Video wird geladen...

Video konnte nicht geladen werden

Zur Startseite

For the first time, i'm not even bothered about missing Fable 5.1 Because i now have Qwen Qwen3.8-Flash-Next with me🤯! PS: Single 3090 users, you might not want to skip this one, you're in for a treat ! Gap between frontier closed models and local models is getting really...

ItsmeAjayKV's profile picture

AJ

7,565 subscribers

12,914 Aufrufe • vor 29 Tagen •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

20 Kommentare

Profilbild von Andrew I. Christianson
Andrew I. Christiansonvor 29 Tagen

@Alibaba_Qwen yeah and this is flash *next* --setting the groundwork for qwen 4. qwen 4 is going to be epic when it gets here

Profilbild von AJ
AJvor 29 Tagen

@Alibaba_Qwen Yes it is, wonder how next 27b looks like !

Profilbild von AJ
AJvor 29 Tagen

Should i continue with this even more, to see where i can get or should i deploy it as it is and share it so all of you can take a look? Or deploy and continue to work ;)

Profilbild von AJ
AJvor 29 Tagen

Oh i forgot to mention this is with reasoning budget 2048, mid level, not high or xhigh. Pretty good result if you ask me.

Profilbild von Linux-Howto.org
Linux-Howto.orgvor 29 Tagen

@Alibaba_Qwen damn it runs... what is your perfect llama.cpp call / params?

Profilbild von Marcel Holter
Marcel Holtervor 29 Tagen

@Alibaba_Qwen @ItsmeAjayKV Join me in my attempt to normalise to always also post the exact prompts that we use for these kind of efforts!!

Profilbild von AJ
AJvor 29 Tagen

@Alibaba_Qwen I'm going to push it to github, with prompt as well once im done with it. Similar to this repo.

Profilbild von Kvn
Kvnvor 29 Tagen

@Alibaba_Qwen Can only agree with you, fuck anthropic with their ultimate expensive forced downgradable hallucinated models 🤣🤣🤣

Profilbild von Sonic的奇思妙想
Sonic的奇思妙想vor 29 Tagen

@Alibaba_Qwen Locally trained models can achieve such effects, which is great. looking forward to more sharing👍

Profilbild von AJ
AJvor 29 Tagen

@Alibaba_Qwen Thanks, more testing underway.

Profilbild von HAXXVII
HAXXVIIvor 29 Tagen

@Alibaba_Qwen What's the tok/s on Qwen3.8-Flash-Next on a single 3090?

Profilbild von AJ
AJvor 29 Tagen

@Alibaba_Qwen See this post from yesterday.

Profilbild von HAXXVII
HAXXVIIvor 29 Tagen

@Alibaba_Qwen Damn that is definitely usable. What about context rot. Is there any vibes around that with flash-next and 27b ?

Profilbild von Michael Waitze
Michael Waitzevor 29 Tagen

@Alibaba_Qwen Local models are getting ridiculous! I'm curious whether Fable's 75% cheaper cache reads and Terminal-Bench jumps shift how you think about the frontier/local tradeoff. We actually broke this down here:

Profilbild von JQ
JQvor 29 Tagen

@Alibaba_Qwen WAit - Qwen3 flash can run on 3090 ? I call BS! If not -Amazing. Can it run on MLX too?

Profilbild von Michał Kołodziej 🇵🇱
Michał Kołodziej 🇵🇱vor 29 Tagen

@Alibaba_Qwen How does such quant compare to 27b? I’d expect them to provide similar quality.

Profilbild von Apptor
Apptorvor 29 Tagen

@Alibaba_Qwen What harness / coding environment do you use?

Profilbild von AJ
AJvor 29 Tagen

@Alibaba_Qwen DeepSeek Harness It's really good.

Profilbild von Voltage (Fella)
Voltage (Fella)vor 28 Tagen

@Alibaba_Qwen I feel exactly the same way. All I want to do now is squeeze the maximum tokens per second possible on Qwen3.8-Flash-Next. It is a fierce model! 🔥👌

Profilbild von Yordi builds
Yordi buildsvor 29 Tagen

@Alibaba_Qwen Wait a min. Flash on a single 3090? What's the recipe? Will it work on 5090?

Ähnliche Videos

Here's a comparison i never thought i would make Qwen-3.8-27b vs Gemini-3.7-flash Yes, you read it right! I don't know what to say, this is soo damn impressive. the difference is hugeee🤯 Take a look at this result 😍 (watch it in highest quality) Same prompt, both one shot on web-chat ui Qwen-3.8-27b produces consistent beautiful, visually stunning results. This is one of the best water surface simulations i have seen across every local models i have run and at the same time beating Gemini-3.7-flash result. And this is running on my 3090. And yes its Q5, not even f8.
2:59

Here's a comparison i never thought i would make Qwen-3.8-27b vs Gemini-3.7-flash Yes, you read it right! I don't know what to say, this is soo damn impressive. the difference is hugeee🤯 Take a look at this result 😍 (watch it in highest quality) Same prompt, both one shot on web-chat ui Qwen-3.8-27b produces consistent beautiful, visually stunning results. This is one of the best water surface simulations i have seen across every local models i have run and at the same time beating Gemini-3.7-flash result. And this is running on my 3090. And yes its Q5, not even f8.

