many have failed their citizens through corruption, mismanagement, and repression. I write about this daily, and I have gone to prison three times in my lifetime for doing so. I have had to leave my country to save my life for doing so. But to reduce a complex, multi-layered issue to “it is African leaders” is intellectually lazy and historically dishonest. It ignores history, economics, and global power dynamics. As for Malema, whether you agree with him or not, his political skill lies in identifying how political and economic narratives are shaped and who benefits from them. He is pointing out that anger is often redirected away from the very systems of inequality and towards vulnerable people, migrants, who did not create those conditions. If you want a serious conversation, then deal with the full picture. Migration is about history, economics, governance, and global inequality. Blaming one factor while ignoring the rest is not analysis at all, it is deceitful propaganda. The economically and intellectually illiterate are often the easiest targets of political propaganda, precisely because they are fed simple, emotionally satisfying explanations for complex structural problems. They are told who to blame for their suffering, migrants, neighbouring countries, or vague notions of “outsiders”, while the real drivers, historical dispossession, entrenched economic inequality, and elite collusion, are deliberately obscured. In Southern Africa, and particularly in South Africa, this manifests in xenophobic narratives that blame Zimbabweans or Mozambicans for unemployment and poverty, when in reality those conditions are rooted in a long standing economic architecture that concentrated wealth and ownership in very few hands. It is easier to turn the poor against the poor than to confront systems that benefit those in power. What is often forgotten in this debate is that the political elites of colonial South Africa and Rhodesia worked in concert to sustain a repressive regional system that enriched a minority while extracting labour and resources from the rest. Your former apartheid Prime Minister John Vorster says it in this video in a very tactful manner. That logic has not disappeared at all, it has merely changed form. Today, segments of the political elite in both South Africa and Zimbabwe continue to operate in ways that protect entrenched economic interests while the majority remain economically marginalised. South Africa was the only true white settler “home”, where wealth, infrastructure, and industry were concentrated, while territories like Southern Rhodesia (Zimbabwe), Northern Rhodesia (Zambia), and Nyasaland (Malawi) functioned largely as economic outposts, feeding capital, labour, and raw materials into that system. The Federation of Rhodesia and Nyasaland was presented as a project of regional integration, but in practice it reinforced patterns of extraction, with mining in Zambia, agriculture in Zimbabwe, and labour flows from Malawi all tied into a broader economic network dominated by South African capital. The same remains to this very day. The tragedy of focusing on Julius as the messenger rather than the message is something I speak about regularly, the need to focus on ideas and not personalities. You do not have to like Julius Malema. You do not have to agree with everything he says. All you need to do is focus on his message and interrogate it critically. I am not enslaved to Julius Malema’s ideas. I pick and choose what I agree with, and I am able to articulate a reasoned argument for both what I support and what I reject. You should do the same. One of the most powerful weapons of colonialism was the deliberate fragmentation of black people into small Bantustans, into isolated villages where communities were conditioned to view the next village with suspicion. In Rhodesia we had “reserves” and “keeps.” People from other Bantustans were treated as outsiders. That mentality was never dismantled, it still exists today. The idea of seeing others with suspicion simply because of an arbitrary line, a colonial border, remains deeply entrenched. Many do not fully appreciate how powerful and enduring that mentality and conditioning is. Yet when you look at the descendants of colonialists, they do not view each other through those same lenses. White Zimbabweans move into South Africa without attracting the same hostility because of the economic architecture that allows them to stay away from the so called lumpen. White people from across the world come and settle with ease in South Africa. In fact, one of the most visible figures advocating for the secession of the Western Cape is a British citizen, yet there is no comparable outrage from black South Africans. The same energy of protests and marches that is directed at fellow Africans is rarely directed there. That is not accidental at all, it is well designed that way. It speaks to the protection afforded by entrenched economic power and privilege, but also to a deeper psychological conditioning in how black people are taught to see each other and to see whiteness. This will not disappear overnight. It may not even disappear in my lifetime. But the task is to keep planting the seeds of awareness and unity. As Bob Marley said, you give your more to get your little. What you do today may seem small, but in time it contributes to something much larger, especially if there is collective effort to confront and resolve these divisions. One of the most important things colonialists understood was that education is the key to discernment, to the ability to interrogate and understand issues such as those I raised in this essay. That is precisely why they restricted access to it. Only a few black people were allowed meaningful education, and the consequences of that exclusion remain with us today, not only in South Africa but across much of the continent. We did not dismantle the systems that underpinned colonialism. We largely inherited them, changed the faces at the top, and continued to operate within the same structures. So I will end by saying this, if anyone truly wants change on the issues being debated, you must fix the foundation. You cannot repair window panes when the foundation itself is cracking. Immigration, whether legal or illegal, will always exist, but it is sustained not by foreigners alone, but by the system itself. When Zimbabweans cross the border without passports, they are often enabled by South Africans within a broken system. When documents are obtained illegally, it is again the system that enables it. When Zimbabwe’s political crisis persists without free and fair elections, regional dynamics, including South Africa’s political and economic interests, often play a role in sustaining that status quo. There is a web of political and economic interests that mirrors, in some respects, the relationships that existed during the colonial and apartheid eras. As long as those interests remain, there is little incentive for those in power to confront injustice decisively. The corruption and governance failures in Zimbabwe are real and significant, but they are part of a broader structural problem. The real issue is the foundation. If black South Africans were living well, with access to quality education, meaningful employment, and economic security, they would not be marching in the streets. The anger you see today is not simply about immigration. It is a reflection of an economic structure that has remained fundamentally unchanged, even after 1994. Repression underpinned by racism in Rhodesia effectively came to an end when South Africa shifted its position and recognised that the system was no longer sustainable. The same principle applies today. Repression underpinned by political corruption in Zimbabwe will begin to end the day South Africa, the regional power whether one accepts it or not, decides that the current situation is no longer acceptable. Zimbabwe’s crisis has, over time, been treated as a largely domestic issue rather than a regional one, yet the political and economic realities of Southern Africa make that distinction artificial. What happens in Zimbabwe does not exist in isolation, it is shaped, sustained, and, at times, enabled by regional dynamics, particularly South Africa’s stance. This may be an uncomfortable truth, but history consistently shows that regional power centres play a decisive role in determining outcomes. Ignoring that reality does not change it, it only delays the moment when it must be confronted. It was convenient then for John Vorster and successive apartheid regimes to continue using illegal migrants as a source of cheap labour in South Africa for menial jobs. It remains the same today. As I have said, the political and economic architecture of the apartheid era largely remains in place. What has changed are the political faces, the white faces that held power then and the black faces that hold office today, often operating within and alongside the same entrenched economic structures. Whether one accepts it or not, that is the reality of our politics in the region and of the economic architecture that continues to shape it. There is a reason why certain political actors avoid critically engaging with the structural drivers of immigration, particularly those that sustain flows of cheap labour. There is also a reason why figures like Helen Zille often emphasise the need to document illegal immigrants in South Africa, that position can be understood within the broader context of preserving an economic order that has long depended on controlling and managing labour rather than fundamentally transforming the conditions that produce it. That economic order is rooted in historical structures of concentrated power that shaped not only South Africa, but the wider region more than a century ago. How black Africans view themselves is often reflected in how they respond to political messages. It is why some are quick to criticise Julius Malema for positions that are, in substance, not fundamentally different from those expressed by Helen Zille. On immigration, there is significant overlap in what has been said across the political spectrum, including by the DA and the EFF. Yet the EFF is frequently viewed through a lens of hostility, in part because it is a black-led party, and that perception shapes the reaction it receives. As a result, some black citizens, influenced by long-standing narratives, direct harsher and more emotive criticism towards it. When similar points are made by figures like Helen Zille, the response is often markedly different. That contrast speaks to deeper historical conditioning and the psychological legacy of colonialism. It has not disappeared, and changing it will take time. The fundamental difference, however, lies in the intent and framing of their messages. Malema’s position on immigration is part of a broader effort to confront and address the structural inequalities created by colonial rule. Zille’s position, by contrast, can be seen as operating within and reinforcing an existing economic framework that has its roots in that same colonial architecture which feeds off cheap migrant labour. However, you can't fix the broken system by chasing away immigrants, legal or illegal, you can only empower black South Africans by allowing them to own the means of production and not fighting in the streets for crumbs. Have a lovely weekend.
