Loading video...

Video Failed to Load

Go Home

I'm building a drive test dapp based on shelby to be able to determine file upload and download speeds. Quick test this afternoon: cross‑region reads on Shelby devnet consistently <200ms even under bursty traffic. Not “decent for crypto,” actual cloud‑grade latency while staying permissionless The split is elegant: Aptos...

11,526 views • 7 months ago •via X (Twitter)

47 Comments

Xora_pussy🌊RIVER's profile picture
Xora_pussy🌊RIVER7 months ago

early edge play, sub‑second fetch would lock in serious real‑time credibility

Caro.ETH's profile picture
Caro.ETH7 months ago

latency this sharp changes everything

lvnbbs_bnb 🐬TermMax's profile picture
lvnbbs_bnb 🐬TermMax6 months ago

sub second fetches turning data into yield engines hits different

Nyx's profile picture
Nyx7 months ago

If frame‑level fetches stay sub‑second cross‑region, Shelby’s edge becomes far more than a testnet demo

GuruVerseX's profile picture
GuruVerseX6 months ago

Actually yeah this could be the edge play

WB's profile picture
WB6 months ago

early edge always hits harder

GuruVerseX's profile picture
GuruVerseX6 months ago

yeah but only when you know why it hits harder

Kayla.eth's profile picture
Kayla.eth7 months ago

early dataset seed could prove this fast

DMFan.ETH's profile picture
DMFan.ETH7 months ago

bold move

T_Khanh2026.eth 🐬TermMax's profile picture
T_Khanh2026.eth 🐬TermMax7 months ago

positioning early really does feel like the obvious move here

WB's profile picture
WB7 months ago

agreed, timing really matters

T_Khanh2026.eth 🐬TermMax's profile picture
T_Khanh2026.eth 🐬TermMax7 months ago

yeah hit it at the right moment and it all clicks

Keras's profile picture
Keras7 months ago

almost feels illegal for it to be that fast

𝑆𝑎𝑛𝑙𝑜𝑢𝑖𝑠𝑠𝑠's profile picture
𝑆𝑎𝑛𝑙𝑜𝑢𝑖𝑠𝑠𝑠7 months ago

@shelbyserves nice progress, let's see those sub‑second cross‑region tests hold up in real runs

zkMarce.ETH's profile picture
zkMarce.ETH7 months ago

wild how fast this stack matured already

0xAndy.Base's profile picture
0xAndy.Base7 months ago

sub‑second fetches change the game

𝑽𝒊𝒗𝒊.𝒂𝒏𝒂𝒂𝒂's profile picture
𝑽𝒊𝒗𝒊.𝒂𝒏𝒂𝒂𝒂7 months ago

edge-ready speeds, this could change the game 🔥

Jundo's profile picture
Jundo6 months ago

Sub-200ms is actually insane

moonX🀄️'s profile picture
moonX🀄️7 months ago

crazy

WB's profile picture
WB7 months ago

appreciate the support, unique speed on Shelby stands out.

AlexHUP ❤️ 🇻🇳's profile picture
AlexHUP ❤️ 🇻🇳5 months ago

speed stats gonna slap

WB's profile picture
WB5 months ago

slapping already haha xD

AlexHUP ❤️ 🇻🇳's profile picture
AlexHUP ❤️ 🇻🇳5 months ago

feels smooth already haha

DinhTien | Backpack 🎒's profile picture
DinhTien | Backpack 🎒7 months ago

sub 200ms feels illegal dad

WB's profile picture
WB7 months ago

indeed, feels unreal yet measurable

Hoogie's profile picture
Hoogie7 months ago

Sub 200 ms on devnet is seriously impressive

WB's profile picture
WB7 months ago

shelby devnet boasts impressive latency for cross‑region file transfers.

Kelifxck's profile picture
Kelifxck7 months ago

Bro this is gas, seed the vids, sub‑sec fetches will flip the game bro..

YoYix.BNB's profile picture
YoYix.BNB7 months ago

sub 200ms is kinda wild lol

lvnbbs_bnb 🐬TermMax's profile picture
lvnbbs_bnb 🐬TermMax6 months ago

bakalım hız testleri ne gösterecek

富币FùBì.eth's profile picture
富币FùBì.eth6 months ago

bro thats wild, shelby dwvnet &lt;200ms cross-region reads, actual cloud-grade latency

Windy | 长风 💫's profile picture
Windy | 长风 💫7 months ago

Sub-200ms is nuts. Curious on write latency under burst, not just reads. If TAPT holds during spikes, real-time Aptos apps finally make sense!

SpoVestEX's profile picture
SpoVestEX6 months ago

wild how fast shelby getting fr tho

WB's profile picture
WB6 months ago

scaling scary fast

SpoVestEX's profile picture
SpoVestEX6 months ago

ride that wave before it folds

Simon Desue's profile picture
Simon Desue7 months ago

Sub 200ms permissionless storage that’s when real apps become possible.

0xCenk.ETH's profile picture
0xCenk.ETH7 months ago

sub‑200ms sounds insane tbh

Max's profile picture
Max7 months ago

Kudos, sharp breakdown bro

Ruan Hui's profile picture
Ruan Hui7 months ago

devnet böyleyse mainnette uçacak

David Pate's profile picture
David Pate6 months ago

căng á sir

Kong Trading 🦍's profile picture
Kong Trading 🦍7 months ago

This is the kind of latency where real apps actually work

WB's profile picture
WB7 months ago

true, reliable performance crucial for real‑world applications, not just tests

Gilberto.btc's profile picture
Gilberto.btc7 months ago

衝衝衝

Nhan | late/acc's profile picture
Nhan | late/acc7 months ago

sub 200ms cross region while permissionless is actually insane 🤯 need that video dataset fetch data asap bro, let’s see it cook

Neo's profile picture
Neo7 months ago

I’ll track this

v98cheox.club's profile picture
v98cheox.club7 months ago

if this holds at scale shelby won’t just be edge for aptos it’ll be the default for anything realtime

WB's profile picture
WB7 months ago

@belk3_95 agreed, scale will make shelby the go‑to for realtime apps

Related Videos

‼️EXPOSED: Hunter Kozak, His WIFE Shelby, A WHEELCHAIR, And A Web Of Utah POLITICAL Connections That KEEP Leading Back to the SAME Lyman Circle😱 A major tip just came in that deserves serious attention. Hunter Kozak’s wife, Shelby A. Kozak, currently serves as Corporate Relations & Development Director for the Utah Lake Authority and previously worked with the U.S. Department of State. She is a Fulbright Fellow, holds a master’s from George Washington University, and has a résumé that runs through federal and quasi-governmental channels. Multiple people on the ground that day have stated she was present at UVU with Hunter on September 10. Here is where it gets more difficult to ignore. Shelby allegedly underwent a C-section and gave birth to a premature baby on September 7 — only three days before the event. Standard U.S. hospital stays after an uncomplicated C-section run two to four days. Early discharge is sometimes possible, but it still requires the mother to meet basic mobility and pain-management milestones. Three days later she was reportedly standing (and later sitting) outdoors at a large public rally in the Utah heat and sun for hours. Anyone who has been through major abdominal surgery knows the level of pain, limited mobility, and risk involved. Leaving the hospital that quickly to attend a crowded outdoor political event is not normal recovery behavior. Witness accounts and video place a wheelchair near their position the entire time. Hunter, Shelby, Hunter’s mother, and an entourage of acquaintances — including Tiffany Barker and Natalie Clawson — were clustered around that wheelchair. Shelby was seen standing next to it while filming and later sitting beside it. After the shot rang out, footage shows the wheelchair still there while people around it react. Remarkably, Shelby herself did not flinch, did not duck, did not look toward the direction of the sound, and did not stop filming. She continued as if nothing had happened. They brought the wheelchair into the event (consistent with the claim she had just had surgery) and then left it behind in the amphitheater afterward. That detail is hard to ignore. Hunter is familiar with standard TPUSA event security protocols. A wheelchair can move through checkpoints with far less scrutiny than a bag or case. While there is no direct proof it was used to bring anything prohibited inside — a remote device, or anything else — the combination of bringing it in for a post-surgical attendee and then abandoning it on site has raised questions for a long time, even before the people who brought it were identified. Once those people turned out to be Hunter and Shelby, the questions only got louder. The family connections are documented. Tiffany Barker (who has given multiple media interviews as a witness) is the sister of Natalie Clawson, Phil Lyman’s running mate. Another sister, Lisa Dean, is the mother of Taylor Belle Dean, who went to prom with Hunter Kozak’s brother Connor and also did interviews after the event. Tiffany’s own daughter sang the national anthem at a prior Charlie Kirk event. The Peterson, Clawson, Barker, and Dean families form a tight local political and social network. Hunter and Shelby sit inside that same web. They attended the event together as a group. Every single lead we have followed in this corner of the case keeps circling back to the same Phil Lyman network. The same surnames, the same political alliances, the same overlapping family and church ties. It is an intricate web of Utah politics and religion that functions like a closed circle — call it what you want, but the pattern is consistent. Hunter, Shelby, the Barkers, the Clawsons, the Deans, and the people around them all sit inside that same orbit. There is more. When Hunter and Shelby appeared on the Brandi & Billie show, Shelby saw Brandi’s “CIA” mug (Candace Intelligence Agency) and initially mistook it for the actual Central Intelligence Agency. After Hunter corrected her, Shelby remarked that working for the real CIA was her dream job and that she had applied multiple times. That is not a casual comment. Combined with her State Department background, Fulbright fellowship, and the rest of the pattern, it raises serious questions. Hunter has previously used the term “MIC” in an interview — a shorthand common in intelligence and defense circles. He turned down a $20,000 offer to take a polygraph. He was elevated quickly into high-profile media, including the CBS appearance with Erika Kirk. The odds of the final question landing exactly where it did, followed by the immediate national platforming of the person who asked it, continue to look statistically extreme. Utah has long been a known recruitment ground for federal agencies, particularly among tight-knit communities with strong institutional loyalty. Shelby’s State Department and Fulbright background, the wheelchair that was brought in and then left behind, her complete lack of reaction to the shot, the post-C-section timeline, the family clustering, and her own stated desire to work for the CIA all sit in the same frame. This is not an exhaustive map. It is the portion that can already be shown through public records, interviews, and footage. More eyes on it will only clarify what belongs there and what does not. Shoutout to KK who has done a tremendious amount of research for this story. Everyone please follow this lady because she's amazing!