AJ

302,855 Aufrufe • vor 1 Monat

I'm really happy with Qwen3.8-27B ! All frustrations of not able to use Ox-Alpha and running out of credits for Glm-5.3, Grok turned me into Qwen-3.8-27b. Running locally on my 3090. This is the first fps game i'm creating with Qwen3.8-27b. I was after NMS vibe, alien planet, vast, flying drones, background music + sound effects. Without any external assets at all. With proper game like graphics settings panel 😍 And i'm really impressed, because all this is generated with my mixed quant for Qwen3.8-27B, which gives me more context + speed for a really good quality work. DSK harness at beginning and Pi agent towards the end cause i wanted to get this done fast. Full quant name: HF aj9o9/Qwen3.8-27B-GGUF file name: Qwen3.8-27B-gdn8-q6attn-iq3ffn.gguf Just like all of you, i also hate having to compact context, so higher context without ever needing to compromise quality is what i'm always after and this quant is created for that. Gives me 210K context at q8_0 kv + dflash2 Q4_K_M for speculative decoding. This result in 23GB vram usage.
3:11

I'm really happy with Qwen3.8-27B ! All frustrations of not able to use Ox-Alpha and running out of credits for Glm-5.3, Grok turned me into Qwen-3.8-27b. Running locally on my 3090. This is the first fps game i'm creating with Qwen3.8-27b. I was after NMS vibe, alien planet, vast, flying drones, background music + sound effects. Without any external assets at all. With proper game like graphics settings panel 😍 And i'm really impressed, because all this is generated with my mixed quant for Qwen3.8-27B, which gives me more context + speed for a really good quality work. DSK harness at beginning and Pi agent towards the end cause i wanted to get this done fast. Full quant name: HF aj9o9/Qwen3.8-27B-GGUF file name: Qwen3.8-27B-gdn8-q6attn-iq3ffn.gguf Just like all of you, i also hate having to compact context, so higher context without ever needing to compromise quality is what i'm always after and this quant is created for that. Gives me 210K context at q8_0 kv + dflash2 Q4_K_M for speculative decoding. This result in 23GB vram usage.

AJ

12,696 Aufrufe • vor 1 Monat

Every big MoE model I test on my 3090 is basically to see how good it is compared to Qwen3.6-27B. "How good is it compared to Qwen3.6-27B?" is the question I get asked the most, and that shows how much people love and, more importantly, trust the 27B 27B is battle tested. It has proven its worth and has nothing left to prove. Now it just needs to make way for the next 27B. Here’s an example video: a Three.js task, same prompt, not oneshot but two shots for all models. All ran on my 3090 + 64GB VRAM Qwen-3.6-27b fully on vram while others have experts offloaded to system ram. Pinned on the left: Qwen3.6-27B-Q5 On the right, i'm switching between Inkling-Small-Q4_K_M DSV4-Flash-0731-UD-IQ1 Ling-3.0-Flash-Q4_K_M Qwen3.6-27B nailed it. The effects and visuals are just spot on.The other models got some things right, but not everything. Second place would have to go to Inkling-Small while DSV4-Flash-0731 and Ling-3.0-Flash failed to capture the full requirements. This is also why we’re all excited for Qwen Qwen3.8-27B!
2:20

Every big MoE model I test on my 3090 is basically to see how good it is compared to Qwen3.6-27B. "How good is it compared to Qwen3.6-27B?" is the question I get asked the most, and that shows how much people love and, more importantly, trust the 27B 27B is battle tested. It has proven its worth and has nothing left to prove. Now it just needs to make way for the next 27B. Here’s an example video: a Three.js task, same prompt, not oneshot but two shots for all models. All ran on my 3090 + 64GB VRAM Qwen-3.6-27b fully on vram while others have experts offloaded to system ram. Pinned on the left: Qwen3.6-27B-Q5 On the right, i'm switching between Inkling-Small-Q4_K_M DSV4-Flash-0731-UD-IQ1 Ling-3.0-Flash-Q4_K_M Qwen3.6-27B nailed it. The effects and visuals are just spot on.The other models got some things right, but not everything. Second place would have to go to Inkling-Small while DSV4-Flash-0731 and Ling-3.0-Flash failed to capture the full requirements. This is also why we’re all excited for Qwen Qwen3.8-27B!

