Will Whitney

So sick! Better world models are going to unlock so many use cases like this

Environment: “34th Street–Penn Station” Character: “Discarded pack of cigarettes” Genie 3:

Very excited for people to be able to play with Genie! Seeing something you worked on in the wild is so different from playing with it by yourself

I got early access to Project Genie from @GoogleDeepMind ✨ It's unlike any realtime world model I've tried - you generate a scene from text or a photo, and then design the character who gets to explore it. I tested dozens of prompts. Here are the standout features 👇

Come work with Joe and you’ll sit by me, @poolio, @jon_barron, @RuiqiGao, @holynski_, @tkipf, and many other fine folks!

@soumithchintala Congratulations on all you've accomplished so far! You did a lot to create a community around FAIR and NYU in the early days of my PhD, and it was a unique experience to watch you and the team hack on early PyTorch.

Leaving Meta and PyTorch I'm stepping down from PyTorch and leaving Meta on November 17th. tl;dr: Didn't want to be doing PyTorch forever, seemed like the perfect time to transition right after I got back from a long leave and the project built itself around me. Eleven years at Meta. Nearly all my professional life. Making many friends for life. Almost eight years leading PyTorch, taking it from nothing to 90%+ adoption in AI. Walking away from this was one of the hardest things I've ever done. But I'm leaving with a full heart. PyTorch handles exascale training now. It powers foundation models that are redefining intelligence. It's in production at virtually every major AI company. It's taught in classrooms from MIT to rural India. The tools I dreamed about making accessible? They are. The barrier to entry I wanted to lower? It's almost gone. To be clear, there’s so much more to do. As long as AI evolves at a breakneck pace, PyTorch will continue to play catch up. Obsessing over the yet-to-come sometimes makes us forget how much we’ve already done. To everyone who built this with me—who believed research should be joyful, that tools should be elegant, that open source changes everything—thank you. This wasn't my journey. It was ours. What's next for me? Something small. Something new. Something I don't fully understand yet. Something uncomfortable. I could have moved to something else inside Meta. But I needed to know what's out there. I needed to do something small again. I couldn't live with the counterfactual regret of never trying something outside Meta. It's very hard to leave. I probably have one of the AI industry’s most leveraged seats, I lead the software layer that powers the entire AI industry. Every major AI company and hardware vendor are on a speed dial. This kind of power is really hard to give up. But curiosity ultimately won out in my head. Keep making AI delicious and accessible. I'll be watching. Probably filing issues. Definitely staying involved. Is PyTorch going to be okay? I don't want to be doing PyTorch forever. I don't want to be like Guido or Linus— bound to a single thing for decades. Last November, coinciding with the birth of my daughter, I started planning my exit with Aparna. My goal was to leave PyTorch in a good and stable place. By this August, during the second half of my parental leave, I knew: Edward, Suo, Alban, Greg, John, Joe and Jana were ready. The team faced hard people, product, technical and organizational problems and didn’t feel the need to lean back on me to solve these for them (unlike in the past). The product story they crafted for the PyTorch Conference was coherent—really coherent. The things I'd flagged red were turning healthy. The project didn't need me anymore. Unlike 2020-2022 (when I stepped down to go do robotics and came back when Lin, Dima and Dwarak left), I have strong confidence that this time PyTorch is truly resilient. The most aligned culture carriers of PyTorch – Greg, Alban, Ed, Jason and Joe are at the decision table now, and people with strong value alignment – Suo, John and Jana have joined them at the table. And there’s a long list of equally value-aligned people willing to sit at the table should any of these people leave. There are many little things that make up my confidence on the people – John worked on Julia and open-source for a very long time (in fact we hacked a Torch.jl in 2015), Suo has been the strongest systems builder and strategic partner I’ve had for the past two years, and Jana worked on resilient core systems for a very long time, I’ve had long technical and organizational discussions with her over the past few months that give me confidence. And the product lineup and execution in 2025 should be sufficient evidence for any remaining doubt. I’m confident that this band of PyTorchers are going to do exceptionally well. PyTorch might change in flavor because I no longer impose my own taste from the top, but I’m confident that the values are going to stay intact and the product is going to be awesome. My time at Meta The early years of FAIR were absolutely magical. I was part of a small family of absolutely brilliant people building state-of-the-art AI out in the open. From working on GANs with Emily Denton, Rob Fergus, Leon Bottou, Martin Arjovsky and the (now legendary) Alec Radford to building Starcraft bots with Gabriel Synnaeve, to building the first FAIR Cluster with Howard Mansell, to working on object detection with Adam Lerer and Piotr Dollar, to building PyTorch. It was more fun than I can describe in words. 