6:23

You are confusing cause and effect, and in doing so, you are missing the deeper structural reality that is the central point of the argument. Migration in Africa does not happen in a vacuum. People do not wake up and casually decide to leave their homes, families, and countries for no apparent reason. Movement of people across colonial borders is driven by economic collapse, political instability, conflict, and governance failures, yes, but also by powerful historical forces that shaped those very conditions in the first place. Not everyone has the third eye to see those historical forces at play unless they read, comprehend and follow ideas and not populist demagoguery. Apartheid was not just a South African policy that ended in 1994. Its effects still live with South Africans to this very day. It was part of a wider political and more importantly economic system of racial capitalism that structured the region’s economy. What you fixed in 1994 was only the political and not the economic side of it. South Africa was designed as the industrial hub, while neighbouring countries were deliberately underdeveloped and turned into labour reserves for South Africa’s economy. Migrant labour from countries like Zimbabwe, Mozambique, and Lesotho was not an accident at all, it was built into the system. It was designed that way and remains so to this very day. The owners of the means of production then remain the owners of the means of production today. Black people are largely still workers. You have a few token black individuals at the top, but the majority remain little more than exploited labour. So when people move from their countries today, they are often moving along routes that were created decades ago. The inequality between South Africa and its neighbours did not emerge overnight, and it is not simply the result of “African leaders” in isolation of other key factors. It is the continuation of a historical economic design that concentrated wealth in one place and poverty in others. That does not absolve African governments of responsibility. Many have failed their citizens through corruption, mismanagement, and repression. I write about this daily, and I have gone to prison three times in my lifetime for doing so. I have had to leave my country to save my life for doing so. But to reduce a complex, multi-layered issue to “it is African leaders” is intellectually lazy and historically dishonest. It ignores history, economics, and global power dynamics. As for Malema, whether you agree with him or not, his political skill lies in identifying how political and economic narratives are shaped and who benefits from them. He is pointing out that anger is often redirected away from the very systems of inequality and towards vulnerable people, migrants, who did not create those conditions. If you want a serious conversation, then deal with the full picture. Migration is about history, economics, governance, and global inequality. Blaming one factor while ignoring the rest is not analysis at all, it is deceitful propaganda. The economically and intellectually illiterate are often the easiest targets of political propaganda, precisely because they are fed simple, emotionally satisfying explanations for complex structural problems. They are told who to blame for their suffering, migrants, neighbouring countries, or vague notions of “outsiders”, while the real drivers, historical dispossession, entrenched economic inequality, and elite collusion, are deliberately obscured. In Southern Africa, and particularly in South Africa, this manifests in xenophobic narratives that blame Zimbabweans or Mozambicans for unemployment and poverty, when in reality those conditions are rooted in a long standing economic architecture that concentrated wealth and ownership in very few hands. It is easier to turn the poor against the poor than to confront systems that benefit those in power. What is often forgotten in this debate is that the political elites of colonial South Africa and Rhodesia worked in concert to sustain a repressive regional system that enriched a minority while extracting labour and resources from the rest. Your former apartheid Prime Minister John Vorster says it in this video in a very tactful manner. That logic has not disappeared at all, it has merely changed form. Today, segments of the political elite in both South Africa and Zimbabwe continue to operate in ways that protect entrenched economic interests while the majority remain economically marginalised. South Africa was the only true white settler “home”, where wealth, infrastructure, and industry were concentrated, while territories like Southern Rhodesia (Zimbabwe), Northern Rhodesia (Zambia), and Nyasaland (Malawi) functioned largely as economic outposts, feeding capital, labour, and raw materials into that system. The Federation of Rhodesia and Nyasaland was presented as a project of regional integration, but in practice it reinforced patterns of extraction, with mining in Zambia, agriculture in Zimbabwe, and labour flows from Malawi all tied into a broader economic network dominated by South African capital. The same remains to this very day. The tragedy of focusing on Julius as the messenger rather than the message is something I speak about regularly, the need to focus on ideas and not personalities. You do not have to like Julius Malema. You do not have to agree with everything he says. All you need to do is focus on his message and interrogate it critically. I am not enslaved to Julius Malema’s ideas. I pick and choose what I agree with, and I am able to articulate a reasoned argument for both what I support and what I reject. You should do the same. One of the most powerful weapons of colonialism was the deliberate fragmentation of black people into small Bantustans, into isolated villages where communities were conditioned to view the next village with suspicion. In Rhodesia we had “reserves” and “keeps.” People from other Bantustans were treated as outsiders. That mentality was never dismantled, it still exists today. The idea of seeing others with suspicion simply because of an arbitrary line, a colonial border, remains deeply entrenched. Many do not fully appreciate how powerful and enduring that mentality and conditioning is. Yet when you look at the descendants of colonialists, they do not view each other through those same lenses. White Zimbabweans move into South Africa without attracting the same hostility because of the economic architecture that allows them to stay away from the so called lumpen. White people from across the world come and settle with ease in South Africa. In fact, one of the most visible figures advocating for the secession of the Western Cape is a British citizen, yet there is no comparable outrage from black South Africans. The same energy of protests and marches that is directed at fellow Africans is rarely directed there. That is not accidental at all, it is well designed that way. It speaks to the protection afforded by entrenched economic power and privilege, but also to a deeper psychological conditioning in how black people are taught to see each other and to see whiteness. This will not disappear overnight. It may not even disappear in my lifetime. But the task is to keep planting the seeds of awareness and unity. As Bob Marley said, you give your more to get your little. What you do today may seem small, but in time it contributes to something much larger, especially if there is collective effort to confront and resolve these divisions. One of the most important things colonialists understood was that education is the key to discernment, to the ability to interrogate and understand issues such as those I raised in this essay. That is precisely why they restricted access to it. Only a few black people were allowed meaningful education, and the consequences of that exclusion remain with us today, not only in South Africa but across much of the continent. We did not dismantle the systems that underpinned colonialism. We largely inherited them, changed the faces at the top, and continued to operate within the same structures. So I will end by saying this, if anyone truly wants change on the issues being debated, you must fix the foundation. You cannot repair window panes when the foundation itself is cracking. Immigration, whether legal or illegal, will always exist, but it is sustained not by foreigners alone, but by the system itself. When Zimbabweans cross the border without passports, they are often enabled by South Africans within a broken system. When documents are obtained illegally, it is again the system that enables it. When Zimbabwe’s political crisis persists without free and fair elections, regional dynamics, including South Africa’s political and economic interests, often play a role in sustaining that status quo. There is a web of political and economic interests that mirrors, in some respects, the relationships that existed during the colonial and apartheid eras. As long as those interests remain, there is little incentive for those in power to confront injustice decisively. The corruption and governance failures in Zimbabwe are real and significant, but they are part of a broader structural problem. The real issue is the foundation. If black South Africans were living well, with access to quality education, meaningful employment, and economic security, they would not be marching in the streets. The anger you see today is not simply about immigration. It is a reflection of an economic structure that has remained fundamentally unchanged, even after 1994. Repression underpinned by racism in Rhodesia effectively came to an end when South Africa shifted its position and recognised that the system was no longer sustainable. The same principle applies today. Repression underpinned by political corruption in Zimbabwe will begin to end the day South Africa, the regional power whether one accepts it or not, decides that the current situation is no longer acceptable. Zimbabwe’s crisis has, over time, been treated as a largely domestic issue rather than a regional one, yet the political and economic realities of Southern Africa make that distinction artificial. What happens in Zimbabwe does not exist in isolation, it is shaped, sustained, and, at times, enabled by regional dynamics, particularly South Africa’s stance. This may be an uncomfortable truth, but history consistently shows that regional power centres play a decisive role in determining outcomes. Ignoring that reality does not change it, it only delays the moment when it must be confronted. It was convenient then for John Vorster and successive apartheid regimes to continue using illegal migrants as a source of cheap labour in South Africa for menial jobs. It remains the same today. As I have said, the political and economic architecture of the apartheid era largely remains in place. What has changed are the political faces, the white faces that held power then and the black faces that hold office today, often operating within and alongside the same entrenched economic structures. Whether one accepts it or not, that is the reality of our politics in the region and of the economic architecture that continues to shape it. There is a reason why certain political actors avoid critically engaging with the structural drivers of immigration, particularly those that sustain flows of cheap labour. There is also a reason why figures like Helen Zille often emphasise the need to document illegal immigrants in South Africa, that position can be understood within the broader context of preserving an economic order that has long depended on controlling and managing labour rather than fundamentally transforming the conditions that produce it. That economic order is rooted in historical structures of concentrated power that shaped not only South Africa, but the wider region more than a century ago. How black Africans view themselves is often reflected in how they respond to political messages. It is why some are quick to criticise Julius Malema for positions that are, in substance, not fundamentally different from those expressed by Helen Zille. On immigration, there is significant overlap in what has been said across the political spectrum, including by the DA and the EFF. Yet the EFF is frequently viewed through a lens of hostility, in part because it is a black-led party, and that perception shapes the reaction it receives. As a result, some black citizens, influenced by long-standing narratives, direct harsher and more emotive criticism towards it. When similar points are made by figures like Helen Zille, the response is often markedly different. That contrast speaks to deeper historical conditioning and the psychological legacy of colonialism. It has not disappeared, and changing it will take time. The fundamental difference, however, lies in the intent and framing of their messages. Malema’s position on immigration is part of a broader effort to confront and address the structural inequalities created by colonial rule. Zille’s position, by contrast, can be seen as operating within and reinforcing an existing economic framework that has its roots in that same colonial architecture which feeds off cheap migrant labour. However, you can't fix the broken system by chasing away immigrants, legal or illegal, you can only empower black South Africans by allowing them to own the means of production and not fighting in the streets for crumbs. Have a lovely weekend.

