I think we are hosting another compute party next week at SF. My job is to show up and have fun. Pin @Midnight_Captl for invite.

Today we’re releasing Underdog Bark 3.0, the #1 frontier speech model It's a tiny 163 MB download as accurate as NVIDIA’s Parakeet (15x bigger). It beats OpenAI's Whisper large-v3 (57x bigger) It's fast- transcribes 1 min of audio in <0.25s on a Macbook. On recorded meetings (AMI), Bark 3.0 gets 7.82% of words wrong. Parakeet gets 9.42%. OpenAI's Whisper large-v3 gets 13.63% error rate Bark runs on both your Laptop & Phone. It’s free and open source under CC-BY 4.0 Try private on-device meeting notes & dictation, powered by Bark, in Underdog today: underdog.ai/bark

Biological intelligence to bootstrap robotic intelligence? If robotics is about to enter its golden age, @MeckaAI is one of the companies feeding it, by sensorizing humans to train robots. The biggest challenge robotics faces is the data flywheel cold start problem. LLMs got a free cold start from the internet: trillions of tokens, already written. Robotics gets nothing. Someone has to go create the training data to jumpstart the system. @joshavata is the co-founder and CEO of Mecka, the data and deployment company for physical AI. Josh read every robotics paper he could find in late 2024, visited the labs, and made a bet: humans and humanoids look the same, so humans could act as the training fleet. Egocentric data, first-person video of people doing real work, could become vital fuel. Two years ago, the biggest egocentric dataset was Ego4D at 4,000 hours. The field is now in the millions of hours, thanks in part to Mecka. We’re probably still orders of magnitude of training data away from robotics nirvana. But every rung on the physical intelligence ladder unlocks real commercial tasks. And eventually, Josh imagines a world with an order of magnitude more robots than people, of a plurality of shapes and sizes and use cases, as the Hitchhiker’s Guide to the Galaxy foretold :) A few things we cover in this conversation. - Why robotics has no cold start: pixels, actions and logic, and why pre-training data is a real market in robotics in a way it never was for LLMs - What “teaching robots about carbon life” means: in sci-fi the final frontier is organic. The human brain runs on 20 watts and self-heals - The egocentric bet: teleop couldn’t scale, humans and humanoids share a form, so Waymo-fy the humans. The first pilot was aunties and uncles recording themselves cooking on a mobile app - Force data: video can’t tell a robot how hard to squeeze an egg. Wrist cameras are a hack. Mecka built the gloves instead - Inside the data factory: a consumer app that streams your tasks, a B2B engine recruiting across “the taxonomy of all physical human work,” 100-gigabit upload lines, hundreds of petabytes - Millions of hours vs. Tesla’s decade on two degrees of freedom, and why it’s a slow creep up, not a GPT-3 moment - The home has the most data and is the hardest place to deploy. Laundry folding and assembly-line pick-and-place get solved in the next year or two - Useful robots already exist: hotel delivery and cleaning bots in China and Southeast Asia, industrial cleaning in the US. Humanoids are the terminal state, not the first step - Deployment diffusion: a modern systems integrator whose installed robot does something next week it couldn’t do last week. If Mecka ever helps deploy a robot that mines an asteroid, they’ve done their job - The data pot of soup: a million hours of human data, tens of thousands of UMI, a few hours of in-domain teleop. Tactile and simulation haven’t entered the mix yet - Move 37 for robotics: AlphaGo went behavior cloning, then self-play, then magic. The catch is the rules of the game are the entire physical world, and we don’t have the simulator yet - Will the AGI labs steamroll robotics? “It takes a lot of courage to scale up data with imperfect information.” - The Kawasaki Corleo, the robot panther-dog-horse on Mecka’s conference room wall, and why the future of robotics looks like Hitchhiker’s Guide to the Galaxy Timestamps: 0:00 – Introduction 1:46 – What Mecka is: pixels, actions and logic 2:45 – Is robot data economically feasible without the internet? 4:15 – Teaching robots about carbon life 5:27 – Cloud kitchens, stablecoins and the path to robotics 7:03 – Entering robotics with a fresh lens 8:45 – What egocentric data is and why Mecka bet on it 10:09 – The first pilot: aunties and uncles with a mobile app 10:42 – Force data and tactile gloves 11:55 – Inside the operations of a data factory 14:41 – Heat-mapping the taxonomy of human work 15:42 – Quality gates, fraud and the McDonald’s POV videos 17:44 – Earning a margin as a data intermediary 19:33 – Millions of hours vs. Tesla’s decade 20:55 – Why it isn’t binary: unlocking commercial tasks along the way 22:02 – Home data is plentiful; home robots are the hardest 23:24 – Useful robots can be deployed today 24:35 – When robotics gets an FSD-style flywheel 26:42 – Deployment diffusion and the modern systems integrator 28:21 – Favorite vertical robot categories; mining an asteroid 30:08 – The hardware slate: Waymo-fying people 30:55 – EgoVerse 32:04 – The data pot of soup: the converged training mix 33:54 – Move 37 for robotics: AlphaGo and Federer 36:02 – World models and the training mix 36:58 – Will scaling the current mix get us to the promised land? 39:05 – Will the AGI labs steamroll robotics? 40:50 – Sci-fi roots and the Kawasaki Corleo 42:44 – How many robots in a decade?

Congratulations, Xona team!

Local inference is the future, so we made local Perplexity. Enjoy!

This is great

You can now run Perplexity Computer fully local on your Mac with no cloud based inference, without WiFi too. Enjoy!

Cancer patients with the same diagnosis often have very different tumor biology. Molecular testing can reveal underlying differences, but that's slow and expensive. Introducing Holon-S1, our platform to predict genes and biomarkers from routine pathology slides. • beats two other models at predicting spatial gene activity in a controlled comparison • has identified potentially novel biomarkers in breast cancer and colon cancer • evaluated across 16 organ types and more than 15 scanners, using slides from both formalin-fixed, paraffin-embedded and fresh-frozen tissue. We believe Holon-S1 has the potential to help researchers find new biomarkers in slides they already have, discover treatment targets, and bring more personalized cancer care within reach!

Kyle Shanahan says a 51% analytics recommendation won’t convince him to go for it because he gets one shot, not an hour at a blackjack table “I don’t want to hear about it if it’s 51% of the time we should go for it.” “As I always say, I’m not sitting here playing Blackjack for an hour with no one else at the table, and eventually I’ll get 51% of it. Like, I have this one try, and I have a feel for it. I’ve been doing it too long.” “But what I want to know from him, when it’s not a feel, when it, no, this is 98%. We don’t even need to talk about it. So we don’t need to talk about it. Tell me when it’s that, and I’ll just do it.”

First multiplayer personal agent - let's go! Congrats @ekuyda and @wabi team!

There are two really interesting bets here: - Agents in human group chats require a new social contract. A dedicated app lets you establish it with agents in the room from day one. - Personal agents need more than a chat interface. Rich mini-apps give people a shared place to plan, play and get things done together. Very cool to see the personal agents space heat up with thoughtful bets in very different directions. Congratulations @ekuyda & team!

This is extremely well crafted. You have to try it out

System 1

Aline started working at age 12. Over the next three decades she built La Barrique, a restaurant in Paris’s 10th arrondissement, from scratch. She cooked, served, and became part of the neighborhood. On October 17, rioters broke in, destroyed the place, punched her in the face, and set the building on fire. By morning her restaurant was ashes. 34 years of work, vaporized in a single night. Obviously, Aline is an immigrant, one who went to work every day, contributed to the economy and served her community. One of liberalism's fatal flaws is the fear of drawing distinctions. Distinctions between men and women. Distinctions between cultures. Distinctions between countries. Because apparently noticing any differences means that you are violating your commitment to equality; it means that you are discriminating, and in the hierarchy of evils in the liberal world, there is no worse sin. This is literally baked in to the French national motto: Liberté, Égalité, Fraternité Égalité means equality, specifically social or political equality. Which means that drawing any distinctions is the same as discrimination. But this is completely untethered from reality because these distinctions are REAL. And if you craft and enact policies that ignore these distinctions, the chickens (ok since this is France, the Gallic rooster) will come home to roost.

Best team in consumer AI, and a radical non-obvious yet totally obvious take on the personal assistant space. Check it out!

proud 🙏🏻

Check out @wabi!

This deserves a RT for comedic value

Larry David is one of America’s unlikeliest icons, and one of history’s purest founders. In our latest profile, Correspondent-at-Large @blakejharrisNYC—who spent six years talking to Larry for his new book—explains what makes Larry so special, and why it took him so long to break through. Larry was a mediocre student with a middling social life. His first two jobs—selling bras and driving cabs—ended quickly. At 25, he called his mother from a hotel room: “Ma, nothing good is ever gonna happen to me.” She urged him to become a mailman. Soon after, he found what he thought would be his calling: stand-up comedy. His material was funny, but he rarely finished a set. If somebody in the audience looked at their watch, or didn’t laugh when he thought they should, Larry stormed off. 14 years later, aged 39, he got a break. He was asked to write his first sitcom script for a big new Showtime series. Only an eighth of what he wrote survived the edit, but the episode was still his. He was broke, and the writing credit could have sparked his career. But Larry refused to put his name on it. Even with Seinfeld, Larry nearly hit self-destruct. The pilot performed worse than almost any other show NBC had ever tested. Still, the network ordered four episodes, one of the smallest orders in TV history. Larry barely made it through them. When the studio installed a showrunner to oversee the scripts, Larry endured the man’s notes, then said “NO. I’m not doing one thing,” and walked out. In the summer of 1991, with the show finally starting to work, NBC wanted Jerry and Elaine to get married and have a baby. Larry quit in the parking lot and wished everyone good luck. Jerry stood with Larry, and NBC caved. Larry’s friend and comedian Bill Scheft put it this way: “He doesn’t [act this way now] because he has fuck-you money. It’s the opposite. He has the fuck-you money because he’s always acted this way.” Read the full profile by Blake, author of Emperor of Nothing: The Making of Larry David. colossus.com/article/many-fa…

Big fan. Photos are a mess with an occasional gold nugget. Well done @marissamayer!