Three years ago, our founders set out on a mission to build a metadata-free future and bring the full power of analytics to images and video data. Today, we’re thrilled to announce that we’re accelerating a major milestone toward this goal with $30M in Series B funding! 🎉

🧵As a mech interp hater (😅), even I have to admit, this demo is cool! Replace 'Golden Gate Bridge' with a sensitive attribute like race/gender, or a medical concept, and it is clear why this is powerful--if it is ever reliable.

Please ignore anything coming from .Percy Liang account, in-progress of getting the account back And don't click on any of the links!

In which I try to make sense of most of machine learning: blog.alexalemi.com/kl-is-all-you-…

I finally got back into my Twitter account...so that's what I get for not having 2FA.🤦 Someone commented how the scam Tweet wasn't all that compelling, and someone else found that fortunately ChatGPT can help! Safety is hard...

Meet Cyrus, Jocelyn, and Thay 📽️ Their Knight-Hennessy scholar journeys look different, but are bound by a common thread: a desire to make a lasting impact through their pursuit of #highered at Stanford University. Hope you’re inspired by their stories✨Full video: bit.ly/3VwalGO

I’m joining the Columbia Computer Science faculty as an assistant professor in fall 2025, and hiring my first students this upcoming cycle!! There’s so much to understand and improve in neural systems that learn from language — come tackle this with me!

✨ Introducing 𝐎𝐩𝐞𝐧𝐕𝐋𝐀 — an open-source vision-language-action model for robotics! 👐 - SOTA generalist policy - 7B params - outperforms Octo, RT-2-X on zero-shot evals 🦾 - trained on 970k episodes from OpenX dataset 🤖 - fully open: model/code/data all online 🤗 🧵👇

I’m disgusted by the demonization of DEI by tech bros. Unless you’re intentional abt diversity, the natural forces of your referral n/ws leads to ppl hiring more folks who look like them. You have to go out of your way to find ppl who are talented but don’t have access.That’s DEI

arxiv.org/abs/2407.08351 LM performance on existing benchmarks is highly correlated. How do we build novel benchmarks that reveal previously unknown trends? We propose AutoBencher: it casts benchmark creation as an optimization problem with a novelty term in the objective.

Super excited to share our work on LOTUS, a query engine for reasoning over large corpuses of data with LLMs! Joint work w/ the amazing Sid Jha, Matei Zaharia & Carlos Guestrin Read the paper: arxiv.org/abs/2407.11418 Try out the code: github.com/stanford-futur… 🧵👇

i always prefer to rebase because what can be should be unburdened by what has been

Nitarshan Rajkumar Nicolas Moës Jeffrey Ladish Neel Guha Jessica Newman Tobin South Alex Pentland Sanmi Koyejo Mykel Kochenderfer Robert Trager And here’s the full paper link: arxiv.org/abs/2407.14981 (for some reason, the shortened link in the first post didn’t show up as preview 🥲)

I'm joining Cornell University this fall as an Assistant Professor of Computer Science! Looking forward to work with students and colleagues Cornell Computer Science, @cornellCIS on generative models, controllable generation, and creative applications like #musictechnology

Language models keep improving, but not for everyone. Sang T. Truong and Sanmi Koyejo are featured in the The New York Times article nytimes.com/2024/07/26/tec…, discussing our collaboration with Unlimited Research Group of AI on language models for Vietnamese and our thoughts on global language gap & interventions

Tonight, I'm flying out to attend #ACL2024NLP! Come to Poster Session 4 on August 13 at 10:30AM to discuss our paper 💥Mission: Impossible Language Models💥 I'm also happy to just make a new friend or be your personal photographer at your poster session! See you in Bangkok 🇹🇭

Check out my new USC Robotics blog post that takes a retrospective look at the development and impact of pyribs! rasc.usc.edu/blog/pyribs-ac…

Super excited to announce what we have been working on in the last six months - Agent Q is out now! This is a framework for self-supervised agent reasoning and search that can self-correct and autonomously improve by self-play and RL on real tasks on the real internet! 👇

if anyone saw me repost elon musk… know now and forever that was a mistake