Hopewell Chin’ono

21,194 次观看 • 4 个月前

🟢GIVEAWAY🟢 Best comments or memes about this whole circus + RT this post. 10 winners will each get $50💎 (For evidence, supporting materials, and context, read both articles and watch the video included in the article I posted yesterday) Housebets.com & Porchy pay your debts A few people told me they did not fully understand the first article because there were too many moving parts: leaderboard accounts, rewards, weekly dates, monthly bonus, Tequity, game categories, withdrawals, Provably Fair, seed changes, migration, support tickets, ledgers and founder messages. Fair enough. The evidence is already there, and I still recommend reading the full articles and, above all, watching the video, because the video shows the reward system failing live. But this text is the cleaner version: the full story explained in plain English, without assuming the reader knows anything about crypto casinos, leaderboards or lossback systems. From all the evidence I’ve gathered, the Housebets story is not a normal “player lost money” complaint. It looks like a full transparency failure across the whole product: leaderboard, rewards, withdrawals, game categories, Provably Fair / Tequity mapping, support, migration and founder response. Housebets sold itself as a rewards-first casino: public leaderboards, weekly/monthly bonuses, fast withdrawals, VIP treatment and Provably Fair games. But every time I asked for the records behind those systems, snapshots, ledger entries, weekly cycles, GGR/NGR, slider logs, PF seed mapping, Tequity round IDs, withdrawal approval logs, the answer became some version of “forwarded to the relevant department.” This started long before the public dispute. I was not some random angry player who appeared after one bad session. In January I was helping Housebets and giving product feedback. I literally told support on 27 January that I was “testing the website for George,” while already dealing with a non-instant withdrawal and a 100% welcome bonus that had not applied. Support even asked me for “proof about your testing job.” The same chat shows the advertised 100% Welcome Bonus, the bonus not applying, and support saying the withdrawal needed internal confirmation instead of being instant. The welcome bonus issue never looked clean. Housebets advertised a 100% Welcome Bonus up to $1,000 on first deposit; I deposited, contacted support, and the bonus did not apply. Then support effectively turned a first-deposit bonus into a second-deposit workaround because the first one had not been applied properly. On 31 January I came back after another deposit and told them the bonus still had not been applied, even though I had already followed support’s instructions. Edward replied that he had “forwarded” the concern to the team. The same 100% welcome bonus was still being advertised in March. By April, the rewards system was already showing serious problems. I had the weekly slider at 100% lossback and told support I had lost money but the weekly did not appear. Jacky said the weekly was generated every Thursday at 00:01 UTC and gave actual internal figures: GGR $6,250, Total Bonus $6,083.99, NGR $168.31. So Housebets clearly had internal calculations when it wanted to explain why something might not pay. But when I later asked for full calculations, those same numbers suddenly became impossible to produce. Then on 18–19 April, the rewards page was bugged and would not let me claim. Support could see a pending weekly bonus of $717.37, but I could not claim it from the UI. Tee said it had been forwarded to the relevant department. That $717.37 later appears in the bonus ledger as Rakeback (20 Apr) 717.37089061, so I am not saying that specific one stayed unpaid forever. The point is worse: already in April, support could see a pending weekly reward while the player-facing reward page did not work. For a casino built around rewards, that is not a small bug. That is the product. In May, the UI and account data kept failing basic trust checks. On 8 May, I deposited 400 USDT; support said it had been credited, but I could not see it, and the proposed fix was to log out, clear cookies and cache. On 16 May, I asked why total deposits and withdrawals had disappeared from the menu; support said the platform was “in continuous evolution.” On 17 May, I asked for my total deposits and withdrawals, and support said they did not have direct access to that consolidated summary and would email it. That full official ledger did not arrive. So when Housebets later defends itself with UI screenshots, remember: this was the same UI where deposits could be credited but invisible, totals disappeared, rewards pages bugged, and support could not access consolidated account totals. Withdrawals were also not what was advertised. On 16 May, I asked why a crypto withdrawal was pending if withdrawals were supposed to be instant. Tee answered: “A few withdrawals require manual approval,” then added, “Our withdrawals are typically instant but…” That matters because a few days later the withdrawal delay became real damage. On 25 May, I told support before a match that I needed the funds to place a time-sensitive bet on another site in less than 20 minutes. I explained I wanted to bet around 60k at odds of 2.55. The withdrawal did not arrive in time. Later I told them the bet won and that I missed around 90k in profit because Housebets took more than two hours despite being warned before the match started. Jacky said he would raise the compensation case to the VIP team. Nobody resolved it. This was not one delayed withdrawal either. In my formal complaint I reconstructed several withdrawal delays: 23 May 02:55 → 08:03, around 5h08m; 25 May 03:05 → 08:09, around 5h04m; 17 May 03:54 → 08:02, around 4h08m; 18 May 04:46 → 08:11, around 3h25m; 16 May 05:23 → 08:12, around 2h49m. That is not “instant withdrawal.” And if later marketing says withdrawals are much faster now, the obvious question is: if this was the faster version, what did slow look like? The Provably Fair / Tequity side was another major issue. On 17 May I asked support how to verify an old Blackjack round. I did not ask for a generic explanation of Provably Fair; I asked where I could see the server seed, client seed, nonce and result for previous games. Support sent me to bet history, mentioned RTP, gave a generic PF explanation and showed the current Dice seed screen. When I said that did not let me verify previous games, they told me to clear cookies/cache. After doing that, I saw a new client seed and nonce 1 even though I had not played with that seed pair. I asked if Housebets changes seeds on every login. Support could not answer and told me to contact VIP. That seed/session behaviour is important. I later recorded video evidence around the seed changing after clearing cookies/cache and asked for the exact mapping: Housebets account ID → Tequity/provider player ID → session/currency context → seed pair → server seed hash → revealed server seed → client seed → nonce/cursor → raw outcome → final result. Housebets cannot sell Provably Fair if the player cannot verify historical bets, and “contact VIP” is not a verification algorithm. On 24 May, I asked for raw verification data for a specific Tequity Blackjack round: Round ID e1648d60-0da1-4433-a5ab-9ae39f5302e3, Blackjack, Tequity, bet amount 11,346 USDT, client seed O3YBZF7LBu, server seed hash starting 712875.... I asked for revealed server seed, nonce, full result JSON, card draw order and verification algorithm. I also asked about an apparent duplicate-card/deck question. Tee replied: “I don’t have the answers to your questions right now, but I’m forwarding your request to the relevant department.” That same day, I asked for a full audit of six Dice bets of 11,400 USDT each, total 68,400 USDT. I requested bet IDs, provider round IDs, roll results, seed data, balance ledger, request/session logs, security logs, retry flags, provider records and a full technical reconciliation. Tee replied: “I will forward this to the relevant department.” So when I asked for raw data, the answer was not data. It was forwarding. Again. There were also many large loss clusters that required reconciliation because of those unresolved PF, Tequity, category, RTP and session questions. In my complaint I listed clusters such as 25 May 02:17–02:54 Blackjack around 169,932 USDT; 16 May 12:31–13:26 Dice around 90,571.92 USDT; 26 May 02:48–03:58 Mines around 89,199 USDT; 24 May 06:20–06:21 Dice at 68,400 USDT; 26 May 00:11–01:41 Blackjack around 59,910 USDT; 25 May 22:51–22:59 Dice around 59,576 USDT; and several more between 40k and 56k. I am not saying every losing cluster proves manipulation by itself. I am saying that when PF mapping, provider logs, RTP/HE, category mapping and seed/session behaviour are unresolved, these sequences need a real reconciliation. The leaderboard is where the story becomes very hard for Housebets to explain. Around 19–20 May, two new accounts, elmourabut and lucasmartirini, appeared and started climbing every day at a vertiginous pace. Not normal slow leaderboard growth. Not a casual player building volume over time. They were created around that period and then started rising with huge wagering in a way that looked extremely unnatural for brand new accounts. By 29 May, I was first on both weekly and monthly leaderboards, and those two accounts were directly behind me with huge volume. In the monthly leaderboard screenshots, I was around $3.33M wagered, while elmourabut was around $1.29M and lucasmartirini around $1.08M. In the weekly leaderboard, I was around $1.096M, while those two accounts were around $635k and $578k. They were not normal accounts sitting at the bottom; they were directly behind me, applying pressure. In my formal complaint I recorded that elmourabut joined on 19 May and lucasmartirini on 20 May, that they showed zero visible withdrawals, large deposits/wagering and significant card-game volume, and I asked Housebets to confirm they were not staff, test, QA, admin, house-controlled, affiliate-controlled, internally funded, promotional, bonus-only or multi-account related accounts. This matters because a leaderboard is not passive. It is gamification. It makes players defend rank. When two new accounts appear behind you with hundreds of thousands or more than a million in volume, you are pressured to keep wagering. In my case, the disputed deposit sequence from 25 May 22:23 to 26 May 02:09 totals 91,168.375326 USDT. That sequence begins with 1,000.00 at 22:23 and continues with repeated deposits until 2,879.148969 at 02:09. The video later shows why those dates matter: there were deposits coming in, no gameplay withdrawal offsetting the sequence, a balance basically at zero, and later a leaderboard prize shown as P/L. I formally asked Housebets to confirm those two leaderboard accounts were real and eligible, and also to preserve wager logs, transaction records, balance adjustment logs, account flags, leaderboard calculation snapshots, support ticket logs, Telegram/email records and internal notes. Edward said he forwarded the request. In the same thread, he added that they were “working on fixing an issue regarding the weekly bonuses,” and then said the weekly countdown was “not currently on Thursday evenings.” So the leaderboard issue and the weekly bonus issue are linked in time and support context. After that, Housebets confirmed by email that elmourabut and lucasmartirini were “legitimate and eligible accounts.” That email is the trap door. If they were legitimate and eligible, they should have remained in the leaderboard with their volume. If they were not, Housebets should never have confirmed them as legitimate and eligible. After that confirmation, the accounts disappeared from the leaderboard or stopped appearing in the positions their previous wagering required. I went back to support on 30 May and wrote: “There has been a material post-confirmation leaderboard change involving two accounts that Housebets had already confirmed as legitimate and eligible. I need the exact reason, timestamp, logs, and recalculation basis.” Edward said the matter was flagged and that I could expect a prompt response. I am still waiting for the actual explanation. Why did they disappear? My read is simple: because every hour that passed, there was more evidence around those accounts. They had been created around the same period, they were climbing at a speed that looked anything but human, they showed no visible withdrawals in the data I could see and reported, they appeared to be generating huge volume in unclear game categories, and the games/categories tied to that volume did not even make sense from the player-facing UI. When I started asking what they were actually playing, what Card meant, whether the volume was Tequity / UnOriginals / House Games, what RTP and house edge applied, and where the logs were, the questions became uncomfortable. Keeping those accounts visible became harder than removing them. So they disappeared. The game category issue made the leaderboard even more suspicious. On 30 May, I asked support why my own stats showed almost all my volume under Slots / Tragamonedas when I did not play real slots. I told them: “i dont play 3$ in unoriginals,” “i played all 3M in unoriginals,” and “ive never play slots.” I asked what “Card” was, where that game was, what RTP and house edge it had. Monica said Card was mainly Blackjack, Baccarat and Poker variants. Marcus later said the team was investigating why it showed that I mostly played slots when I had not. He could not give the exact game, RTP, HE, provider, category mapping or contribution logic. That matters because those same unclear categories were connected to leaderboard volume. If the site cannot clearly explain whether volume is Slots, Card, UnOriginals, House Games, Blackjack, Baccarat, Always 9 Baccarat or Tequity, then the leaderboard is not auditable for the player. I even asked which UnOriginals those two accounts were playing, and support told me to look at Live Bets. That is not an answer. I was not asking for gossip; I was asking what exact games generated leaderboard volume, what RTP/HE applied and whether that volume was eligible. There is also an earlier leaderboard-related precedent: Porchy had already told me in February that I would lose leaderboard places if I did not rename, because too many people were messaging support saying the site was not being fair due to my name and it “doesn’t make us look good.” That matters because it suggests leaderboard positioning was not treated as a sacred, untouchable system when public perception was involved. If leaderboard positions can be threatened for image reasons, then later claims that everything is purely automatic deserve scrutiny. Then Porchy made the leaderboard situation worse. Instead of producing logs or snapshots, he later said the leaderboard had “abusers” on it, that they were removed to help other players, and that it never affected me. Later he said they paid every single person, “even these abusers,” then called me “begging for money.” That creates a direct contradiction: Housebets confirmed the accounts as legitimate and eligible, then Porchy referred to leaderboard “abusers.” If they were abusers, why were they confirmed as legitimate and eligible? If they were eligible, why did they disappear? If they never affected me, where are the historical snapshots proving that? Once those accounts disappeared, Housebets paid the leaderboard prizes. On 1 June, the bonus ledger shows two Leaderboard entries: 5,007.46111706 and 1,001.49222341, totaling 6,008.95334047. That part was paid. But then Act Two started: the weekly and monthly rewards did not appear as separate ledger entries. The same bonus ledger shows those two 1 June entries as Leaderboard only, not Monthly Bonus, not Weekly Reload, not Lossback. The weekly timeline is a mess. On 28 May, the dashboard / UI said the weekly bonus was claimable every Thursday at 00:01 UTC, and the monthly was available on the 1st at 00:01 UTC. That same night I told support the weekly had shown as available, then reset to 6 days without paying. Later I sent screenshots and wrote: “1M wagered and 0.2$.” Jacky said he had raised the issue to the technical team. So the weekly failure was reported live, not reconstructed after the fact. The next day, 29 May, Edward said they were fixing an issue regarding weekly bonuses and that the weekly countdown was “not currently on Thursday evenings.” Then on 1 June, Spencer said the May weekly bonuses were 7th, 14th, 21st, and then due to migration the weekly moved to Monday, so there was one on the 25th on the new platform. He also said the 25 May weekly covered gameplay from 21–24 May, and that tech was looking at that plus the monthly bonus. The ledger does show a 25 May 02:10 Rakeback entry of 1,996.08334791, which likely corresponds to that 21–24 May weekly. But my major loss sequence starts about 20 hours later, on 25 May at 22:23, and continues until 26 May at 02:09. So the 25 May weekly cannot cover those losses. If weekly was still Thursday, the 25/26 losses should have been in the 28 May weekly. But the bonus ledger on 28 May shows only two tiny Rakeback entries, 0.28373945 and 0.00280958. If weekly moved to Monday because of migration, those losses should have appeared in the next weekly after 25 May. But on 1 June the ledger only shows Leaderboard entries. Then the final video shows the next Weekly Reload reaching zero, paying nothing and resetting to 6d 23h. So the same loss sequence appears to fall into no paid weekly cycle. The 4 June support conversation makes this even more ridiculous. After I recorded the weekly reset video, I asked support a very simple question: what were the last weekly dates/cycles? The dashboard / support flow again said weekly bonuses are claimable every Thursday at 00:01 UTC. Jacky confirmed: “Weekly bonuses can be claimed every Thursday at 00:01 UTC in the Rewards tab,” and added that if not claimed by the following Wednesday at 23:59 UTC, it expires. But when I asked for the exact last four dates, Jacky said he had to check with the relevant department. When I pressed again, he said, “Sorry, As I am only a CS, Let me raise your concerns to relevant department.” I asked whether support did not have the information or simply could not answer. He replied: “Do you have any other concerns?” They use weekly cycles to decide whether to pay, but support cannot explain the weekly cycle. The monthly is missing too. The dashboard / UI said the monthly bonus is based on activity and VIP level from the previous month and is available on the 1st at 00:01 UTC. In May I had more than 3,258,023.0829 wagered according to the formal complaint data. I also have proof/video that the monthly slider was set to 50/50. On 1 June, Spencer first told me I had claimed the Monthly Bonus at 1:12am BST around the same time as the monthly leaderboard reward. I immediately said I only received leaderboard prizes. Then Spencer changed the answer: “Our tech team are still actively working on issues regarding the monthly bonuses.” So first the monthly was claimed, then tech was still fixing it. The ledger still shows no Monthly Bonus entry. Housebets then seems to rely on “up overall” as a defence. But the video and ledger show why that does not work. My weekly/monthly profile later showed around +6,008 P/L with 0 deposits, 0 wagered and around 6,008 in bonuses. That number matches exactly the two 1 June Leaderboard payments. So the UI is showing leaderboard rewards as P/L. Then support used “up overall” to say I was not eligible for weekly lossback. That is not a clean lossback calculation. That is using a leaderboard reward as apparent profit to deny a lossback that should be based on actual eligible losses. There were also smaller reward-confusion issues along the way. On 22 May I asked for all pending bonuses,weekly, monthly, rakeback, level-up, anything, and support said the internal team would manually verify whether everything had been credited correctly and email me. On 24 May, I asked about level-up rewards because the reward looked like $3,500 for Pearl; support clarified it was $3,500 total across all Pearl levels, $500 per level. These are not the core issues, but they are part of the same pattern: rewards marketing, unclear UI, manual verification, emails that do not arrive, and players having to chase basic explanations. Then there is the migration. On 25 May, after the delayed withdrawal, missing VIP contact and unresolved issues, support told me my account would be moved to the new platform and that this upgrade would offer a better withdrawal process and fix many issues. Before that migration, I explicitly requested that no account data, internal data, logs, balance history, bonus history, bet history, provider records or pending issues be deleted. The response: “Your request has been relayed to the relevant department.” Again, forwarding. But if the old data is safe, Housebets should provide the old leaderboard snapshots, old weekly states, old bonus logs, old Tequity mapping and old withdrawal approval logs. The founder response did not fix anything. When Porchy finally engaged, he did not provide the records. He framed the settlement request as “so you want $100,000?” and asked whether I needed it or else I was going to post on X. I had already made clear this was not money for silence; I asked for logs, snapshots, withdrawal records, calculations and a counter-calculation if Housebets disagreed. He later referred to “abusers,” told me I was “up overall,” said “You are begging for money,” and suggested I “just do this to casinos.” Still no ledger. Still no weekly calculation. Still no monthly entry. Still no PF/Tequity mapping. Still no leaderboard snapshots. Another player also contacted me with screenshots pointing to similar categories of issues: private deals, leaderboard payout disputes, migration/account merge problems, missing history and a tiny monthly bonus despite claimed losses. I am not using that player’s case as the foundation of my claim without his full ledger, but it matters because it suggests the same type of opacity may not be isolated: private VIP/reward deals, leaderboard eligibility, monthly bonus calculations, migration and unclear history. If Housebets has private deals that affect leaderboard eligibility or rewards, it must explain how those deals interact with public leaderboards. So the overall picture is this: Housebets sold a public leaderboard and rewards system that pressured real wagering. Two new accounts appeared directly behind me with huge volume, were confirmed as legitimate and eligible, then disappeared after I asked for logs and questioned game categories. Housebets could not explain the exact games, RTP, house edge or category mapping behind the volume. The accounts were later framed by Porchy as “abusers,” contradicting the earlier eligibility confirmation. Once Housebets paid me the leaderboard prizes, those prizes were shown as P/L, and that contaminated P/L was then used to claim I was “up overall” and not eligible for lossback. At the same time, my real 25 May 22:23 → 26 May 02:09 loss sequence of 91,168.375326 USDT appears in no clean weekly cycle. The 25 May weekly covered 21–24 May according to Spencer, so it cannot cover that loss sequence. The 28 May weekly showed only tiny Rakeback entries and was already reported as broken. The 1 June ledger shows only Leaderboard entries. The later video shows Weekly Reload reaching zero, paying nothing and resetting. And when I ask support for the exact weekly calendar, they cannot answer and send it to the relevant department. The monthly is the same story. The dashboard / UI says it is based on activity and VIP. I had more than 3.25M wagered in May. Spencer first says I claimed it, then says tech is still working on monthly bonuses. The ledger shows no Monthly Bonus. If Housebets says I was not eligible, they need to show the formula, slider history, cycle, GGR/NGR, eligible loss/activity, deductions and ledger result. If they cannot, “not eligible” is just another label. And this opens another can of worms: Tequity / provider configuration. Housebets cannot hide behind “the provider” whenever something goes wrong. The player does not deposit with Tequity. The player does not withdraw from Tequity. The player does not speak to Tequity support. The player does not compete in a Tequity leaderboard. The player plays on Housebets, with a Housebets wallet, Housebets UI, Housebets rewards, Housebets leaderboard and Housebets support. 1/2
