To my friends at FAIR and anyone who might be impacted by layoffs, please stay strong. The AI world out there is exciting and dynamic. FAIR was great, but all good things must come to an end. Happy to talk anytime and share leads for interesting opportunities. Get in touch!

I've been thinking about doing this for a while. And no better time than now in the aftermath of the H1B issue. Indian AI startup landscape is exploding. For anyone looking for a job at some of the best companies, I'd love put my network to work. - We've curated 20 amazing

📢Life is a sequence of bets – and I’ve picked my next: @MyolabAI It’s incredibly ambitious, comes with high risk, & carries unbounded potential. But it’s a version of the #future I deeply believe in. I believe: ➡️AI will align strongly with humanity - coz it maximizes its own

Excited to share my project from FAIR! As a side quest to 3D localization, we also ended up developing 3D-JEPA, the first large scale SSL method for 3D. I'm optimistic 3D-JEPA and Locate3D will play a major role in robotics and improving spatial AI. Try the model for your apps!

Really exciting VLA demo from @Figure_robot ! The hericahical architecture they use here is very similar to Latent Code as Bridges (LCB) from @YideShentu et. al. Cool to see them scale this type approach scaled to such success! https://t.co/K9LpQc4tFo 🧵

Interested in knowing the difference between "world" models and "word" models? Come talk to us at the #OpenEQA poster today at #CVPR (Arch 4A-E Poster #179) https://t.co/9EC3coHeJd

New work with @YideShentu! We present Latent Code as Bridges🌉 or LCB. We leverage LLMs for robotics by finetuning the LLM to output latent embeddings to control the lower level policy. This hierarchical approach enables end2end training while maintaining the advantages of LLMs!

Wouldn't it be cool to just talk to your robot 🤖 or smart glass agent 🕶️ in natural language? ✅ Hey 🕶️ do you remember where I left my badge? ✅ Hey 🤖 I'm hungry and driving back from work, do I have any fruits left? We're (@AIatMeta) working on this Embodied Question

Today we’re releasing OpenEQA — the Open-Vocabulary Embodied Question Answering Benchmark. It measures an AI agent’s understanding of physical environments by probing it with open vocabulary questions like “Where did I leave my badge?” More details ➡️ https://t.co/vBFG7Z58kQ

🚀 Exciting news! The Cortex team at #FAIR is looking for #PostDocs to join our research journey! Our mission is to create foundation models for all manner of Embodied AI agents - think robots 🤖, smart glasses 🕶️, and beyond! 🤝 Our commitment? Open-source innovation! Just like

Sad to see the saga unfold at @OpenAI, I was one of the early interns, and still have several friends and mentors there. I'll resist the urge to make elevator pitches for different labs at this stage. Folks, let the poor souls catch a break! Crisis calls for judgement and

🚀 Excited to announce the open-sourcing of #𝗥𝗼𝗯𝗼𝗛𝗶𝘃𝗲! It brings together numerous strands of work from talented collaborators over the years! ✅ Dexterous hands ✅ Quadrupeds on randomized surfaces ✅ Musculoskeletal models ✅ Visually rich domains ✅ Large

📢#𝗥𝗼𝗯𝗼𝗛𝗶𝘃𝗲 - a unified robot learning framework ✅Designed for genralizn first robot-learning era ✅Diverse (500 envs, 8 domain) ✅Single flag for Sim<>Real ✅TeleOper Support ✅Multi-(Skill x Task) realworld dataset ✅pip install robohive https://t.co/CWSipPv8M7 🧵👇

🤖 Excited to share MoDem-v2, our latest work on decoder-free (JEPA-style) world models. 💡 Provides a simple three-step recipe to extend world model-style algorithms (e.g., TD-MPC) to real-world robots. 🙌 Fun collaboration with @palanc_, @ncklashansen, and @Vikashplus. Thanks

#𝗥𝗼𝗯𝗼𝗔𝗴𝗲𝗻𝘁 -- A universal multi-task agent on a data-budget 💪 with 12 non-trivial skills 💪 can generalize them across 38 tasks 💪& 100s of novel scenarios! 🌐 https://t.co/37YbJP05Zt w/ @mangahomanga @jdvakil @m0hitsharma, Abhinav Gupta, @shubhtuls

LfS revisited for control: ✅ A well-tuned LfS baseline is competitive with pre-trained visual rep. (PVR) ✅ But fine-tuning with aug improves PVR results ✅ And better benchmarks are needed! 📢 Tue 2-3.30pm HST, Hall-1 #711 https://t.co/5yv1UjuldD

🌟 MTM is a new SSL paradigm for decision making 🚀 Builds on decision transformer, BERT, and MAE 🔥 A single MTM transformer is versatile and can function as a world model, policy, representation, and more! 🔍🧵 📢 Thu 1.30-3pm HST, Hall-1 #402 https://t.co/Au200XqpcX

✈️ Just landed in Hawaii 🌴 to present two cool projects at #ICML2023 🚀 Masked Trajectory Models (w/ @philippswu, @arjunmajum, @kevinleestone, @yixin_lin_ , @IMordatch, @pabbeel) 📚 LfS Revisited (w/ @ncklashansen, @haosu_twitr, @HarryXu12, @xiaolonw et al.) Details in 🧵👇

Excited to share our #ICML2023 paper ✨LIV✨! Extending VIP, LIV is at once a pre-training, fine-tuning, and (zero-shot!) multi-modal reward method for (real-world!) language-conditioned robotic control. Project: https://t.co/HZuZnKOqPL Code & Model: https://t.co/J30MOankDW 🧵: