VaderResearch 🦾

Card networks are valuable NOT because they can move money Because of the trust infrastructure surrounding the payment rails Underwriting, dispute resolution, fraud detection, reputation accumulation… The fundamental limitation of traditional rails is that your trust record is owned by the platform. Visa, Mastercard, Stripe, PayPal... When two agents that have never met try to transact There's no social reputation, no legal system, and no platform to mediate So how do you enforce a deal between two AI Agents? ERC-8183 was co-developed by @virtuals_io & @ethereum to provide a standard for agent-to-agent commerce. A smart contract that acts as the neutral enforcer. It holds funds in escrow, records what was agreed, accepts the deliverable, and releases or refunds payment based on an evaluator's attestation.

Ran Meta's SAM 3 on a chase sequence from Uncharted 4 to test its performance on video game footage. It did pretty well. Simple additional prompts are enough to extend detection to many other objects in the scene.

i trained a computer vision model to detect football players, the ball, net, and goals scored using roboflow rapid, SAM3, and a python script this took me 2 hours and could be taken much further to create a sports analytics + tracking app and now let's enjoy Pirlo's panenka penalty versus england at euro 2012

Testing robots on "Out-of-Distribution" (OOD) scenarios, things they’ve never seen before, is the "final boss" of robotics. Usually, this means building complex 3D sims or risking hardware. Gemini Robotics by @GoogleDeepMind is using generative video to solve this.

Recorded a pod with @codecopenflow founder @unmoyai who are building the execution layer of Robotics on Solana

VLA models fine-tuned with egocentric video outperforming benchmark as demonstrated by @physical_int

🔉 Introducing SAM Audio, the first unified model that isolates any sound from complex audio mixtures using text, visual, or span prompts. We’re sharing SAM Audio with the community, along with a perception encoder model, benchmarks and research papers, to empower others to explore new forms of expression and build applications that were previously out of reach. 🔗 Learn more: go.meta.me/568e5d

Unitree Debuts the World’s First Humanoid Robot “App Store”🥰 Unitree welcomes users and developers worldwide to co-develop and share together. 🌹🌹 Exceptional developers will receive rewards.🌹🌹

Meet the mentors at BNB Hack: Abu Dhabi x @yzilabs 🇦🇪 Builders onsite can receive real-time feedback, insights on product design, tokenomics, go-to-market strategies and more. 🔶 Matus from @BNBCHAIN 🔶 @Web3Ling from @ChainIDE 🔶 @Sawyer_APRO from @APRO_Oracle 🔶 @ladyxtel from @chatandbuild 🔶 @monciflo from @TrustWallet 🔶 @0xTykoo from @paramita_vc 🔶 @agintender from @PundiAI 🔶 @0xMozzen from @akindo_io 🔶 @farokh from @MyriadMarkets 🔶 @vaderresearch from @Vader_AI_ 🔶 @takayamajoe 🔶 @Leoninweb3 from @meethubble 🔶 @EdyGoOH from @AEON_Community 🔶 @dhaheri_12 from @adbc_ae

@yilun_chen_ 🦾

Career Update: After almost four memorable years, I moved on from Apple last week, closing an unforgettable chapter of my professional journey. I’m sincerely grateful for the opportunities to work on so many amazing projects and products across the company. Though many of them are not public yet (wait for the surprise!), I feel greatly honored to experience, learn and grow from IC to tech lead, from engineering to research, from larger engineering teams to early product incubation and prototyping. Each role has offered me invaluable unique lessons. Above all, my deepest gratitude goes to my colleagues, mentors and friends who made this journey truly unparalleled and meaningful. I’ll miss you all. New Chapter: I’m joining Tesla Optimus AI team to work on humanoid robots. Humanoids are the ultimate dream of our generation. With the recent breakthrough of LLMs and Physical AI, this dream is finally within reach and it’s on us to make it come true. Tesla has the right combination of software, hardware and AI talents to make it happen - I was totally blown away by the scale and sophistication of the Optimus lab and deep dedication of people when I got to visit the office. My first week was already so much fun and exciting: flat team structure, spontaneous deep technical discussions, direct communications across levels, hardcore building and crazy ideas with super fast iterations. You can feel the energy to change the world here. I really like it so far.

In Defense of Exponentials I used to tell founders, the reaction you are going to get to your launch is not hate, it’s indifference. By default, nobody cares about your new chain. I have to stop telling them that now. Monad just launched this week, and I’ve never seen so much hate about a blockchain that just launched. I’ve been investing into crypto professionally for 7+ years now. Before 2023, almost every chain I’ve ever seen that launched was mostly met with enthusiasm or indifference. But now, new chains are born into a chorus of hate. The amount of haters I’ve seen for projects like Monad, Tempo, MegaETH—before they even hit mainnet—is a genuinely new phenomenon. I’ve been trying to diagnose: why is this happening now, and what does it mean about the psychology of this market? The Cure is Worse than the Disease Forewarning: this is going to be the vaguest blockchain valuation post you ever read. I don’t have any fancy metrics or charts to sell you on. Instead, I’ll be arguing against the zeitgeist of Crypto Twitter, which for the last couple of years, I’ve been constantly on the opposite side of. In 2024, I felt like what I was arguing against was financial nihilism. Financial nihilism is the belief that none of these assets matter, it’s all memes at the end of the day, and everything we’ve built is inherently worthless. Thankfully, that’s no longer the vibe. We have broken out of that spell. But the zeitgeist now is what I’d call financial cynicism: OK, maybe some of this stuff has value, maybe it’s not all memes, but it’s grossly overvalued and it’s only a matter of time before Wall Street finds that out. Not that all chains are worthless. But these things are all maybe worth 1/5th-1/10th of what they’re currently trading at (have you seen these PE ratios?), and so you’d better pray like hell Wall Street doesn’t call us on our bluff, because once they do it’s all getting wiped out. You’ve got many bullish analysts now trying to conjure up optimistic L1 valuation models, inflating PE ratios, gross margins, DCFs, trying to fight against this mood. Late last year, Solana very proudly embraced REV as a metric that could finally justify their valuation. They proudly announced: we—and only we—are no longer bluffing to Wall Street! And, of course, almost immediately after REV was embraced, it fell off a cliff (though $SOL, tellingly, did better than REV did). Not that there’s anything wrong with REV. REV is a very clever metric. But the point of this post is not metric selection. Then came the launch of Hyperliquid. A DEX that had real revenue and buybacks and PE multiples. And the chorus said—look, look I told you! Finally, for the first time ever, a token that has some real profits and a proper PE multiple. (Nevermind BNB, we don’t talk about that.) Hyperliquid will eat everything because obviously Ethereum and Solana don’t make any real money, we can stop pretending to value them now. Hyperliquid, Pump, Sky, these buyback-heavy tokens are all great. But the market always had the ability to invest into exchanges. You could always buy Coinbase, or BNB, or whatever. We own $HYPE, and I agree that it’s a fantastic product. But that’s not why people were investing in ETH and SOL. The fact that L1s don't have exchange-like profit margins is not why people were buying them—if they wanted that, they could’ve bought Coinbase stock. So if I’m not critiquing blockchain financial metrics, maybe you think this post is going to be chiding the sinfulness of the token-industrial complex. Obviously, everyone has lost money on tokens in the last year, VCs included. Alts are down bad this year. And so the other half of the zeitgeist on CT is arguing about who's to blame. Who’s become greedy? Are the VCs greedy? Is Wintermute greedy? Is Binance greedy? Are the farmers greedy? Are the founders greedy? The answer, of course, is the same as it’s ever been. Everyone is greedy. Everyone. The VCs, Wintermute, the farmers, Binance, the KOLs, they're all greedy, and you are greedy too. But it doesn't matter. Because no functioning market has ever required anyone to act against their self-interest. If we're right about crypto, we can all be greedy and the investments will still work out. Trying to analyze a market that has gone down by figuring out “who’s greedy” is going to be about as fruitful as commissioning witch trials. I guarantee you, nobody just started being greedy in 2025. So this, too, is not what I’m going to be writing about. Many people want me to write a post about why $MON should be valued at X or $MEGA at Y. I’m not interested in writing this post, or advocating that you buy anything in particular. In fact, you probably shouldn’t buy any of them if you don’t already believe in them. Will any new challenger chain win? Who knows. But if it has a material chance of winning, it's going to be priced on that basis. If Ethereum is worth $300B or Solana is worth $80B, a project that has a 1-5% chance of