22:49

🟢GIVEAWAY🟢 Best comments or memes about this whole circus + RT this post. 10 winners will each get $50💎 (For evidence, supporting materials, and context, read both articles and watch the video included in the article I posted yesterday) Housebets.com & Porchy pay your debts A few people told me they did not fully understand the first article because there were too many moving parts: leaderboard accounts, rewards, weekly dates, monthly bonus, Tequity, game categories, withdrawals, Provably Fair, seed changes, migration, support tickets, ledgers and founder messages. Fair enough. The evidence is already there, and I still recommend reading the full articles and, above all, watching the video, because the video shows the reward system failing live. But this text is the cleaner version: the full story explained in plain English, without assuming the reader knows anything about crypto casinos, leaderboards or lossback systems. From all the evidence I’ve gathered, the Housebets story is not a normal “player lost money” complaint. It looks like a full transparency failure across the whole product: leaderboard, rewards, withdrawals, game categories, Provably Fair / Tequity mapping, support, migration and founder response. Housebets sold itself as a rewards-first casino: public leaderboards, weekly/monthly bonuses, fast withdrawals, VIP treatment and Provably Fair games. But every time I asked for the records behind those systems, snapshots, ledger entries, weekly cycles, GGR/NGR, slider logs, PF seed mapping, Tequity round IDs, withdrawal approval logs, the answer became some version of “forwarded to the relevant department.” This started long before the public dispute. I was not some random angry player who appeared after one bad session. In January I was helping Housebets and giving product feedback. I literally told support on 27 January that I was “testing the website for George,” while already dealing with a non-instant withdrawal and a 100% welcome bonus that had not applied. Support even asked me for “proof about your testing job.” The same chat shows the advertised 100% Welcome Bonus, the bonus not applying, and support saying the withdrawal needed internal confirmation instead of being instant. The welcome bonus issue never looked clean. Housebets advertised a 100% Welcome Bonus up to $1,000 on first deposit; I deposited, contacted support, and the bonus did not apply. Then support effectively turned a first-deposit bonus into a second-deposit workaround because the first one had not been applied properly. On 31 January I came back after another deposit and told them the bonus still had not been applied, even though I had already followed support’s instructions. Edward replied that he had “forwarded” the concern to the team. The same 100% welcome bonus was still being advertised in March. By April, the rewards system was already showing serious problems. I had the weekly slider at 100% lossback and told support I had lost money but the weekly did not appear. Jacky said the weekly was generated every Thursday at 00:01 UTC and gave actual internal figures: GGR $6,250, Total Bonus $6,083.99, NGR $168.31. So Housebets clearly had internal calculations when it wanted to explain why something might not pay. But when I later asked for full calculations, those same numbers suddenly became impossible to produce. Then on 18–19 April, the rewards page was bugged and would not let me claim. Support could see a pending weekly bonus of $717.37, but I could not claim it from the UI. Tee said it had been forwarded to the relevant department. That $717.37 later appears in the bonus ledger as Rakeback (20 Apr) 717.37089061, so I am not saying that specific one stayed unpaid forever. The point is worse: already in April, support could see a pending weekly reward while the player-facing reward page did not work. For a casino built around rewards, that is not a small bug. That is the product. In May, the UI and account data kept failing basic trust checks. On 8 May, I deposited 400 USDT; support said it had been credited, but I could not see it, and the proposed fix was to log out, clear cookies and cache. On 16 May, I asked why total deposits and withdrawals had disappeared from the menu; support said the platform was “in continuous evolution.” On 17 May, I asked for my total deposits and withdrawals, and support said they did not have direct access to that consolidated summary and would email it. That full official ledger did not arrive. So when Housebets later defends itself with UI screenshots, remember: this was the same UI where deposits could be credited but invisible, totals disappeared, rewards pages bugged, and support could not access consolidated account totals. Withdrawals were also not what was advertised. On 16 May, I asked why a crypto withdrawal was pending if withdrawals were supposed to be instant. Tee answered: “A few withdrawals require manual approval,” then added, “Our withdrawals are typically instant but…” That matters because a few days later the withdrawal delay became real damage. On 25 May, I told support before a match that I needed the funds to place a time-sensitive bet on another site in less than 20 minutes. I explained I wanted to bet around 60k at odds of 2.55. The withdrawal did not arrive in time. Later I told them the bet won and that I missed around 90k in profit because Housebets took more than two hours despite being warned before the match started. Jacky said he would raise the compensation case to the VIP team. Nobody resolved it. This was not one delayed withdrawal either. In my formal complaint I reconstructed several withdrawal delays: 23 May 02:55 → 08:03, around 5h08m; 25 May 03:05 → 08:09, around 5h04m; 17 May 03:54 → 08:02, around 4h08m; 18 May 04:46 → 08:11, around 3h25m; 16 May 05:23 → 08:12, around 2h49m. That is not “instant withdrawal.” And if later marketing says withdrawals are much faster now, the obvious question is: if this was the faster version, what did slow look like? The Provably Fair / Tequity side was another major issue. On 17 May I asked support how to verify an old Blackjack round. I did not ask for a generic explanation of Provably Fair; I asked where I could see the server seed, client seed, nonce and result for previous games. Support sent me to bet history, mentioned RTP, gave a generic PF explanation and showed the current Dice seed screen. When I said that did not let me verify previous games, they told me to clear cookies/cache. After doing that, I saw a new client seed and nonce 1 even though I had not played with that seed pair. I asked if Housebets changes seeds on every login. Support could not answer and told me to contact VIP. That seed/session behaviour is important. I later recorded video evidence around the seed changing after clearing cookies/cache and asked for the exact mapping: Housebets account ID → Tequity/provider player ID → session/currency context → seed pair → server seed hash → revealed server seed → client seed → nonce/cursor → raw outcome → final result. Housebets cannot sell Provably Fair if the player cannot verify historical bets, and “contact VIP” is not a verification algorithm. On 24 May, I asked for raw verification data for a specific Tequity Blackjack round: Round ID e1648d60-0da1-4433-a5ab-9ae39f5302e3, Blackjack, Tequity, bet amount 11,346 USDT, client seed O3YBZF7LBu, server seed hash starting 712875.... I asked for revealed server seed, nonce, full result JSON, card draw order and verification algorithm. I also asked about an apparent duplicate-card/deck question. Tee replied: “I don’t have the answers to your questions right now, but I’m forwarding your request to the relevant department.” That same day, I asked for a full audit of six Dice bets of 11,400 USDT each, total 68,400 USDT. I requested bet IDs, provider round IDs, roll results, seed data, balance ledger, request/session logs, security logs, retry flags, provider records and a full technical reconciliation. Tee replied: “I will forward this to the relevant department.” So when I asked for raw data, the answer was not data. It was forwarding. Again. There were also many large loss clusters that required reconciliation because of those unresolved PF, Tequity, category, RTP and session questions. In my complaint I listed clusters such as 25 May 02:17–02:54 Blackjack around 169,932 USDT; 16 May 12:31–13:26 Dice around 90,571.92 USDT; 26 May 02:48–03:58 Mines around 89,199 USDT; 24 May 06:20–06:21 Dice at 68,400 USDT; 26 May 00:11–01:41 Blackjack around 59,910 USDT; 25 May 22:51–22:59 Dice around 59,576 USDT; and several more between 40k and 56k. I am not saying every losing cluster proves manipulation by itself. I am saying that when PF mapping, provider logs, RTP/HE, category mapping and seed/session behaviour are unresolved, these sequences need a real reconciliation. The leaderboard is where the story becomes very hard for Housebets to explain. Around 19–20 May, two new accounts, elmourabut and lucasmartirini, appeared and started climbing every day at a vertiginous pace. Not normal slow leaderboard growth. Not a casual player building volume over time. They were created around that period and then started rising with huge wagering in a way that looked extremely unnatural for brand new accounts. By 29 May, I was first on both weekly and monthly leaderboards, and those two accounts were directly behind me with huge volume. In the monthly leaderboard screenshots, I was around $3.33M wagered, while elmourabut was around $1.29M and lucasmartirini around $1.08M. In the weekly leaderboard, I was around $1.096M, while those two accounts were around $635k and $578k. They were not normal accounts sitting at the bottom; they were directly behind me, applying pressure. In my formal complaint I recorded that elmourabut joined on 19 May and lucasmartirini on 20 May, that they showed zero visible withdrawals, large deposits/wagering and significant card-game volume, and I asked Housebets to confirm they were not staff, test, QA, admin, house-controlled, affiliate-controlled, internally funded, promotional, bonus-only or multi-account related accounts. This matters because a leaderboard is not passive. It is gamification. It makes players defend rank. When two new accounts appear behind you with hundreds of thousands or more than a million in volume, you are pressured to keep wagering. In my case, the disputed deposit sequence from 25 May 22:23 to 26 May 02:09 totals 91,168.375326 USDT. That sequence begins with 1,000.00 at 22:23 and continues with repeated deposits until 2,879.148969 at 02:09. The video later shows why those dates matter: there were deposits coming in, no gameplay withdrawal offsetting the sequence, a balance basically at zero, and later a leaderboard prize shown as P/L. I formally asked Housebets to confirm those two leaderboard accounts were real and eligible, and also to preserve wager logs, transaction records, balance adjustment logs, account flags, leaderboard calculation snapshots, support ticket logs, Telegram/email records and internal notes. Edward said he forwarded the request. In the same thread, he added that they were “working on fixing an issue regarding the weekly bonuses,” and then said the weekly countdown was “not currently on Thursday evenings.” So the leaderboard issue and the weekly bonus issue are linked in time and support context. After that, Housebets confirmed by email that elmourabut and lucasmartirini were “legitimate and eligible accounts.” That email is the trap door. If they were legitimate and eligible, they should have remained in the leaderboard with their volume. If they were not, Housebets should never have confirmed them as legitimate and eligible. After that confirmation, the accounts disappeared from the leaderboard or stopped appearing in the