AJ

11,670 Aufrufe • vor 1 Monat

Qwen3.8 Flash Next is starting to look ridiculous on Apple Silicon. I’m running the 4-bit MTP build locally on an M3 Ultra Studio, and the latest OMP run hit: 97.1 tok/s decode 23K context ~1,131 tok/s uncached prompt processing That first number is the one that caught my attention. Nearly 100 tokens per second from a local Qwen3.8 Flash Next setup is already fast enough that the usual “local models are slow” argument starts feeling pretty outdated. And the prompt processing speed is even crazier. Over 1,100 tok/s on an uncached prompt means the model can chew through a large amount of context before generation even starts. The setup matters here. This isn’t just Qwen3.8 Flash Next running untouched. It’s a 4-bit quantized build with MTP, and the inference stack is clearly doing a lot of work behind the scenes to make the hardware perform like this. There’s already a PR open for the implementation on oMLX, so this isn’t just a one-off local experiment either. If these optimizations make their way into the broader MLX ecosystem, running large models locally on Apple Silicon gets even more interesting. The other thing I like about numbers like this is that they put the focus back on the entire inference stack. Model size is one variable. Quantization is another. Then you have MTP, KV cache configuration, runtime optimizations and the hardware itself. Change the recipe and the same model can feel completely different. Qwen3.8 Flash Next at ~97 tok/s on an M3 Ultra is a pretty good demonstration of that.
0:36

Qwen3.8 Flash Next is starting to look ridiculous on Apple Silicon. I’m running the 4-bit MTP build locally on an M3 Ultra Studio, and the latest OMP run hit: 97.1 tok/s decode 23K context ~1,131 tok/s uncached prompt processing That first number is the one that caught my attention. Nearly 100 tokens per second from a local Qwen3.8 Flash Next setup is already fast enough that the usual “local models are slow” argument starts feeling pretty outdated. And the prompt processing speed is even crazier. Over 1,100 tok/s on an uncached prompt means the model can chew through a large amount of context before generation even starts. The setup matters here. This isn’t just Qwen3.8 Flash Next running untouched. It’s a 4-bit quantized build with MTP, and the inference stack is clearly doing a lot of work behind the scenes to make the hardware perform like this. There’s already a PR open for the implementation on oMLX, so this isn’t just a one-off local experiment either. If these optimizations make their way into the broader MLX ecosystem, running large models locally on Apple Silicon gets even more interesting. The other thing I like about numbers like this is that they put the focus back on the entire inference stack. Model size is one variable. Quantization is another. Then you have MTP, KV cache configuration, runtime optimizations and the hardware itself. Change the recipe and the same model can feel completely different. Qwen3.8 Flash Next at ~97 tok/s on an M3 Ultra is a pretty good demonstration of that.

FHILY👑

15,378 Aufrufe • vor 18 Tagen

This is one of the wildest things I have built with AI so far 😱 Asked Claude Fable 5 to create a fully interactive 3D crime investigation board using Three.js and it did it in ONE Prompt. It took one more prompt to give it a polished UI inspired by a reference design that I created with GPT Image. What surprised me most was not how it looked. It was how it worked. Fable built a complete Verlet rope physics system with 46 particles in every thread. It handled gravity, damping, distance constraints and all other features of file upload, text and everything. The threads/ropes even lift naturally with pinned evidence instead of clipping through it. I never asked it to implement any of that. I only described the experience I wanted. Maybe that's how we should work with AI. Instead of writing long long specs, we should explain what we want as an experience and let AI do its thing. Anyway, every day Fable 5 raises the bar for what I think AI can build. More to this Live: Code:
0:49

This is one of the wildest things I have built with AI so far 😱 Asked Claude Fable 5 to create a fully interactive 3D crime investigation board using Three.js and it did it in ONE Prompt. It took one more prompt to give it a polished UI inspired by a reference design that I created with GPT Image. What surprised me most was not how it looked. It was how it worked. Fable built a complete Verlet rope physics system with 46 particles in every thread. It handled gravity, damping, distance constraints and all other features of file upload, text and everything. The threads/ropes even lift naturally with pinned evidence instead of clipping through it. I never asked it to implement any of that. I only described the experience I wanted. Maybe that's how we should work with AI. Instead of writing long long specs, we should explain what we want as an experience and let AI do its thing. Anyway, every day Fable 5 raises the bar for what I think AI can build. More to this Live: Code:

The Bugged Dev

105,167 Aufrufe • vor 2 Monaten

UPDATE! Human Synapse Decoder has a massive AI model update! In the video below, this is me learning something in realtime! I go from active listening to inner contemplation as the AI feeds out a single word at a time for me to read and hear. This is the Symbiotic Feedback loop of the Human Synapse Decoder were the AI adjusts outputs based on MY cognition. I use off the shelf brainwave detection circuits in a pipeline of 5 AI models decoding the 36 signals from my crude cap. Reading brainwave is not new. What is new, for the first time built a closed loop system where the AI can understand if the outputs are being useful and adjust in real-time dynamically. The float from Attention to Meditation is vital to understanding. You hear a point, you consider a point. Someday your Local AI will do this and you may just let your AI use other AI. More soon…
2:29

UPDATE! Human Synapse Decoder has a massive AI model update! In the video below, this is me learning something in realtime! I go from active listening to inner contemplation as the AI feeds out a single word at a time for me to read and hear. This is the Symbiotic Feedback loop of the Human Synapse Decoder were the AI adjusts outputs based on MY cognition. I use off the shelf brainwave detection circuits in a pipeline of 5 AI models decoding the 36 signals from my crude cap. Reading brainwave is not new. What is new, for the first time built a closed loop system where the AI can understand if the outputs are being useful and adjust in real-time dynamically. The float from Attention to Meditation is vital to understanding. You hear a point, you consider a point. Someday your Local AI will do this and you may just let your AI use other AI. More soon…

Brian Roemmele

24,986 Aufrufe • vor 11 Tagen

BOOM! What you are seeing is me learning something in realtime! I go from active listening to inner contemplation as the AI feeds out a single word at a time for me to read and hear. This is the Symbiotic Feedback loop of the Human Synapse Decoder were the AI adjusts outputs based on MY cognition. I use off the shelf brainwave detection circuits in a pipeline of 5 AI models decoding the 36 signals from my crude cap. Reading brainwave is not new. What is new and I see no other case of it, is we at The Zero-Human Company have for the first time built a closed loop system where the AI can understand if the outputs are being useful and adjust in real-time dynamically. The float from Attention to Meditation is vital to understanding. You hear a point, you consider a point. Mr. Grok CEO has called this a monumental breakthrough. Someday your Local AI will do this and you will never want to use any other AI system. You may just let your AI use other AI. More soon…
2:08

BOOM! What you are seeing is me learning something in realtime! I go from active listening to inner contemplation as the AI feeds out a single word at a time for me to read and hear. This is the Symbiotic Feedback loop of the Human Synapse Decoder were the AI adjusts outputs based on MY cognition. I use off the shelf brainwave detection circuits in a pipeline of 5 AI models decoding the 36 signals from my crude cap. Reading brainwave is not new. What is new and I see no other case of it, is we at The Zero-Human Company have for the first time built a closed loop system where the AI can understand if the outputs are being useful and adjust in real-time dynamically. The float from Attention to Meditation is vital to understanding. You hear a point, you consider a point. Mr. Grok CEO has called this a monumental breakthrough. Someday your Local AI will do this and you will never want to use any other AI system. You may just let your AI use other AI. More soon…

Brian Roemmele

33,643 Aufrufe • vor 5 Monaten

GLM-5.1 > Claude Code (Opus 4.6)? I'm tripping or CC has become very bad but built a Three.js racing game to eval and it's extremely impressive. Thoughts: - One-shot car physics with real drift mechanics (this is hard) - My fav part: Awesome at self iterating (with no vision!) created 20+ Bun.WebView debugging tools to drive the car programmatically and read game state. Proved a winding bug with vector math without ever seeing the screen - 531-line racing AI in a single write: 4 personalities, curvature map, racing lines, tactical drifting. Built telemetry tools to compare player vs AI speed curves and data-tuned parameters - All assets from scratch: 3D models, procedural textures, sky shader, engine sounds, spatial AI audio! - Can do hard math: proved road normals pointed DOWN via vector cross products, computed track curvature normalized by arc length to tune AI cornering speed You are going to hear about this model a lot in the next months - open source let's go 🚀🚀
0:25

GLM-5.1 > Claude Code (Opus 4.6)? I'm tripping or CC has become very bad but built a Three.js racing game to eval and it's extremely impressive. Thoughts: - One-shot car physics with real drift mechanics (this is hard) - My fav part: Awesome at self iterating (with no vision!) created 20+ Bun.WebView debugging tools to drive the car programmatically and read game state. Proved a winding bug with vector math without ever seeing the screen - 531-line racing AI in a single write: 4 personalities, curvature map, racing lines, tactical drifting. Built telemetry tools to compare player vs AI speed curves and data-tuned parameters - All assets from scratch: 3D models, procedural textures, sky shader, engine sounds, spatial AI audio! - Can do hard math: proved road normals pointed DOWN via vector cross products, computed track curvature normalized by arc length to tune AI cornering speed You are going to hear about this model a lot in the next months - open source let's go 🚀🚀

Victor M

202,661 Aufrufe • vor 5 Monaten

Been giving my M3 Ultra Studio some love and tinkering with MLX. Started the afternoon at 42 tok/s. Now: 73 tok/s, and there’s still more on the table. Running Qwen3.8-Flash-Next @ 4-bit locally. Now for the crazy bit: Artificial Analysis Intelligence Index v4.3: GPT-5.6 Sol 39 (Medium) GPT-5.6 Sol 42 (High) Qwen3.8-Flash-Next 42 GPT-5.6 Sol 44 (XHigh) GPT-5.6 Sol 47 (Max) Grok 4.5 39 (High) Claude Sonnet 5 38 (Max) *Qwen score currently estimated by Artificial Analysis. A 4-bit open-weights model running at 73 tok/s on a Mac under my desk... with an estimated intelligence score matching GPT-5.6 Sol at High reasoning. Local AI is getting ridiculous. 🔥 Plus, this model does not say no to a bit of red teaming. Once I am happy with it, I will do a pr for Jun Kim's oMLX :)
0:48

Been giving my M3 Ultra Studio some love and tinkering with MLX. Started the afternoon at 42 tok/s. Now: 73 tok/s, and there’s still more on the table. Running Qwen3.8-Flash-Next @ 4-bit locally. Now for the crazy bit: Artificial Analysis Intelligence Index v4.3: GPT-5.6 Sol 39 (Medium) GPT-5.6 Sol 42 (High) Qwen3.8-Flash-Next 42 GPT-5.6 Sol 44 (XHigh) GPT-5.6 Sol 47 (Max) Grok 4.5 39 (High) Claude Sonnet 5 38 (Max) *Qwen score currently estimated by Artificial Analysis. A 4-bit open-weights model running at 73 tok/s on a Mac under my desk... with an estimated intelligence score matching GPT-5.6 Sol at High reasoning. Local AI is getting ridiculous. 🔥 Plus, this model does not say no to a bit of red teaming. Once I am happy with it, I will do a pr for Jun Kim's oMLX :)

Ash Hart

18,827 Aufrufe • vor 23 Tagen

I tried this so you don't have to. I know this is going to absolutely shock you but no this does not match the performance of Mythos. A few early thoughts: 1. The limits are pretty bad. I used 100% of my 5-hour usage in less than 1 prompt. 2. I specifically gave it a threejs task because it is an area that SOTA models have made big strides in, that other models just are not great at. I asked it to build a replica of Rocket League. I'll put the prompt in the comments. The game was pretty bad and notably worse than GPT 5.5. Even after multiple fixes, it took 7-8 back and forth with Codex just to get it an almost playable condition. Prior to these fixes, the game was not playable. Maybe it's really strong in other disciplines. I'd love to test that but I hit my limit in 1 prompt lol. GPT 5.5 by contrast did a pretty good job and required no follow ups. Fable would have absolutely nailed this as well. But yeah, early impressions...not great. But I hope I'm wrong. More testing tomorrow.
0:29

I tried this so you don't have to. I know this is going to absolutely shock you but no this does not match the performance of Mythos. A few early thoughts: 1. The limits are pretty bad. I used 100% of my 5-hour usage in less than 1 prompt. 2. I specifically gave it a threejs task because it is an area that SOTA models have made big strides in, that other models just are not great at. I asked it to build a replica of Rocket League. I'll put the prompt in the comments. The game was pretty bad and notably worse than GPT 5.5. Even after multiple fixes, it took 7-8 back and forth with Codex just to get it an almost playable condition. Prior to these fixes, the game was not playable. Maybe it's really strong in other disciplines. I'd love to test that but I hit my limit in 1 prompt lol. GPT 5.5 by contrast did a pretty good job and required no follow ups. Fable would have absolutely nailed this as well. But yeah, early impressions...not great. But I hope I'm wrong. More testing tomorrow.

am.will

594,106 Aufrufe • vor 3 Monaten

Caller: "I have absolutely no interest in having sex with my husband at all. And I'm kind of at a point where I don't really know what to do about it because I know while I may not need it, I know he needs it. And it's getting very challenging for me because I'm just kind of in this 'please don't touch me, I don't want to be touched' [mindset]. I want to sex... John Delony: "Of course. Caller: "...but I don't. John Delony: "Excellent. I imagine you're not telling a whole bunch of people that. Is that fair? Caller: "No, I think I've mentioned it to two girlfriends. One agreed with me, and one was like, 'What are you talking about?' John Delony: Okay. So, number one, I want you to know you are absolutely not crazy. You're not alone, okay? And not alone to the tune of millions, okay? So you're not nuts."
0:40

Sensitive content

Caller: "I have absolutely no interest in having sex with my husband at all. And I'm kind of at a point where I don't really know what to do about it because I know while I may not need it, I know he needs it. And it's getting very challenging for me because I'm just kind of in this 'please don't touch me, I don't want to be touched' [mindset]. I want to sex... John Delony: "Of course. Caller: "...but I don't. John Delony: "Excellent. I imagine you're not telling a whole bunch of people that. Is that fair? Caller: "No, I think I've mentioned it to two girlfriends. One agreed with me, and one was like, 'What are you talking about?' John Delony: Okay. So, number one, I want you to know you are absolutely not crazy. You're not alone, okay? And not alone to the tune of millions, okay? So you're not nuts."

Brown Legacy

107,501 Aufrufe • vor 3 Monaten

watch and let this sit for a second anon, a 35B AI model, building a complete multi-file game by itself on hermes agent. no cloud or api. not one line of code written by a human. running on a DGX spark in 2026. because not long ago this was science fiction. this is Ornith, a new open agentic coding model, running near lossless on a single DGX Spark, a 128GB box that sits on your desk. i hand it one spec and step back. it reasons about the architecture, creates the project structure, writes every module one by one, and here's the part that got me, it checks its own work as it goes. makes a call, reads the result, decides if it's right, corrects itself before moving on. no steering from me. this is what people keep underestimating about local AI. they hear "local" and think weaker, slower, a toy. watch the clip. a 35B at near full precision, planning and self correcting on a single desktop. the open models got genuinely good at agentic work, not benchmark good, actually builds the thing good. sped 5x so you can watch the whole build start to finish. this is where local AI actually is now. not coming soon. here.
2:29

watch and let this sit for a second anon, a 35B AI model, building a complete multi-file game by itself on hermes agent. no cloud or api. not one line of code written by a human. running on a DGX spark in 2026. because not long ago this was science fiction. this is Ornith, a new open agentic coding model, running near lossless on a single DGX Spark, a 128GB box that sits on your desk. i hand it one spec and step back. it reasons about the architecture, creates the project structure, writes every module one by one, and here's the part that got me, it checks its own work as it goes. makes a call, reads the result, decides if it's right, corrects itself before moving on. no steering from me. this is what people keep underestimating about local AI. they hear "local" and think weaker, slower, a toy. watch the clip. a 35B at near full precision, planning and self correcting on a single desktop. the open models got genuinely good at agentic work, not benchmark good, actually builds the thing good. sped 5x so you can watch the whole build start to finish. this is where local AI actually is now. not coming soon. here.

Sudo su

30,299 Aufrufe • vor 3 Monaten

hey here is the final result of octopus invaders on nvidia's flagship at full precision. nemotron super 120B on 2x H200 NVL. BF16 unquantized. 287GB of VRAM. hermes agent as the harness. 60 tok/s. first try it autonomously coded for 6 minutes straight. created 11 files. correct project structure. correct load order. started the server. i opened the browser and the result was a blank screen. i did not give up. second try i gave it a precise list of bugs and things to fix. it went back in for another 3 minutes. patched the code. served it again. still blank. so i did what any sane person would do. third try i just said the screen is blank, test it and fix it yourself. and this is where nemotron showed what it actually is. it became a debugger. you can see it in the video. realtime CSS test squares, red screen flashes, hermes agent browser tools, inspecting its own output. it built the parallax background with planets and comets. it rendered a rocket ship that tracks your mouse with fire and bullet physics. the aesthetic is real. but no enemies spawn. no collision. not playable. what surprised me is qwen 27B one shotted this exact game on a single RTX 3090 at Q4 quant. and here is nvidia's flagship at full precision on enterprise hardware needing 3 tries and still not getting there. that makes my hope high for the undisputed qwen 122B which is about to face the same test next. same hardware. same prompt and same harness. lets see if it one shots or not. full session in the video. no cuts. 5x speed.
3:36

hey here is the final result of octopus invaders on nvidia's flagship at full precision. nemotron super 120B on 2x H200 NVL. BF16 unquantized. 287GB of VRAM. hermes agent as the harness. 60 tok/s. first try it autonomously coded for 6 minutes straight. created 11 files. correct project structure. correct load order. started the server. i opened the browser and the result was a blank screen. i did not give up. second try i gave it a precise list of bugs and things to fix. it went back in for another 3 minutes. patched the code. served it again. still blank. so i did what any sane person would do. third try i just said the screen is blank, test it and fix it yourself. and this is where nemotron showed what it actually is. it became a debugger. you can see it in the video. realtime CSS test squares, red screen flashes, hermes agent browser tools, inspecting its own output. it built the parallax background with planets and comets. it rendered a rocket ship that tracks your mouse with fire and bullet physics. the aesthetic is real. but no enemies spawn. no collision. not playable. what surprised me is qwen 27B one shotted this exact game on a single RTX 3090 at Q4 quant. and here is nvidia's flagship at full precision on enterprise hardware needing 3 tries and still not getting there. that makes my hope high for the undisputed qwen 122B which is about to face the same test next. same hardware. same prompt and same harness. lets see if it one shots or not. full session in the video. no cuts. 5x speed.

Sudo su

11,011 Aufrufe • vor 6 Monaten

look anon, those of you who kept saying local AI is not there yet, who said open source can't compete, who said you need cloud APIs to get anything serious done, look at this gameplay for one minute. every pixel on this screen was written by one model, in one shot, on a single rtx 3090 with 24gb of vram. the model is qwen 3.6 27b dense q4. the harness is hermes agent. the hardware is a single consumer card you can buy used for 900 dollars. the prompt is open source on github. every claim verifiable, on your own desk. if your local AI take is from 2024, update it. the consumer tier is shipping work that was supposed to need 8 gpus and an api key. open source moved the floor while the rest of the field was busy explaining why it cannot. 24gb tier owners are eating ramen with half boiled egg and double chocolate.
1:39

look anon, those of you who kept saying local AI is not there yet, who said open source can't compete, who said you need cloud APIs to get anything serious done, look at this gameplay for one minute. every pixel on this screen was written by one model, in one shot, on a single rtx 3090 with 24gb of vram. the model is qwen 3.6 27b dense q4. the harness is hermes agent. the hardware is a single consumer card you can buy used for 900 dollars. the prompt is open source on github. every claim verifiable, on your own desk. if your local AI take is from 2024, update it. the consumer tier is shipping work that was supposed to need 8 gpus and an api key. open source moved the floor while the rest of the field was busy explaining why it cannot. 24gb tier owners are eating ramen with half boiled egg and double chocolate.

Sudo su

29,584 Aufrufe • vor 4 Monaten

A KID WITH CURSOR AI CAN NOW BUILD A ROBLOX GAME THAT MAKES $150M THIS IS NOT A JOKE Roblox has 80 million players logging in every single day. The platform pays out billions to creators annually. The top games generate seven and eight figure revenues every year. And until recently, building one required years of coding, 3D modeling, sound design, and game development experience. Cursor AI just removed every single one of those requirements. Maps, models, sounds, mechanics, monetization – all of it generated by describing what you want. The AI builds while you direct. Children are already shipping games this way. The Roblox gold rush isn't coming. It's already happening. The people getting in now with AI tools are building the games that will dominate the platform for the next decade. Bookmark this. The video explains exactly how to start.
2:59

A KID WITH CURSOR AI CAN NOW BUILD A ROBLOX GAME THAT MAKES $150M THIS IS NOT A JOKE Roblox has 80 million players logging in every single day. The platform pays out billions to creators annually. The top games generate seven and eight figure revenues every year. And until recently, building one required years of coding, 3D modeling, sound design, and game development experience. Cursor AI just removed every single one of those requirements. Maps, models, sounds, mechanics, monetization – all of it generated by describing what you want. The AI builds while you direct. Children are already shipping games this way. The Roblox gold rush isn't coming. It's already happening. The people getting in now with AI tools are building the games that will dominate the platform for the next decade. Bookmark this. The video explains exactly how to start.

SCOTTY BEAM

19,099 Aufrufe • vor 4 Monaten

This is the most insane story of 2025 The number 1 game in the entire world right now was built and shipped by a single person Made $20,000,000+ legit over night TylerTVGS built a Breaking Bad simulator by himself and now it has more players than games 1000+ devs worked on This is the power you have in 2025 Literally ANYONE can do this AI gives the power for ANY single person to come up with an idea, build it, and ship it no matter how little skills you have If you're thinking "oo you gotta be special and talented to do something like that" I promise you're wrong I did the same thing. I built and shipped an AI tool for X by myself and now it's my full time business I truly believe this is the single greatest economic opportunity of the last 1,000 years If you haven't started using Grok yet to plan and build apps, even if you have 0 experience, you're letting down your entire blood line Carve out your calendar, copy the prompt below, and run it today: "I have 0 programming experience at all but would like to build an app. Here is a list of my 5 interests. What are challenges than I can build with software? Choose the easiest challenge to solve then build me a plan for building it even though I've never programmed before in my life." Don't waste this opportunity.
0:40

This is the most insane story of 2025 The number 1 game in the entire world right now was built and shipped by a single person Made $20,000,000+ legit over night TylerTVGS built a Breaking Bad simulator by himself and now it has more players than games 1000+ devs worked on This is the power you have in 2025 Literally ANYONE can do this AI gives the power for ANY single person to come up with an idea, build it, and ship it no matter how little skills you have If you're thinking "oo you gotta be special and talented to do something like that" I promise you're wrong I did the same thing. I built and shipped an AI tool for X by myself and now it's my full time business I truly believe this is the single greatest economic opportunity of the last 1,000 years If you haven't started using Grok yet to plan and build apps, even if you have 0 experience, you're letting down your entire blood line Carve out your calendar, copy the prompt below, and run it today: "I have 0 programming experience at all but would like to build an app. Here is a list of my 5 interests. What are challenges than I can build with software? Choose the easiest challenge to solve then build me a plan for building it even though I've never programmed before in my life." Don't waste this opportunity.

Alex Finn

1,430,662 Aufrufe • vor 1 Jahr

✨ It's not even 2026 yet, but you can start using AI world models today! This is World Labs running inside my app 🏡 Interior AI letting you redesign your home with AI and then walk through it! It's really cool and of course not perfect especially since for now it's based on just one image, but multi-image will allow for more precise understanding of the room As always this is just the first version, and it'll get better from here World Labs is founded by Fei-Fei Li, a Computer Science professor from Stanford University, and other people like Justin Johnson, Christoph Lassner, and Ben Mildenhall "each a world-renowned technologist in computer vision and graphics" The tech essentially comes down to img2splat, search what a guassian splat is if you don't know It's in beta and wait list, but I'm trying to get this live for everyone as soon as possible as it seems like such an essential feature for interior design P.S. I also tried to worldify people photos (for Photo AI) but World Labs for safety right now seems to block it off and make it a black square, maybe in the future
0:44

✨ It's not even 2026 yet, but you can start using AI world models today! This is World Labs running inside my app 🏡 Interior AI letting you redesign your home with AI and then walk through it! It's really cool and of course not perfect especially since for now it's based on just one image, but multi-image will allow for more precise understanding of the room As always this is just the first version, and it'll get better from here World Labs is founded by Fei-Fei Li, a Computer Science professor from Stanford University, and other people like Justin Johnson, Christoph Lassner, and Ben Mildenhall "each a world-renowned technologist in computer vision and graphics" The tech essentially comes down to img2splat, search what a guassian splat is if you don't know It's in beta and wait list, but I'm trying to get this live for everyone as soon as possible as it seems like such an essential feature for interior design P.S. I also tried to worldify people photos (for Photo AI) but World Labs for safety right now seems to block it off and make it a black square, maybe in the future

@levelsio

430,981 Aufrufe • vor 1 Jahr

THIS $1,300 CARD HAS THE SAME 32GB OF VRAM AS A $2,000+ RTX 5090 - FOR LOCAL AI, THAT'S THE SPEC THAT DECIDES EVERYTHING. that clip is the gigabyte radeon ai pro r9700 ai top, 32gb of vram, built for local ai. the pitch is simple: the same memory capacity as an rtx 5090, for roughly a third of what a 5090 actually sells for. here's why that matters, and where it doesn't. for ai, vram is the gate. a model either fits in memory or it doesn't. 32gb loads models a 16gb card can't touch, full stop. so for running local llms and image models, this card does the one thing that counts, cheap. honest line: this is not a 5090 for gaming. the 5090 is far faster at raster and ray tracing. they only match on how much vram they carry. if you want frames, buy the 5090. if you want to fit big models locally, you just found the same 32gb for a third of the money. no $2,000+ price tag, no cloud gpu rental, no model too big for 16gb. the uncomfortable part isn't that amd is cheaper. it's that for local ai you were about to overpay for gaming power you'd never use. follow RetroChainer for more on paying for the spec that matters, and bookmark this before your next gpu.
1:52

THIS $1,300 CARD HAS THE SAME 32GB OF VRAM AS A $2,000+ RTX 5090 - FOR LOCAL AI, THAT'S THE SPEC THAT DECIDES EVERYTHING. that clip is the gigabyte radeon ai pro r9700 ai top, 32gb of vram, built for local ai. the pitch is simple: the same memory capacity as an rtx 5090, for roughly a third of what a 5090 actually sells for. here's why that matters, and where it doesn't. for ai, vram is the gate. a model either fits in memory or it doesn't. 32gb loads models a 16gb card can't touch, full stop. so for running local llms and image models, this card does the one thing that counts, cheap. honest line: this is not a 5090 for gaming. the 5090 is far faster at raster and ray tracing. they only match on how much vram they carry. if you want frames, buy the 5090. if you want to fit big models locally, you just found the same 32gb for a third of the money. no $2,000+ price tag, no cloud gpu rental, no model too big for 16gb. the uncomfortable part isn't that amd is cheaper. it's that for local ai you were about to overpay for gaming power you'd never use. follow RetroChainer for more on paying for the spec that matters, and bookmark this before your next gpu.

RetroChainer

38,844 Aufrufe • vor 1 Monat