2015 and 2016 were probably the most productive and professionally enjoyable years of my life. I’ll probably romanticize this period of my life forever. When I joined FAIR, I had massive impostor syndrome, and the first 3 months were very very difficult. I can’t credit Andrew Tulloch enough for being the most thoughtful, kind and welcoming mentor, without whom I wouldn’t have made it. I’m so damn bullish for Meta just from the fact that he’s back. --- My time on PyTorch was special. I loved every part of building it—designing it, managing it, being the PM, TL, comms lead, doc engineer, release engineer, squashing bugs, growth hacking, turning it into a coherent product with hundreds of people, transitioning it to industry stakeholdership – the whole nine yards. To the core PyTorch team at Meta: the engineers, researchers, open-source maintainers, docs writers, CI infrastructure folks, hardware partners, the community builders. To the hundreds more inside and outside Meta—thank you. You turned a library into a movement. There are too many people to credit and thank, but I can't not mention Adam Paszke, Sam Gross, Greg Chanan, Joe Spisak, Alban Desmaison, Edward Yang, Richard Zou, Tongzhou Wang, Francisco Massa, Luca Antiga, Andreas Köpf, Zach DeVito, Zeming Lin, Adam Lerer, Howard Mansell and Natalia Gimelshein. And Schrep. They made the launch happen. And so many more people became centrally important later: Lu Fang, Xiaodong Wang, Junjie Bai, Nikita Shulga, Horace He, Mark Saroufim, Jason Ansel, Dmytro Dzhulgakov, Yangqing Jia, Geeta Chauhan, Will Constable, Briah Hirsh, Jane Xu, Mario Lezcano, Piotr Balecki, Yinghai Lu, Less Wright, Andrew Tulloch, Bruce Lin, Woo Kim, Helen Suk, Chris Gottbrath, Peng Wu, Joe Isaacson, Eli Uriegas, Tristan Rice, Yanan Cao, Elias Ellison, Animesh Jain, Peter Noordhuis, Tianyu Liu, Yifu Wang, Lin Qiao and hundreds more. It’s criminal of me to not take the space to list out everyone else I should be mentioning here. PyTorch is nothing without its people ❤️. The most joyful moments of building PyTorch was meeting users eager to share their happiness, love and feedback. I remember a grad student coming to me at Neurips 2017, in a slurring emotional voice he said he’d been trying to make progress on his research for 3 years but within 3 months of using PyTorch he made so much progress that he was ready to graduate. That moment made it tangible that what we do matters, a lot, to a lot of people, even if you don't constantly hear from them. I do miss the intimacy of the PyTorch community, with a 300 person conference that felt like an extended family gathering, but I feel that’s a small price to pay considering the scale of impact PyTorch is truly having today – yes the Conference is now 3,000 people where market-moving deals get brokered, but it’s helping orders of magnitude more people to do their best AI work. I miss the intimacy, but I'm proud of that growth. --- To Mark Zuckerberg and Mike Schroepfer, who believed that open-sourcing is fundamentally important and is a sound business strategy. This is so hard to understand for most people within the course of business, but we’ve run lock-step on this strategy without ever having to discuss it. Without you two, neither FAIR nor PyTorch would’ve happened. And those mean so much to me. To Yann LeCun and Rob Fergus, for building the magical early FAIR that I so revere. To Aparna Ramani, a leader that I find so rare at Meta in her ability to hold a really high bar for the org, technically brilliant with the span to discuss deep infra systems and industry-strategy within the same conversation and for being an absolute execution-machine! I’ve learned so much from you. To Santosh, Kaushik, Delia, Oldham and Ben for being so welcoming to Infra. For someone coming over from FAIR with a wildly different culture, you all made me feel at home and made me part of the family, and thank you for that. To all my managers who've championed me through the PSC video game – Serkan, Howard, Jerome, Abhijit, Yoram, Joelle, Aparna and Damien – I owe you a lifetime of drinks. --- Signing off for now. —Soumith

@E0M Not having to go around the world to get to a policy would be great

@E0M My bet is still on learning a world model from plentiful passive data before sprinkling in some actions, which is why I’m working on Genie. But I’d love to be proven wrong

LLMs talk about compute as the "oil" of modern AI. In AI robotics, it's not compute, it's data. Data is the limiting factor to general purpose robotic intelligence.

GEN-0 models exhibit strong scaling laws, in which more pretraining data and compute consistently (and predictably) improve downstream post-training performance of the model across many tasks. 🔬📈🧠

@sedielem That is more my take. Which leads to good questions! How do we increase distribution modeling capacity? Improve the generative process? Or are images / videos inescapably an eg 1T parameter problem?

@wfwhitney One perspective on guidance is that it allows you to cover up underfitting, provided that you don't care too much about diversity😁 The popularity of GANs before diffusion models took over also indicates that most of us are very happy to make this trade!

I contemplated whether I should post this, because it seems kind of obvious. But it's often taken for granted, so we might underestimate the impact: e.g. these days, diffusion papers don't usually show samples without guidance anymore (figures from GLIDE arxiv.org/abs/2112.10741)

Every smart home product except lights would be better as a manual product operated by a robot

@AlexGDimakis And more generally labeling data for one model with other learned models has been very successful

@AlexGDimakis It’s useful as a UI for humans too, and that’s the first place I’d expect this to show up. But say you want to train an agent in e.g. a Genie environment. It’s already common to use VLMs for success detection in robotics, where we also don’t have a ground truth reward

The future of AI is models that generate graphical interfaces. Instead of the linear, low-bandwidth metaphor of conversation, models will represent themselves to us as computers: rich visuals, direct manipulation, and instant feedback.

@scychan_brains I guess an interesting question is what the closest thing to a battery might be. The most general-purpose artifact you can spend your ephemeral compute on to make it widely valuable later. Large model pretraining seems like a good candidate

@wfwhitney Ah yes, computation itself is a different beast (I guess I failed to be clear that I was talking more about the other two.. tweeting is hard 🥲), but yeah totally agree with you that it's ephemeral!

Before money, human societies tended to use credit/debit (and minimal bartering). But most ended up developing money 💵 What's the next step in this evolution? Could compute take the place of money, in the age of AGI? Here's a mini-analysis based on definitions of money and historical examples: scychan.github.io/2025/07/27/com… 🧵👇

@scychan_brains Computation is when you bring together hardware and power at the same time in the right way and get outputs. Interestingly you can't "store" this at all, e.g. save up your datacenter for a month and then spend it all at once

@scychan_brains Yes, that's the durability row. I'm distinguishing computation from e.g. GPUs or credits: - GPUs / datacenters are only the capital part of compute. Ongoing costs (e.g. power) are just as significant and scale differently - Credits aren't new, they're just scrip like gift cards

@scychan_brains It’s more like electricity on the grid in that it must be created and consumed at the same moment

@scychan_brains Also though, money is nonperishable. You can collect money today and spend it next week, and that timeshifting is crazy valuable. Compute doesn’t really have that; you can’t pull out one you made earlier