positions their previous wagering required. I went back to support on 30 May and wrote: “There has been a material post-confirmation leaderboard change involving two accounts that Housebets had already confirmed as legitimate and eligible. I need the exact reason, timestamp, logs, and recalculation basis.” Edward said the matter was flagged and that I could expect a prompt response. I am still waiting for the actual explanation. Why did they disappear? My read is simple: because every hour that passed, there was more evidence around those accounts. They had been created around the same period, they were climbing at a speed that looked anything but human, they showed no visible withdrawals in the data I could see and reported, they appeared to be generating huge volume in unclear game categories, and the games/categories tied to that volume did not even make sense from the player-facing UI. When I started asking what they were actually playing, what Card meant, whether the volume was Tequity / UnOriginals / House Games, what RTP and house edge applied, and where the logs were, the questions became uncomfortable. Keeping those accounts visible became harder than removing them. So they disappeared. The game category issue made the leaderboard even more suspicious. On 30 May, I asked support why my own stats showed almost all my volume under Slots / Tragamonedas when I did not play real slots. I told them: “i dont play 3$ in unoriginals,” “i played all 3M in unoriginals,” and “ive never play slots.” I asked what “Card” was, where that game was, what RTP and house edge it had. Monica said Card was mainly Blackjack, Baccarat and Poker variants. Marcus later said the team was investigating why it showed that I mostly played slots when I had not. He could not give the exact game, RTP, HE, provider, category mapping or contribution logic. That matters because those same unclear categories were connected to leaderboard volume. If the site cannot clearly explain whether volume is Slots, Card, UnOriginals, House Games, Blackjack, Baccarat, Always 9 Baccarat or Tequity, then the leaderboard is not auditable for the player. I even asked which UnOriginals those two accounts were playing, and support told me to look at Live Bets. That is not an answer. I was not asking for gossip; I was asking what exact games generated leaderboard volume, what RTP/HE applied and whether that volume was eligible. There is also an earlier leaderboard-related precedent: Porchy had already told me in February that I would lose leaderboard places if I did not rename, because too many people were messaging support saying the site was not being fair due to my name and it “doesn’t make us look good.” That matters because it suggests leaderboard positioning was not treated as a sacred, untouchable system when public perception was involved. If leaderboard positions can be threatened for image reasons, then later claims that everything is purely automatic deserve scrutiny. Then Porchy made the leaderboard situation worse. Instead of producing logs or snapshots, he later said the leaderboard had “abusers” on it, that they were removed to help other players, and that it never affected me. Later he said they paid every single person, “even these abusers,” then called me “begging for money.” That creates a direct contradiction: Housebets confirmed the accounts as legitimate and eligible, then Porchy referred to leaderboard “abusers.” If they were abusers, why were they confirmed as legitimate and eligible? If they were eligible, why did they disappear? If they never affected me, where are the historical snapshots proving that? Once those accounts disappeared, Housebets paid the leaderboard prizes. On 1 June, the bonus ledger shows two Leaderboard entries: 5,007.46111706 and 1,001.49222341, totaling 6,008.95334047. That part was paid. But then Act Two started: the weekly and monthly rewards did not appear as separate ledger entries. The same bonus ledger shows those two 1 June entries as Leaderboard only, not Monthly Bonus, not Weekly Reload, not Lossback. The weekly timeline is a mess. On 28 May, the dashboard / UI said the weekly bonus was claimable every Thursday at 00:01 UTC, and the monthly was available on the 1st at 00:01 UTC. That same night I told support the weekly had shown as available, then reset to 6 days without paying. Later I sent screenshots and wrote: “1M wagered and 0.2$.” Jacky said he had raised the issue to the technical team. So the weekly failure was reported live, not reconstructed after the fact. The next day, 29 May, Edward said they were fixing an issue regarding weekly bonuses and that the weekly countdown was “not currently on Thursday evenings.” Then on 1 June, Spencer said the May weekly bonuses were 7th, 14th, 21st, and then due to migration the weekly moved to Monday, so there was one on the 25th on the new platform. He also said the 25 May weekly covered gameplay from 21–24 May, and that tech was looking at that plus the monthly bonus. The ledger does show a 25 May 02:10 Rakeback entry of 1,996.08334791, which likely corresponds to that 21–24 May weekly. But my major loss sequence starts about 20 hours later, on 25 May at 22:23, and continues until 26 May at 02:09. So the 25 May weekly cannot cover those losses. If weekly was still Thursday, the 25/26 losses should have been in the 28 May weekly. But the bonus ledger on 28 May shows only two tiny Rakeback entries, 0.28373945 and 0.00280958. If weekly moved to Monday because of migration, those losses should have appeared in the next weekly after 25 May. But on 1 June the ledger only shows Leaderboard entries. Then the final video shows the next Weekly Reload reaching zero, paying nothing and resetting to 6d 23h. So the same loss sequence appears to fall into no paid weekly cycle. The 4 June support conversation makes this even more ridiculous. After I recorded the weekly reset video, I asked support a very simple question: what were the last weekly dates/cycles? The dashboard / support flow again said weekly bonuses are claimable every Thursday at 00:01 UTC. Jacky confirmed: “Weekly bonuses can be claimed every Thursday at 00:01 UTC in the Rewards tab,” and added that if not claimed by the following Wednesday at 23:59 UTC, it expires. But when I asked for the exact last four dates, Jacky said he had to check with the relevant department. When I pressed again, he said, “Sorry, As I am only a CS, Let me raise your concerns to relevant department.” I asked whether support did not have the information or simply could not answer. He replied: “Do you have any other concerns?” They use weekly cycles to decide whether to pay, but support cannot explain the weekly cycle. The monthly is missing too. The dashboard / UI said the monthly bonus is based on activity and VIP level from the previous month and is available on the 1st at 00:01 UTC. In May I had more than 3,258,023.0829 wagered according to the formal complaint data. I also have proof/video that the monthly slider was set to 50/50. On 1 June, Spencer first told me I had claimed the Monthly Bonus at 1:12am BST around the same time as the monthly leaderboard reward. I immediately said I only received leaderboard prizes. Then Spencer changed the answer: “Our tech team are still actively working on issues regarding the monthly bonuses.” So first the monthly was claimed, then tech was still fixing it. The ledger still shows no Monthly Bonus entry. Housebets then seems to rely on “up overall” as a defence. But the video and ledger show why that does not work. My weekly/monthly profile later showed around +6,008 P/L with 0 deposits, 0 wagered and around 6,008 in bonuses. That number matches exactly the two 1 June Leaderboard payments. So the UI is showing leaderboard rewards as P/L. Then support used “up overall” to say I was not eligible for weekly lossback. That is not a clean lossback calculation. That is using a leaderboard reward as apparent profit to deny a lossback that should be based on actual eligible losses. There were also smaller reward-confusion issues along the way. On 22 May I asked for all pending bonuses,weekly, monthly, rakeback, level-up, anything, and support said the internal team would manually verify whether everything had been credited correctly and email me. On 24 May, I asked about level-up rewards because the reward looked like $3,500 for Pearl; support clarified it was $3,500 total across all Pearl levels, $500 per level. These are not the core issues, but they are part of the same pattern: rewards marketing, unclear UI, manual verification, emails that do not arrive, and players having to chase basic explanations. Then there is the migration. On 25 May, after the delayed withdrawal, missing VIP contact and unresolved issues, support told me my account would be moved to the new platform and that this upgrade would offer a better withdrawal process and fix many issues. Before that migration, I explicitly requested that no account data, internal data, logs, balance history, bonus history, bet history, provider records or pending issues be deleted. The response: “Your request has been relayed to the relevant department.” Again, forwarding. But if the old data is safe, Housebets should provide the old leaderboard snapshots, old weekly states, old bonus logs, old Tequity mapping and old withdrawal approval logs. The founder response did not fix anything. When Porchy finally engaged, he did not provide the records. He framed the settlement request as “so you want $100,000?” and asked whether I needed it or else I was going to post on X. I had already made clear this was not money for silence; I asked for logs, snapshots, withdrawal records, calculations and a counter-calculation if Housebets disagreed. He later referred to “abusers,” told me I was “up overall,” said “You are begging for money,” and suggested I “just do this to casinos.” Still no ledger. Still no weekly calculation. Still no monthly entry. Still no PF/Tequity mapping. Still no leaderboard snapshots. Another player also contacted me with screenshots pointing to similar categories of issues: private deals, leaderboard payout disputes, migration/account merge problems, missing history and a tiny monthly bonus despite claimed losses. I am not using that player’s case as the foundation of my claim without his full ledger, but it matters because it suggests the same type of opacity may not be isolated: private VIP/reward deals, leaderboard eligibility, monthly bonus calculations, migration and unclear history. If Housebets has private deals that affect leaderboard eligibility or rewards, it must explain how those deals interact with public leaderboards. So the overall picture is this: Housebets sold a public leaderboard and rewards system that pressured real wagering. Two new accounts appeared directly behind me with huge volume, were confirmed as legitimate and eligible, then disappeared after I asked for logs and questioned game categories. Housebets could not explain the exact games, RTP, house edge or category mapping behind the volume. The accounts were later framed by Porchy as “abusers,” contradicting the earlier eligibility confirmation. Once Housebets paid me the leaderboard prizes, those prizes were shown as P/L, and that contaminated P/L was then used to claim I was “up overall” and not eligible for lossback. At the same time, my real 25 May 22:23 → 26 May 02:09 loss sequence of 91,168.375326 USDT appears in no clean weekly cycle. The 25 May weekly covered 21–24 May according to Spencer, so it cannot cover that loss sequence. The 28 May weekly showed only tiny Rakeback entries and was already reported as broken. The 1 June ledger shows only Leaderboard entries. The later video shows Weekly Reload reaching zero, paying nothing and resetting. And when I ask support for the exact weekly calendar, they cannot answer and send it to the relevant department. The monthly is the same story. The dashboard / UI says it is based on activity and VIP. I had more than 3.25M wagered in May. Spencer first says I claimed it, then says tech is still working on monthly bonuses. The ledger shows no Monthly Bonus. If Housebets says I was not eligible, they need to show the formula, slider history, cycle, GGR/NGR, eligible loss/activity, deductions and ledger result. If they cannot, “not eligible” is just another label. And this opens another can of worms: Tequity / provider configuration. Housebets cannot hide behind “the provider” whenever something goes wrong. The player does not deposit with Tequity. The player does not withdraw from Tequity. The player does not speak to Tequity support. The player does not compete in a Tequity leaderboard. The player plays on Housebets, with a Housebets wallet, Housebets UI, Housebets rewards, Housebets leaderboard and Housebets support. 1/2

Dr. W

20,491 次观看 • 3 个月前

Moneytaur study blueprint 🗺️ The process I used to go from not knowing what an order block is to pulling cash from the crypto markets in under 6 months using 🎯 Master concepts. Proof of performance, past 120 days👇 Start date: 09/03/2025 Requirements: - A PC/laptop - Wifi - A basic understanding of trading. ( What candlesticks are, how to actually place trades , etc ) - A free mind - Time or the ability to free up time. Starting: - Structure and routine - Stick to that routine + Pre mortem plan. - Notion / Obsidian setup. The first thing you need to create is a clear routine moulded around how you intend to approach this very large and complex task. This will not be linear and you will naturally adapt it as you progress but especially in the beginning some resemblance of structure each day is vital. This is an individual process but it is important to understand from the beginning that this will require a majority of your free time assuming you work a full time Job or study as a student. For me in the beginning this looked like: - Wake up at 6:30. - Shower - Study/work for 1h 45m before leaving for work. - 09:00 -> 17:00 work - 17:30 Exercise / Train - Eat - 19:00 resume study/work - 22:30 Start to wind down and get ready to sleep. It changed several times over the months and especially now I am full time but this is irrelevant, the only thing that matters is sticking with what you choose. Whatever your own routine may look like, it is important to understand it will inevitably require sacrifice. --- The next thing once you have established a draft framework of your routine is ensuring you will actually stick to that routine. Something I implemented which I found particularly beneficial was the concept of a Pre-Mortem plan. This involves creating several scenarios of a future in which you have failed and working backwards from each of these to find where it went wrong. Here is a video which explains it fully: When I did this I came up with 3 scenarios as well as prevention and cure for each. In the 6 months that followed each scenario presented at some point but I was able to catch them early due to having done this. The last thing is to not over complicate this, don't hyper focus on systems and loose momentum optimizing each detail. Just ensure you do the fucking work. I was a little guilty of the above at times, trying to craft the perfect routine. In reality the person who just gets up, drinks too much coffee and works his ass off out performs the workflow perfectionist who visualizes and repeats affirmations, any day of the week. --- Next you need somewhere to store your notes, journal your trades and build your knowledge. For me this was Obsidian but I have also used Notion before and it is an equally viable option. Whichever one of these you choose be warned you will inevitably want to bang your head against a wall trying to use them for the first few days, but they will both click pretty quick and are 100% better options the word document or paper alternative. Here is my full obsidian setup tutorial: Here is a link to MisterPA 's notion Journal: Here is how I create "Meta-Notes" using obsidian: The process: - How I did it. - How I would do it if doing it again. Now I did things the "hard way" and manually worked my way back through each of MT's tweets starting in 2021, reading every one and logging those that I felt where relevant. You can see in my first post: the very first system I used to do this. I quickly adapted though after about a week and focused less on just logging each relevant tweet but trying to find and focusing on those which contained the most information. There where a lot of charts I looked at then skipped over because especially at the start of his timeline they contained little useful information and my time was better spent finding those where there was something to decode. Now this does not mean skip out on "work" just use your time efficiently. -- If however if I was to start from the beginning again with the goal of levelling up technical understanding as quickly as possible I would take a different approach. To start with I would familiarise myself with all relevant SMC concepts, I have linked the best free recourses for this below 👇 CryptoChase beginner friendly index: Barncore's "The Moneytaur Way" series: Gian's Trading bootcamp playlist: Following this I would then work through all of Taur's subscription posts working backwards, recreating his charts and taking notes on his logic. The subscription feed has the highest value density and least noise. Video example of my notes from his subscription posts 👇: --- Okay so now once you have a basic understanding of concepts and can re-recreate them on charts of your own it is time to put this in to practice. The next step is vigorous backtesting, you can use the trading view tool but I think trade Zella offers a more use friendly option if you pay for the subscription. Especially as it allows you to change timeframes without skipping ahead to candle close time of the timeframe you change too ( like Trading view does ) *my only note would be that their LTF/Micro TF data feed with be different to brokerage charts you will use on Trading view, to start with though you should not be going low enough that this is an issue. When you backtest in this context, treat it like real trading. That means journal and logging like you would if real cash was on the line. Take time, do not rush and focus on quality. Stick to BTC, ETH, Major FX pairs or indices as these assets are less reliant on confluence, backtesting a shitcoin is near useless as whether levels work or not will be highly dependent on Majors PA. Go on HTF, scroll back a couple years and try not too look at chart while doing so and then begin. Start with HTF analysis and work down to 2H or wherever you feel comfortable, chart it fully and then identify setups. Make rough notes / plans and then press play, execute the setups as they hit, log and journal trade management as well as observations and key notes. It is very important to not cheat when you do this, do not skip back and adjust your stoploss because it hit by 0.1%, do not skip back and adjust plan because you missed a block and your TP got frontrun. Instead these are the things you journal, embrace these mistakes because they are the cheapest mistakes you are going to make. Grind this, do it for hours, put some music on and enjoy. To start with focus on HTF's, as you get better and start netting $ on paper you can drop the timeframes and increase the difficulty. HTF = Normal, MTF = Medium, LTF = Hard. Even if you do not intend to day trade, learning how to read the lower TF's that force you to think faster, harder and prepare you for lower win rates / loss streaks can greatly improve your ability on higher TF's. While you are doing this as you start to have concepts click you now want to build up your real trading experience, take a sum of money that you care about but will be okay loosing and dedicate this to live trading. Start taking real trades and expect net losses in the beginning. This is where you will make you 2nd cheapest mistakes. This is also where you can begin to learn about your psychology. You may encounter some elements already in backtesting but the real market is where true colours really start to show. Mental issues are inevitable and part of the game, get used to them and start working to identify and fix them. Reading and applying books like Trading in the Zone and Mental Game of Trading are important and will help a lot but there is no easy fix, for some stuff you I believe you just have to get used to it and it goes away with experience. Losses suck at the beginning but after you loose 100 times you starting getting pretty numb to it, same goes for the winners. To accelerate the learning process, build connections and get advice there is also always the option of private groups, while I never personally chose this route and committed to learning everything through my own endeavours there is no denying that having nearly all the information you need structured and compiled in one place is valuable and can save time. Beyond this having access to real time thoughts and opinions of profitable traders can accelerate performance, however it carries the risk of being a double edged sword if not used properly, if relying on it like a crutch and using it as a substitute for real work you will not succeed. With that said if you take it for what it is, a learning opportunity then I believe it can be very beneficial. I am not a member of, nor affiliated with any paid group. There are now many options available within the community, all run by different people with different styles, tailored to different needs. If I was to make a recommendation though, as a non-member, it would be Albert & Co's 618'ers simply due to the diversity in styles of the traders running it and results I have seen from members I know personally. It is important that as you start to trade with real capital you reduce noise in your social feeds or eliminate it all together. You do not need 5 different opinions, you also do not need 2 people telling you the same thing in their own way so you feel re-assured. What you do need is to develop your independent thinking as a trader and be comfortable making different decisions to others, even traders ahead of yourself if it fits with your system or understanding of market. Taur here is perhaps an exception as this is who you are learning from but down the line a real test of your own ability and independence will be being able to stick with your own plan even when it differs from his. Don't get me wrong, counter trading him is retarded but you must learn to adapt his gift to your own style. This will make sense at some point. The next stage is taking your understanding of specific concepts to higher level as you simultaneously snowball experience. Look back through your journal and review where you lost money and made money, do not over extrapolate from a small sample but start to take notes and observe if trends in performance emerge. This is the beginning of the transition to self reliance, you now understand the strategy but must learn for yourself when and where it works. Here you can also learn more nuanced secondary concepts such as VSA, orderflow etc and add these to your game where appropriate. Do NOT get lost in the sauce though and remember mastery of basics is key. IMO a big focus should be understanding correlation thoroughly but especially on HTF's this is the most important thing and what triggers the majority of large swings where most of your cash will be made and losses recovered. Some people will disagree with me here but IMO you should also not be *focusing* on Odd TF's. These are secondary at best and most people overweight their significance leading to avoidable losses while wondering why price did not care about their 327minute Breaker Block which they think is the key to the market. Study Taurs feed and take note of how he mostly uses: 3M, 1M, 3W, 2W, 1W, 5D, 4D, 3D, 2D, 1D, 12H, 8H, 6H, 4H, 2H, 1H, 30m, 15m + micro time frames. The only thing left is time and repetition, you must show up each day and really do this, for months. Maybe you start to see result's, you catch your first key swing and where able to trade where others froze. Congratulations. Learn from these winners and repeat the actions. Find what assets work best for you, find your style, refine and grow. --- The last thing I will include is a short list of tools or links that can be helpful. - Trading view tutorial: - Dictionary: - Market news Calendar: --- Thank you too all those who have read this, I hope this has been helpful for the beginners who want to start but are just not sure how. 🫶 Don't just bookmark this and move on, start 🙃