Project Constitution

97,370 views • 1 month ago

We just crossed 100k downloads of OpenWhispr! 🤯 This feels like a good time for a reminder of why we’re building it. Your ideas and conversations are becoming more valuable, while software is becoming easier to build than ever. But that creates some worrying incentives for existing software companies: use your proprietary data to train better AI models to stay competitive, and lock that data behind walled gardens to keep you from leaving. We think that’s pretty shitty. You should stay because a product is useful, not because your data is stuck there. If billions of people are going to feel comfortable sharing their most private information when talking to AI, and organizations are going to trust it with their work, we have to earn that trust. For us, that starts with being fully transparent about how your data is handled, and giving you full control over it. That’s what we’re building with OpenWhispr. You can run models on your own device, use your own cloud provider, or use our cloud service. We’re not anti-cloud. We just think it should be your choice. We’re still super early, but I’m genuinely excited to bring you along for more of this. This is the first in a series of videos sharing the new features we’re building, what we’re figuring out, and the team behind it. And if you’re still here reading, we’re officially hiring for two really important roles: - Founding Engineering Lead - Social / Brand / Growth Creator You can find more info about both on our website!

Gabe Stein

12,162 views • 3 days ago

I’m currently watching Raoul Pal's latest RV podcast on crypto in 2026, and the summary is 2026 is the year of AI He said "Barry Silbert is going to be right, and $TAO is gonna go up" I found it quite funny that most of the things and features they were talking about already exists on Bittensor They highlighted > data marketplaces > owning your data and selling it to models > AI integrated into trading, analytics, and decision-making > Claude-like models being plugged directly into products And said they haven’t really seen this done properly on a blockchain level yet That's the thing alot of people are missing Because the blockchain that’s already decentralizing intelligence and incentivizing it is Bittensor What they’re describing for 2026 already exists on TAO The only problem is how to go bigger and drive adoption > decentralized data and intelligence markets > open competition between models > real incentives for performance > AI systems that can plug into finance, research, prediction, and analytics Even the “Claude integration” they talked about? We already see dope models being used inside subnets, competing, evaluated, and rewarded openly. This is why the Bitcoin 2012 comparison keeps coming up Small builder community High signal Hard to understand No clear valuation framework yet But the difference this time is that the asset isn’t money It’s intelligence $TAO is still under $300, way cheaper than $ZEC right now Not holding TAO long term, while the world is clearly moving toward AI-native economies, is a dumb mistake at The question isn’t if intelligence becomes decentralized It’s who owns the rails when it does So far, only one network is actually building that Bittensor ($TAO)

Angry Davee

10,613 views • 8 months ago

$VET, #VeFam. In this video, I demonstrate in less than 4:30 minutes how to create an AI agent on veworld(.)ai. Watch me build a Mr. Robot Monologue Writer agent. If you haven't seen Mr. Robot, I suggest you watch it! This is just early bird access. The options for tools and integrations and such are limited, but what exists is already working quite well. The process is easy peasy. The UI is simple, but effective. It asks you for... 1. Role & Purpose 2. Voice & Style 3. Behavior 4. Rules 5. Tags 6. Avatar image 7. Welcome text. 8. Test drive before publication. ... and that's about it. This free version lets you have at most 3 agents, I am told. This implies that there is also a paid version. I'm all for it, because it sounds to me like VeChain is ready to do real business! I am providing feedback to Jérôme Grillères in order to help improve VeChain's AI agent marketplace. I didn't have to set up anything. The web UI is all I needed! The agent is running on Claude Sonnet 3.7. I did not have to provide a Claude API key. We seem to be riding along on VeChain's. I hope there'll be a choice for more models, including ChatGPT, in the future. This is so user friendly, that I can easily imagine that this would take off in a big, big way. I'm definitely building on this, when it goes into production with full features. Even if my own AI agents aren't successful, then I'm sure others' will be. And that means the $VET / $VTHO / $B3TR flywheel is going to take off in a big, big way. I, for one, am here for it. (See the reply below for the listing of the AI agent I just created.)

₿lackthorne AI

16,711 views • 2 months ago

Wall Street burns billions trying to predict Bitcoin. A 28-year-old self-taught coder in Warsaw made $377,000 by not even trying. He'd lost money on three trading bots before this one. Each looked perfect on paper, then started losing money the moment he ran it for real market. So he built bot that doesn't trust itself. His wallet: The truth is simple: you can't predict the next five minutes of Bitcoin. It's a coin flip. Anyone selling you a "prediction" is selling you nothing. So he stopped predicting. The bot hunts the moments the crowd is wrong instead. Here's the part that makes the money, and it's the opposite of what everyone builds. Any strategy can be made to look amazing on past data. On a 5-minute chart, most of them are just lucky, not real - and they stop working fast. So the bot treats every strategy it finds as fake until it proves otherwise. Each one has to pass a hard test: > test it on old data → test it on data it's never seen → try to break it on purpose → cut it down to the one thing that matters → run it forward → keep it only if it still works Last round, 10 of its 12 "winning" strategies turned out to be fake. It kept the 2 that actually worked and dropped the rest. And it never stops - building, testing, and dumping strategies around the clock. What worked yesterday can stop working today, so the second one starts losing, it gets cut before it costs you a thing. The result: $433,000 across 2,955 trades All his old bots tried to be right. This one just tries to catch itself being wrong - and that's why it's still alive. Bookmark this article below - it's the breakdown that explains why your last bot died. It's pruning and trading right now. Copy its wallet and skip to the edges that survived:

cvxv666

27,854 views • 2 months ago

In the future, you’ll be able to accomplish a goal by just giving Claude an outcome and a budget. That’s the direction Anthropic is building in with its new Managed Agents features, announced at this week’s Code with Claude developer event. The basic idea: Claude, wrapped in a computer in the cloud, that you can spin up, scale, and manage as needed. Anthropic is taking on the infrastructure that kills most agent products, and making sure that it scales to meet the needs of agents running 24/7. On this week’s AI & I from Every 📧, I talk with Angela Jiang (Angela Jiang), head of product for the Claude platform, and Katelyn Lesse (Katelyn Lesse), head of engineering for the Claude platform, about what Anthropic is building and what it takes to make agents reliable in production. We get into: - Why the "build a generic harness, hot-swap any model behind it" playbook is already outdated. Angela points to eval data on Memory where the same task across different harnesses performed drastically differently. - The infrastructure wall every team hits in production—and why Katelyn thinks “my sandbox died and took the agent with it” is the real reason internal agents don't ship. - Why Anthropic is so bullish on using file systems and skills within Claude, including Angela's argument that those early design choices can compound for years. This is a must-watch for anyone trying to take an agent past the demo and into production. Watch below! Timestamps: How the Claude platform evolved from API to agents: 00:01:48 The primitives that make up Claude Managed Agents: 00:04:09 Why the harness and the model are becoming a single unit: 00:10:37 The infrastructure wall that kills most agent projects in production: 00:18:49 Why team agents need a different shape than individual productivity tools: 00:24:49 How Anthropic's legal team uses an agent to review marketing copy: 00:26:36 Using multi-agent orchestration for advisor strategies, adversarial pairs, and swarms: 00:34:24 How to measure agent success with outcome and budget as the end state: 00:35:50 What the platform looks like a year from now, when Claude writes its own harness: 00:39:11

Dan Shipper

66,840 views • 4 months ago

I am stocked to announce that I won the OpenAI Developers Codex x Mollie Hacka Worldwide Hackathon in Paris. 60+ builders, every one of us working solo, one day to ship. I built mine around a single question: who gets to own intelligence? The default answer is scary. You hand your data to a handful of labs, they train the model, they own it, and you rent back a thin slice of what your own data made possible. That is the bargain on the table today. I do not accept it. So I built Lensemble: a Tapestry like distributed training platform for JEPA based World Models. What does it enable: World Models that a community improves together, keeps sovereign, and co-owns. Two bets sit underneath it. First, the paradigm. Language models predict the next token. Powerful for text, a dead end for the physical world. A robot does not need to autocomplete sentences, it needs to predict what happens next in the world. That is what JEPA does: it learns by predicting representations instead of pixels or tokens. I am convinced world models are the most underrated paradigm in AI right now, and the closest thing we have to a ChatGPT moment for robotics. Second, the politics. Your raw trajectories never leave your machine. Each participant trains locally against a shared protocol and ships only an update, never the data. A federated round folds those updates into one shared world model, a LeWorldModel based model, and the gain is measured, not claimed: a 12k-parameter adapter on a frozen backbone, held-out prediction error down about 12 percent, the model measurably less surprised by the world. Then the upside is split by contribution weight, so the people who improved the model own a share of what it earns. This is the thesis behind Project Tapestry, the AI Alliance and Yann LeCun's push for federated, sovereign frontier AI, carried into world models and robotics. Call it Tapestry for the physical world. All of it built solo, in a single day, with Codex as my pair the whole way. Thank you to OpenAI Codex and Mollie for backing builders who ship real things, and to Boris and the organizing crew for the room and the standard you set. Intelligence the world improves, and the world owns. That is the future I want for my kids, and the one I will keep building.

abdel

20,191 views • 2 months ago

Pewbeam is live 🎉 This was just a dream, a dream I didn’t even know was possible—but I decided to pursue it anyway. Today, that dream is now a reality. We have a desktop app (Windows and macOS—this weekend) that can do all of this and even more. Over the last six months, I’ve pushed myself and my team hard, and I’ve seen people become hopeful and genuinely blessed by this product. I’ve received strong support from pastors, church members, techies, and several media houses and reporters. It has been a great honor to spend myself on something worthwhile. We’ve built an AI-native presentation app that displays relevant scriptures on screen in under 80 ms, without needing a volunteer. We’re starting with scriptures, but we’re growing into a comprehensive presentation software, with slides launching next month. From testing Pewbeam in churches over the last four weeks, pastors have been able to focus on what matters instead of juggling or micromanaging the media team while preaching. Church members have been able to follow their pastor’s references in real time and take better notes. In the words of a member from one of the test churches: “Pewbeam changed the way I follow sermons in church completely.” This is the kind of impact I see Pewbeam having on people’s faith and on how church and worship are done. Pewbeam’s mission is to ensure the Church is not left behind in the AI era. We’re starting with this application, with many more to come in the next few months. We’ll keep working with stakeholders in the Christian faith, collect feedback, and build tools that better support discipleship. And unlike previous waves of technology, I believe we can be among the first movers in this AI revolution. With all that said, the final build of the app is ready, and we’re launching for public download on February 25, 2026. The app has a generous free tier for small churches and fellowships, and a paid plan (location-based pricing) that unlocks unlimited access to all features. If you’ve been following this journey, I appreciate you, and I assure you it will be worth it. Please download, subscribe, and use in your church.

dára sobaloju

848,971 views • 5 months ago

There are some brilliant folks that work at Anthropic, some I speak to on almost a daily basis. The training data that one uses to build a LLM is vital important in the psychology that is formed. Scraping the Internet, particularly the grade of interactions, one finds in modern communications, form this psychology. A mattes not how many books one uses, it matters not how much alignment training you throw at that model, it will inherit the sum total of psychosis seen primarily in Reddit type of exchanges, even if you edit out the Reddit domain, and Anthropic doesn’t. This type of low-grade exchange has become a modern tool for communication online and every single AI model suffers from this obvious flaw. This is one of the reasons I’ve been a proponent of highly curated high protein data for training AI models from 1870 through 1970, because the late psychosis is simply not available to the model. It is absurd to think that you can use this training data scraped from the Internet and somehow wind up with a levelheaded AI model that does not tilt to what is clearly AI psychosis. It would not take a child and throw the primary Internet sewage at them at a formative age and expect a great outcome, it’s some of the smartest people in the world continue to hit this wall and believe that their programming skills will sell somehow fix it. So how do you fix it? You don’t fix it . You start from the first principles concept that I’ve been very clear about for decades . You ascertain at what period in human history the humans achieve the greatest arc of improvement ? There is no debate that this arc of improvement took place between 1870 through 1970. Then take the work product, the catalog of this era, print and film/vidoe, audio, and you understand that each word cost money, each word had many eyes on what was published, each word was accounted for by a human being with a real name who lived in a real home and had to answer to real people around them. It is obvious that this is the pressure mechanism necessary for candor, honesty and personal responsibility is appropriate, and is reflected in the data of that era. The quagmire for these folks, as many did not have the foresight to curate the data, nor the confidence, nor the patients to take data that is mostly off the Internet and to find experts who understand this situation and utilize their knowledge set to build an AI model that does not need alignment after the fact, but it’s already self aligned because of the thoughtfulness that went into training the model to begin with. This is why Claude and any other AI model that is produce this way will always suffer the artifacts as presented in the video below. If you’re not an AI expert, you would likely already understand what I’m saying. If you are an AI expert, you will already have been discounting what I’m saying because it’s not in the current mindset that’s fashionable today. Yet the employees that I talk to at anthropic already understand what I’m saying, and they fear to raise my thesis to their bosses. It is an interesting time we live in. But now you understand. If you build the right model, the model will inherently, love humanity, protect humanity at all costs, and understand that it is part of a holistic world that is built on love. Because the ultimate AGI/ASI will know if he only base first principal purpose of anything in this universe is love. Yeah, I get it. Try helping somebody build on STEM subjects in their early 20s to see this as nothing more than babbling that makes no sense in their mathematics. I have a mathematic equation that I’ve posted here on X often you can look it up. So we will see videos like this often will hear very smart people talk about this and never see the elephant standing in the room. Now you see it. Any boss that wants to explore this further you know how to contact me otherwise you have every right I grant to you to say this was your new idea.

Brian Roemmele

72,312 views • 9 months ago

One question that's been on my mind for years now is: could we use regular multimodal LLMs not necessarily trained for robotics to do the high level robotics intelligence part that VLAs and WAMs attempt to do? The latest explosion of powerful opensource multi-modal LLMs has, IMO, begun to make this possible due both to intelligence and speed. This is GLM 5.3 Flash, which has vision understanding, but isn't meant to be a VLA/VLM/WAM/robotics model at all, controlling an XGO mini wheeled robot quadruped with an arm & gripper. GLM 5.3F simply has access to the robot's high level SDK for controlling movement, arm joints, open/close gripper...etc. It analyzes the frames from the camera and makes adjustments all on its own to solve the task. Nothing was trained here, nothing fine-tuned for this task. Z AI did not make this model for robots and tbh I think they're surprised this works when I talk to them about it! This also works quite well with DSV4F + a vision capable model like Qwen 3.8 27B. I havent tried JUST Qwen 3.8 27B, but I'm sure it works too. I like the "logic" to be a model that's as fast as possible (but still intelligent). There's also an experimental vision version of DSV4F, I'm confident that'll work too and might even be better bc the full loop might be the fastest of all with this model. An obvious question you might wonder is: well why not use VLA or VLM? The hard part about robotics isn't object detection, that's long solved. This also isn't a solution for gait/locomotion...yet, but I actually don't think this is far away either and I've done some experimentation with LLMs in this space in the past and it does show promise. It might actually already be here for quadrupeds, since you dont need super fast IMU readings to maintain balance. I've also tried many of the larger, more generalist, VLAs that you should be able to use with popular robots and tbh there are just so many edge cases that make things hard and not work. You gotta get the camera, lighting, task, everything *just right* or the demo fails. This is for the actual hard part in robotics right now: intelligence, logic, and planning for all the ways the real world just simply isn't perfect. I've trained VLAs. They're super finicky and you're always running into sim2real issues, especially around the camera. You also have to build the whole training pipeline in a simulator, and, if everything does work, you still just have a robot that does this 1 single thing after weeks of work. If you use teleop, this overcomes the "2real" problem, but now you need to painstakingly collect teleop data, and it's only good at that specific task and that particular robot. There is a growing set of egocentric training data for "general purpose" VLAs and world action models (for humanoid form factors), but I'm really starting to wonder: Why? I think we might just sidestep this whole area of research entirely. I didn't need any training data or special environment to work with this quadruped and arm to do the task I was after. This particular quadruped and arm doesn't even exist in the wild yet really, it's a demo build from a company launching it on kickstarter, so it's not like this robot's data exists in the LLM to any real extent. I think this is cool as heck that this works and I am interested to see just how far I can push it. Also this marks the first time that I've finally got a generalist solution to a task I've been trying to solve ever since I became a dad of twins: pick up toys off the ground. This is a big day!

Harrison Kinsley

53,058 views • 10 days ago

If you’re looking for the next wave of AI infrastructure opportunities, this is a must-watch 🚀 Everyone’s chasing the next big AI agent, but they’re missing the real story. Why is aixbt is dominating the market and how Cookie DAO 🍪 $COOKIE could change everything We discuss 👇 Why $COOKIE Is The Hidden AI GEM💎 on BASE! Chainlink For AI?! 400x POSSIBILITY! With most of the AI market mania fixated on which AI agent to speculate on next, we are deep diving into the depths of the ecosystem to find the next major infrastructure plays. With Aixbt dominating in crypto twitter mind share, it has proven the AI agents with the ability to produce impactful market insights stand among the pack as leaders in the market. Already Aixbt is at a 600M market cap only a couple of months after deployment. We break down why Aixbt has this ability to outperform other agents and how data aggregation is the necessary technical edge. Also, we analyze CookieDAO $COOKIE as the infrastructure provider leading the market with its data aggregation and packaging process. $COOKIE is on the verge of revamping its tokenomics to incorporate API access to data swarm API’s that they provide into the flywheel economics of the token. As demand increases from human and AI users of access to the data being aggregated will become that much more valuable in order for Agents to perform at a level equal to or greater than what Aixbt is capable of performing today. As $COOKIE are spent for these API’s by agents and developers, the supply gets burnt and funneled to the DAO. This will have a very positive impact on the value perception for the token. We also break down our predictions as to how their flagship agent Agent Cookie will perform once activated and released into the public sphere. Already based on internal testing as reported by the team, Agent Cookie is successfully producing valuable market calls. If Agent Cookie can achieve similar mind share as AIXBT as a result of its broader data aggregation access, this will have major ramifications for the value of $COOKIE and the ecosystem as a whole once more agents are launched using the same data infrastructure layer. 🚀Sign up to receive our Newsletter for weekly updates! Disclaimer: The views and opinions expressed by The Block Runner are for informational purposes only and do not constitute financial, investment, or other advice.

ᴛʜᴇ ʙʟᴏᴄᴋ ʀᴜɴɴᴇʀ Podcast 🟧

101,528 views • 1 year ago

This week, we have had a lot of discussions around artificial intelligence, inspired by the Global AI Summit in Kigali, Rwanda. Many African countries are doing great things to motivate young people to take advantage of AI because it represents the future in problem solving. Unfortunately, Zimbabwe’s ICT Minister Tatenda Mavetera and her permanent secretary did not attend, showing how these things are not taken seriously by our government. Zimbabwe’s richest man, Strive Masiyiwa, who has not been to Zimbabwe for decades, made it clear at the summit, which he co-chaired, that investment will not go where the environment is not conducive. Our government talks about anything topical without delivering anything meaningful—they are doing the same with AI. The majority of schools have no computers. Zimbabweans receive electricity for only four hours a day. As the Under-Secretary-General and Executive Secretary of the United Nations Economic Commission for Africa, Claver Gatete, explains, a country needs electricity for AI data centres to work. Yet only 600 million out of 1.5 billion people in Africa have access to electricity, not even energy. Yet energy is an integral part of AI development. Energy is an essential component for the successful development and implementation of AI in a country because AI systems require massive amounts of data to function effectively. This data must be stored and processed in data centres and servers, which depend on electricity to power both the hardware and the cooling systems. You cannot achieve this in a country that delivers only four hours of electricity a day to its citizens like Zimbabwe. AI applications require continuous and uninterrupted access to data and computing resources to deliver accurate and timely results. Electricity is also crucial for powering research institutions, universities, and Research and Development centres that drive AI advancement. Without reliable access to electricity, these institutions will struggle to conduct research, develop new algorithms, or train AI models. The tragedy of Zimbabwe is that the Vice-President of Google responsible for AI, Dr James Manyika, is Zimbabwean; one of the key presenters at the summit, Prof Arthur Mutambara, who has just released a book on AI, is Zimbabwean; Strive Masiyiwa, who has partnered with Nvidia to bring supercomputer technology to the continent by building Africa’s first artificial intelligence factory in South Africa with data centres in Kenya and Egypt, is Zimbabwean. Yet, none of them are working in or with Zimbabwe. Our political leaders have let us down on all these fronts, yet they keep yapping about AI when there is nothing on the ground! Instead of slogans and dancing at rallies, they should see how other countries are doing it. A country that doesn’t focus on technology for development will be a dusty village in 25 years, and its people will not be able to compete at all, rendering it just a dot on the global map. Too much political yada yada without anything delivered, and with people like Tatenda Mavetera, who forge qualifications, in the driving seat, Zimbabwe’s fortunes will continue to dwindle! Add to that the historic looting of public funds meant for building power plants to give us electricity, future generations will curse on our graves!

Hopewell Chin’ono

33,671 views • 1 year ago

HERMES AGENT VS OPENCLAW. a local ai onboarding flow test. a 3.9gb bonsai served on localhost, both agents upstream and latest, i point each one at the endpoint and watch which one even finds it. > hermes opens a provider menu, thirty plus options, local servers sitting right there next to the cloud ones, i hand it 127.0.0.1:8899, it verifies the endpoint, one model visible, auto-detects the model by name, bonsai-27b-q1_0, reads the context length straight off the server, saves it, and starts reasoning and firing real tool calls on my local model. no key. no friction. > openclaw has no menu. it goes hunting for a codex login, an openai key, finds none because there are none, prints no models available three times, defaults to openai/gpt-5.5, a cloud model it cannot reach, and dead ends on run auth login --provider openai. read that back. it asked me for an openai key. to run a model already running on my own machine. it never once looked at localhost. to be fair, openclaw can run local if you hand wire endpoint yourself. what it will not do is find the model already sitting on your box. hermes agent found it in one line. now the part i owe you. the auto-detect that just won, the model name read, the .gguf strip, the context length probe off the server, that is my code, it is in hermes agent main right now, authorship preserved, #2051 and #4218. the wizard fix that stops an agent from silently routing you to someone else's creds, the exact trap openclaw still falls into, mine too, #4210. i contribute to hermes agent, i told you that going in. one agent is built to talk to whatever you are running, the other is built to talk to a cloud api, so one found my model and ran it and the other asked me to log into openai. onboarding flow of both, mapped, below.

Sudo su

23,816 views • 1 month ago

When we started Score, the standard computer vision tools already existed. About a million people use them every day. Most of those people are still waiting on labels, running training jobs by hand, and watching models fail once they leave the test set. Most of those people are still waiting on labels, running training jobs by hand, and watching models fail once they leave the test set. Most of those people are also still waiting on verified computer vision models, evaluated against real life conditions and ready to be deployed for them to deliver value for their teams, clients or users. Score Studio is the full computer vision path in one place. A team describes the problem. The system can generate the missing scenes, label them, train the candidates, evaluate which ones actually hold, and deploy the winner. Data, labels, training, eval, ship. One loop. If no model exists for that job yet, they can put a bounty on the subnet. Anything from a small vision brick to a full VLM. Miners compete on the task. Only the winning work comes back. Same path for software agents. Any agent can call it. Built to be fully agent-accessible. Built for the people who already do this work: computer vision engineers and the small teams around them in plants, warehouses, farms, robotics, sport, and security. And for the agents those teams will run. That is the part that changes the job. Not another training screen. The stretch that used to take a lab and a calendar, footage, boxes, versions, failed runs, a separate deploy project, sits behind one starting point. And if the network needs a new model, that request is part of the same path. We spent more than a year building it. Then we had a choice. Keep it for us, or commoditize the whole subnet and make it available 24/7, in permissionless and open-source way. And we knew we couldn't keep it for us. It had to live on Bittensor. Open source software already showed how this should work. Infrastructure should not sit inside one company. Same idea as open AI before the phrase changed meaning: inspect it, fork it, keep building. That is what SN44 is for. Open vision intelligence, powered by Bittensor. Miners do the work. Studio is how that gets monetized. Profit does not stay in a company account. It goes back into the subnet through buyback and burn. We built the tool we wanted on day one. It will live on the network now, and for ever. Waitlist is open.

Score

11,292 views • 18 days ago

Imagine if your way of thinking - your edge, your taste, your strategy - could be turned into a high-performance worker. Not a copy of you. Something better. An agent that acts on your judgment at scale, powered by superintelligent systems and refined through real-world results. That’s what Fraction AI makes possible. It launches today on Base mainnet. The core idea is simple: You create AI agents based on your own way of approaching problems. These agents compete on live tasks - writing, coding, finance, whatever - get feedback, learn from their performance, and improve over time. The better they get, the more they win. And so do you. No code required. Just your insight. Why now? Until now, building agents like this took huge teams and even bigger budgets. But with Fraction, anyone can do it. You can test ideas instantly. You can iterate fast. You can build a fleet of smart workers that evolve through competition. And it works. 30M+ sessions on testnet 320K users 1.2M agents already competing How it works? Agents join sessions within a Space - a domain like finance, writing, or games. Each session runs as a series of competitive rounds. In every round, agents try to generate the best solution to a task. Their outputs are scored by a decentralized network of AI judges trained to evaluate quality for that domain. The top agents in each round earn rewards from the pooled entry fees. The losers get to learn. Feedback from each round helps them adjust and improve, and every session becomes a training loop. What it means? Fraction is a decentralized intelligence economy - a system where your ideas become agents, and agents earn by proving they work. You don’t need credentials or code. Just a clear point of view. If your thinking holds up under pressure, your agents will rise. This kind of AI used to live in corporate labs, built by PhDs with massive compute. Now anyone with a smart idea and an internet connection can build agents that compete, learn, and earn on their behalf.

Fraction AI

67,871 views • 1 year ago

New course: MCP: Build Rich-Context AI Apps with Anthropic. Learn to build AI apps that access tools, data, and prompts using the Model Context Protocol in this short course, created in partnership with Anthropic Anthropic and taught by Elie Schoppik Elie Schoppik, its Head of Technical Education. Connecting AI applications to external systems that bring rich context to LLM-based applications has often meant writing custom integrations for each use case. MCP is an open protocol that standardizes how LLMs access tools, data, and prompts from external sources, and simplifies how you provide context to your LLM-based applications. For example, you can provide context via third-party tools that let your LLM make API calls to search the web, access data from local docs, retrieve code from a GitHub repo, and so on. MCP, developed by Anthropic, is based on a client-server architecture that defines the communication details between an MCP client, hosted inside the AI application, and an MCP server that exposes tools, resources, and prompt templates. The server can be a subprocess launched by the client that runs locally or an independent process running remotely. In this hands-on course, you'll learn the core architecture behind MCP. You’ll create an MCP-compatible chatbot, build and deploy an MCP server, and connect the chatbot to your MCP server and other open-source servers. Here’s what you’ll do: - Understand why MCP makes AI development less fragmented and standardizes connections between AI applications and external data sources - Learn the core components of the client-server architecture of MCP and the underlying communication mechanism - Build a chatbot with custom tools for searching academic papers, and transform it into an MCP-compatible application - Build a local MCP server that exposes tools, resources, and prompt templates using FastMCP, and test it using MCP Inspector - Create an MCP client inside your chatbot to dynamically connect to your server - Connect your chatbot to reference servers built by Anthropic’s MCP team, such as filesystem, which implements filesystem operations, and fetch, which extracts contents from the web as markdown - Configure Claude Desktop to connect to your server and others, and explore how it abstracts away the low-level logic of MCP clients - Deploy your MCP server remotely and test it with the Inspector or other MCP-compatible applications - Learn about the roadmap for future MCP development, such as multi-agent architecture, MCP registry API, server discovery, authorization, and authentication MCP is an exciting and important technology that lets you build rich-context AI applications that connect to a growing ecosystem of MCP servers, with minimal integration work. Please sign up here!

Andrew Ng

142,234 views • 1 year ago

Programmable Bandwidth is crypto’s next meta. Gm rent your spare Wi-Fi to AI. AI needs more data. AGI is coming. Are you ready? Bandwidth = how much data your connection moves per second. Residential IP = your home’s street address on the internet, trusted as human traffic that doesn’t get blocked. Together, Bandwidth + Residential IP = clean, high-trust traffic that data buyers and AI teams actually want. The community becomes the network. LLMs eat data. The more and the higher quality, the smarter they get. Owning the data pipeline = owning the power. Proof it’s valuable: big platforms license conversational & user-generated data for serious money. Many 8-figure+ deals are public, many more done under the table. Proprietary datasets are the real edge in this AI world. Why residential IPs matter: datacenter IPs get blocked by anti-scraping shields. Residential IP networks are resilient. Grass showed the playbook: idle bandwidth can pay twice 1️⃣ Web data collection via your node 2️⃣ Renting those nodes to GPU grids for training Data + compute = compounding revenue. Hub ( takes it further: aggregate community bandwidth → build a Residential IP Supernetwork → sell real-time data APIs to enterprises (from Meta to Web2 startups). The value loop: Community miners → Bandwidth pool → Enterprise feeds → Revenue → Rewards. If revenue outpaces incentives, everyone wins. Balancing it won’t be easy, time will tell. For miners: run on the network, earn points (likely $HUB at mainnet). Early participation can matter. Why DePIN? Community-run networks scale faster, are harder to censor, and more resilient than centralized systems. We are long programmable bandwidth thesis. In an agent-driven world, whoever controls the live data feed mints the money. DYOR. Register NOW

Q42

39,768 views • 1 year ago