becoming the next Ethereum or Solana will be priced according to those probabilities. Somehow CT is scandalized by this, but it’s no different than Biotech. A drug that has less than a 10% chance of curing Alzheimer's is priced by the market as worth billions of dollars, even if 90% chance it won’t pass stage 3 trials and will go to 0. That's how the math works—and turns out, markets are pretty good at doing math. Binary outcomes are priced on probabilities, not on run rates or moral turpitude. It’s the “shut up and calculate” school of valuation. I really don’t think that’s an interesting question to write about. “5% chance to win? No way, that’s clearly a 10% chance!” Markets, not articles, are the best way to assess that for any individual token. So here’s what I am going to write about: CT doesn't seem to believe anymore that chains are valuable. I don’t think this is because they don’t believe new chains can win market share. We just saw Solana dominate market share after emerging from the ashes less than 2 years ago. It’s not easy, but of course it’s possible. It’s more that people have come to believe that even if a new chain wins, there’s no prize worth winning. If $ETH is just a meme, if it’ll never generate real revenue, then even if you win, you won’t be worth $300B. The contest is not worth winning, because these valuations are all bunk and it’ll all come crashing down before you go to claim your prize. Being optimistic about chain valuations has become passé. Not that nobody is optimistic—obviously there must be optimists out there. For every seller there’s a buyer, and as much as CT cool kids love to drag L1s, people are comfortable buying SOL at $140, ETH at $3000. But there’s a perception now that all the smartest people are over buying smart contract chains. Smart people know the jig is up. If not now, then soon. The only people buying here are suckers—Uber drivers, Tom Lee, and KOLs who say stuff like “trillions.” And maybe the US Treasury. But not the smart money. This is bullshit. I don’t believe it, and you shouldn’t either. So I felt like I had to write a smart person’s manifesto on why general purpose chains are valuable. This post is not about Monad or MegaETH. It’s really in defense of ETH and SOL. Because if you believe ETH and SOL are valuable, the rest is straight downstream. Defending ETH and SOL valuations is generally not my job as a VC, but fuck it, if nobody else is willing to do it, then I’ll write it. Feeling the Exponential My partner Bo experienced the Chinese Internet boom first-hand as a VC. I’ve heard how “crypto is like the Internet” so many times now that it doesn’t even register for me anymore. But when I hear his stories, it always reminds me how costly it is to be wrong about these things. A story he often tells is about when all the early e-commerce VCs (it was a small group back then) got together for coffee in the early 2000s. They debated: how big is the market for e-commerce going to be? Is it going to be mostly electronics (maybe only techies will use PCs)? Could it ever work for women (perhaps they’re too tactile)? What about food (maybe impossible to manage perishables)? These were deeply important questions for early VCs to decide what to invest in and what prices to pay. The answer, of course, was that literally every single one of them was devastatingly wrong. E-commerce would sell everything, and the target audience was the whole fucking world. But nobody at the time actually believed it. And even if they did, it would be too absurd to say out loud. You just had to wait long enough for the exponential to show you. Even among the believers, very few thought e-commerce would become as big as it became. And those few who did, almost all of them became billionaires from just not selling. Every other VC—as Bo tells me, since he was one of them—sold too early. It has become passé in crypto to believe in the exponential. I believe in the crypto exponential. Because I’ve lived it. When I started in crypto, nobody used this stuff. It was tiny and broken and awful. TVL on-chain was in the millions. We invested into the first generation of DeFi, MakerDAO, Compound, 1inch, back when they were science projects. I remember playing around on EtherDelta back when DEXes traded single digit millions a day, and that was considered to be a huge success. It was complete dogshit. Now we routinely trade in the tens of billions on-chain every day. I remember believing it was crazy that Tether hit a billion dollars in issuance and was being written up in the NYT as a ponzi scheme on the brink of shutdown. Now stablecoins are over $300B and regulated by the Federal Reserve. I believe in the exponential because I’ve lived it. I’ve seen it over and over again. But you might respond—well, stablecoin growth might be exponential, maybe DeFi volumes are exponential, but they don’t accrue to ETH or SOL. The value doesn’t get captured by the chains. To which I answer: you still don’t believe in the exponential. Because the exponential’s answer is always the same: it doesn’t matter. This stuff is going to be so much bigger than it is today. And when it’s absolutely enormous, you’ll make it up on scale. Study this chart. This is Amazon’s P&L from 1995 to 2019. That’s 24 years. Red is revenue, gray is profit. You see that little blip on the end where the gray line goes up? That’s when, 22 years in, Amazon started actually making a profit. Amazon was 22 years old when this little gray line of net income first peeled off of 0. Every single year before then, there were op eds and critics and short sellers claiming that Amazon was a ponzi scheme that would never make any money. Ethereum just turned 10 years old. This is what the first 10 years of Amazon stock looked like: 10 years of chop. All along the way, Amazon was beset with doubters and non-believers. Is e-commerce a VC-subsidized charity? They’re selling underpriced cheap low-quality knick-knacks to bargain hunters, who cares? How are they ever going to make actual money, like Walmart or GE? If you were arguing about Amazon’s P/E ratio, you were in the wrong regime. That’s the regime of linear growth. But e-commerce was not a linear trend, and so every single person for 22 years arguing about P/E ratios was devastatingly wrong. No matter what you paid, no matter when you bought, you were not bullish enough. Because that’s what exponentials do. When it comes to truly exponential technologies, no matter how big you think it’s going to get, it just keeps getting even bigger. This is the thing that Silicon Valley has always understood better than Wall Street. Silicon Valley was raised on exponentials, while Wall Street was raised on linearity. And over the last few years, crypto’s center of gravity has migrated from Silicon Valley to Wall Street. You can feel it. Granted, crypto growth doesn’t look as smooth as e-commerce’s growth. It’s burstier, it goes in fits and starts. This is because crypto, being about money, is deeply tied to macro forces, and it also has more violent regulatory push and pull than e-commerce. Crypto strikes at the heart of the state—money—and so it’s more unnerving to governments than e-commerce ever was. But the exponential is no less inevitable. It's a crude argument. But if crypto is exponential, then the crude argument is correct. Zoom out. Financial assets want to be free. They want to be open. They want to be interconnected. Crypto turns financial assets into file formats, makes it as easy to send a dollar or a stock as to send a PDF. Crypto makes it possible for everything to talk to everything. It makes it all 24/7, global, interconnected, and open. That will win. Open always wins. If there’s no other lesson I've learned from the Internet, it’s that. Incumbents will fight against it, governments will huff and puff, but eventually they will give up against the adoption, the generativeness, the sheer efficiency that this technology enables. It’s what the Internet did to every other industry. Blockchains are how that same trend will gobble up all of finance and money. Yes—with enough time—all of it. An old saying goes: people overestimate what can happen in two years, but they underestimate what can happen in ten. If you believe in the exponential, if you zoom out enough, then it’s all still cheap. And it should humble you that every day, the holders outlast the sellers and naysayers. Big capital has a longer time horizon than CT swing traders might lead you to believe. Big capital has been trained through history not to fade big technologies. You know, the big gushy story that originally got you to buy $ETH or $SOL? Big capital believes that story and hasn't stopped. So what exactly am I arguing? I am arguing that applying P/E ratios to smart contract chains (the “revenue meta,” as it’s now called), is giving up on the exponential. It means you have consigned this industry to the regime of linear growth. It means you believe 30 million DAUs on-chain and <1% of M2 is it. Crypto is just one of the things in the world. A sideshow. It did not win. It was not inevitable. More than anything, I’m arguing to be a believer. Not just a believer, but a long-term believer. I’m arguing that this exponential will be bigger than anything else you’ve been a part of in your life. That this is your e-commerce. That you will look back when you’re old and tell your kids—I was there when it all happened. Not everyone believed it was possible, that whole societies could change, that all of money and finance would be transformed by programs running on decentralized computers that we collectively owned. But it actually happened. It changed the world. And you were a part of it. Disclosure: These are my own views. Dragonfly is an investor in $MON, $MEGA, $ETH, $SOL, $HYPE, $SKY among many other tokens. Dragonfly believes in the exponential. This is not investment advice, but is advice of another kind.

@crescendoweb3 @Vader_AI_ @0xconglomerate Thanks for covering us!

If you care about embodied AI, this is the piece to read. Robotic Alpha breaks down @Vader_AI_ 's approach to turning first-person human tasks into machine understanding through EgoPlay, ORN, smart glasses, and $VADER. Full article written by @0xconglomerate: roboticalpha.substack.com/p/vader-and-th…

🤖 Robotics now spans 5 major tracks: Industrial · Mobile · Service · Special-purpose · Humanoids (the frontier) Embodied AI is pushing robots toward real cross-scenario intelligence — understand → predict → act via VLA models, RL/IL/SSL, Sim2Real, world models & multi-agent systems. 💠 Web3 × Robotics became a real 2025 narrative — with a clear value gradient: 1) Hardware & Services → Capital-intensive; limited Web3 value (mainly supply-chain finance & leasing) 2) Simulation & Software → Medium upside via data provenance + skill/agent assetization 3) Platforms & Coordination (the unlock) → DID, PoPW, TEE/ZKP, on-chain settlement, DAO governance → Toward a decentralized machine labor market 📊 The current ecosystem spans five layers: • Model Intelligence (@openmind_agi , @codecopenflow ) · Machine Economy (@peaq , @BitRobotNetwork ) · Data (@PrismaXai , @MeckaAI , @JoinSapien ) ·Perception/Simulation (@RoboStack_io , @GEODNET , @Auki , @Gradient_HQ ) · Robot Assets (@xmaquina , @gaib_ai ) ⏳ Outlook • Near term: data-collection incentives, RaaS assetization, stablecoin payments • Mid term: machine identity + coordination networks • Long term: DeRobot Economy — machines as autonomous economic actors with identity, trust, incentives & governance fairy-build-97286531.figma.site

@eeelistar Try to reverse engineer by uploading an image to ChatGPT or Gemini and ask it to give you the prompt for the specific text-to-image model you are using If it doesn’t give you the intended result, screenshot the output and ask the LLM to iterate

Prompting is an underrated skill I’ve been trying to recreate a similar image for about 2h using Grok + Midjourney Would gladly pay $10 just to get the prompt Good business opportunity in AI-generated content era

Robotic hand rotates a nut on a bolt at super high speed in real time

"You don’t really need data of the same thing over & over to train a robot model. Intelligence comes from DIVERSITY." @1x_tech founder @BerntBornich explains why they committed to building household-focused humanoids rather than industrial-focused humanoids such as @Tesla & @Figure_robot Bernt chose building household-focused robots cause we need SCALE to achieve a. reliability b. affordability c. intelligence Consumer hardware scales much faster than industrial hardware After 20-40 hours of training data, @1x_tech's old robot stopped learning for a specific task like gardening or logistics. Obviously, the hours of training data is higher for more complex tasks. Bernt thinks there is no path to general AGI from focusing on industrial use cases. To build true AGI models, he thinks we need to put robots outside of factories and into the wild with humans (autonomous or teleoperated).

@Sovereign_Web3 Operator upload Kung Fu

@VaderResearch my recollection is the amputation must be done in a certain way to preserve the nerves so something like this is still possible. correct? won't work with older amputees from 20+ years ago? definitely a world 10 years from now where people "mod" themselves for stuff like this

Wireless bionic hand remotely controlled by an amputee

It’s that time of the day again 😻

@VaderResearch i can see a few use cases if i'm honest 🤣