0:38

Moneytaur study blueprint 🗺️ The process I used to go from not knowing what an order block is to pulling cash from the crypto markets in under 6 months using 🎯 Master concepts. Proof of performance, past 120 days👇 Start date: 09/03/2025 Requirements: - A PC/laptop - Wifi - A basic understanding of trading. ( What candlesticks are, how to actually place trades , etc ) - A free mind - Time or the ability to free up time. Starting: - Structure and routine - Stick to that routine + Pre mortem plan. - Notion / Obsidian setup. The first thing you need to create is a clear routine moulded around how you intend to approach this very large and complex task. This will not be linear and you will naturally adapt it as you progress but especially in the beginning some resemblance of structure each day is vital. This is an individual process but it is important to understand from the beginning that this will require a majority of your free time assuming you work a full time Job or study as a student. For me in the beginning this looked like: - Wake up at 6:30. - Shower - Study/work for 1h 45m before leaving for work. - 09:00 -> 17:00 work - 17:30 Exercise / Train - Eat - 19:00 resume study/work - 22:30 Start to wind down and get ready to sleep. It changed several times over the months and especially now I am full time but this is irrelevant, the only thing that matters is sticking with what you choose. Whatever your own routine may look like, it is important to understand it will inevitably require sacrifice. --- The next thing once you have established a draft framework of your routine is ensuring you will actually stick to that routine. Something I implemented which I found particularly beneficial was the concept of a Pre-Mortem plan. This involves creating several scenarios of a future in which you have failed and working backwards from each of these to find where it went wrong. Here is a video which explains it fully: When I did this I came up with 3 scenarios as well as prevention and cure for each. In the 6 months that followed each scenario presented at some point but I was able to catch them early due to having done this. The last thing is to not over complicate this, don't hyper focus on systems and loose momentum optimizing each detail. Just ensure you do the fucking work. I was a little guilty of the above at times, trying to craft the perfect routine. In reality the person who just gets up, drinks too much coffee and works his ass off out performs the workflow perfectionist who visualizes and repeats affirmations, any day of the week. --- Next you need somewhere to store your notes, journal your trades and build your knowledge. For me this was Obsidian but I have also used Notion before and it is an equally viable option. Whichever one of these you choose be warned you will inevitably want to bang your head against a wall trying to use them for the first few days, but they will both click pretty quick and are 100% better options the word document or paper alternative. Here is my full obsidian setup tutorial: Here is a link to MisterPA 's notion Journal: Here is how I create "Meta-Notes" using obsidian: The process: - How I did it. - How I would do it if doing it again. Now I did things the "hard way" and manually worked my way back through each of MT's tweets starting in 2021, reading every one and logging those that I felt where relevant. You can see in my first post: the very first system I used to do this. I quickly adapted though after about a week and focused less on just logging each relevant tweet but trying to find and focusing on those which contained the most information. There where a lot of charts I looked at then skipped over because especially at the start of his timeline they contained little useful information and my time was better spent finding those where there was something to decode. Now this does not mean skip out on "work" just use your time efficiently. -- If however if I was to start from the beginning again with the goal of levelling up technical understanding as quickly as possible I would take a different approach. To start with I would familiarise myself with all relevant SMC concepts, I have linked the best free recourses for this below 👇 CryptoChase beginner friendly index: Barncore's "The Moneytaur Way" series: Gian's Trading bootcamp playlist: Following this I would then work through all of Taur's subscription posts working backwards, recreating his charts and taking notes on his logic. The subscription feed has the highest value density and least noise. Video example of my notes from his subscription posts 👇: --- Okay so now once you have a basic understanding of concepts and can re-recreate them on charts of your own it is time to put this in to practice. The next step is vigorous backtesting, you can use the trading view tool but I think trade Zella offers a more use friendly option if you pay for the subscription. Especially as it allows you to change timeframes without skipping ahead to candle close time of the timeframe you change too ( like Trading view does ) *my only note would be that their LTF/Micro TF data feed with be different to brokerage charts you will use on Trading view, to start with though you should not be going low enough that this is an issue. When you backtest in this context, treat it like real trading. That means journal and logging like you would if real cash was on the line. Take time, do not rush and focus on quality. Stick to BTC, ETH, Major FX pairs or indices as these assets are less reliant on confluence, backtesting a shitcoin is near useless as whether levels work or not will be highly dependent on Majors PA. Go on HTF, scroll back a couple years and try not too look at chart while doing so and then begin. Start with HTF analysis and work down to 2H or wherever you feel comfortable, chart it fully and then identify setups. Make rough notes / plans and then press play, execute the setups as they hit, log and journal trade management as well as observations and key notes. It is very important to not cheat when you do this, do not skip back and adjust your stoploss because it hit by 0.1%, do not skip back and adjust plan because you missed a block and your TP got frontrun. Instead these are the things you journal, embrace these mistakes because they are the cheapest mistakes you are going to make. Grind this, do it for hours, put some music on and enjoy. To start with focus on HTF's, as you get better and start netting $ on paper you can drop the timeframes and increase the difficulty. HTF = Normal, MTF = Medium, LTF = Hard. Even if you do not intend to day trade, learning how to read the lower TF's that force you to think faster, harder and prepare you for lower win rates / loss streaks can greatly improve your ability on higher TF's. While you are doing this as you start to have concepts click you now want to build up your real trading experience, take a sum of money that you care about but will be okay loosing and dedicate this to live trading. Start taking real trades and expect net losses in the beginning. This is where you will make you 2nd cheapest mistakes. This is also where you can begin to learn about your psychology. You may encounter some elements already in backtesting but the real market is where true colours really start to show. Mental issues are inevitable and part of the game, get used to them and start working to identify and fix them. Reading and applying books like Trading in the Zone and Mental Game of Trading are important and will help a lot but there is no easy fix, for some stuff you I believe you just have to get used to it and it goes away with experience. Losses suck at the beginning but after you loose 100 times you starting getting pretty numb to it, same goes for the winners. To accelerate the learning process, build connections and get advice there is also always the option of private groups, while I never personally chose this route and committed to learning everything through my own endeavours there is no denying that having nearly all the information you need structured and compiled in one place is valuable and can save time. Beyond this having access to real time thoughts and opinions of profitable traders can accelerate performance, however it carries the risk of being a double edged sword if not used properly, if relying on it like a crutch and using it as a substitute for real work you will not succeed. With that said if you take it for what it is, a learning opportunity then I believe it can be very beneficial. I am not a member of, nor affiliated with any paid group. There are now many options available within the community, all run by different people with different styles, tailored to different needs. If I was to make a recommendation though, as a non-member, it would be Albert & Co's 618'ers simply due to the diversity in styles of the traders running it and results I have seen from members I know personally. It is important that as you start to trade with real capital you reduce noise in your social feeds or eliminate it all together. You do not need 5 different opinions, you also do not need 2 people telling you the same thing in their own way so you feel re-assured. What you do need is to develop your independent thinking as a trader and be comfortable making different decisions to others, even traders ahead of yourself if it fits with your system or understanding of market. Taur here is perhaps an exception as this is who you are learning from but down the line a real test of your own ability and independence will be being able to stick with your own plan even when it differs from his. Don't get me wrong, counter trading him is retarded but you must learn to adapt his gift to your own style. This will make sense at some point. The next stage is taking your understanding of specific concepts to higher level as you simultaneously snowball experience. Look back through your journal and review where you lost money and made money, do not over extrapolate from a small sample but start to take notes and observe if trends in performance emerge. This is the beginning of the transition to self reliance, you now understand the strategy but must learn for yourself when and where it works. Here you can also learn more nuanced secondary concepts such as VSA, orderflow etc and add these to your game where appropriate. Do NOT get lost in the sauce though and remember mastery of basics is key. IMO a big focus should be understanding correlation thoroughly but especially on HTF's this is the most important thing and what triggers the majority of large swings where most of your cash will be made and losses recovered. Some people will disagree with me here but IMO you should also not be *focusing* on Odd TF's. These are secondary at best and most people overweight their significance leading to avoidable losses while wondering why price did not care about their 327minute Breaker Block which they think is the key to the market. Study Taurs feed and take note of how he mostly uses: 3M, 1M, 3W, 2W, 1W, 5D, 4D, 3D, 2D, 1D, 12H, 8H, 6H, 4H, 2H, 1H, 30m, 15m + micro time frames. The only thing left is time and repetition, you must show up each day and really do this, for months. Maybe you start to see result's, you catch your first key swing and where able to trade where others froze. Congratulations. Learn from these winners and repeat the actions. Find what assets work best for you, find your style, refine and grow. --- The last thing I will include is a short list of tools or links that can be helpful. - Trading view tutorial: - Dictionary: - Market news Calendar: --- Thank you too all those who have read this, I hope this has been helpful for the beginners who want to start but are just not sure how. 🫶 Don't just bookmark this and move on, start 🙃

Ace

45,538 次观看 • 9 个月前