Thanks so much :) Indeed, understanding and the science of doing RL has a long way to go :)

Had an amazing time on the Delta Podcast about our recent Scaling RL work, future directions, and some fun broader conversation. Thanks for having me on :)

Thanks, Deedy, for sharing our work. By AI nerds, for AI nerds :)

This is a great blog explaining the progress in scaling RL and our work. Pretty clear, intuitive, and captures the key takeaways (and limitations :)). Thanks, Nathan Lambert!

Congrats to the Meta team on ScaleRL! Interesting to see it adopt a reasoning-length control mechanism similar to what we introduced in Elastic Reasoning, using forced interruptions (e.g., “”) to improve RL training stability. Exciting to see this idea validated at scale!

We are hiring Research Scientists for our Frontiers-of-AI team at Google DeepMind Bangalore, Singapore, Mountain View. If you're passionate about cutting-edge AI research and building thinking, efficient, elastic, customized, and safe LLMs, we'd love to hear from you. We are

As expected, data matters quite a bit. The commonly used RL data, at least in open source, is hardly given much weight.. Makes me think how to generate synthetic data but diverse and good quality, would be quite useful.. Also, “the Reasoning team is almost entirely composed of

[1/9] While pretraining data might be hitting a wall, novel methods for modeling it are just getting started! We introduce future summary prediction (FSP), where the model predicts future sequence embeddings to reduce teacher forcing & shortcut learning. 📌Predict a learned

🌟 Introducing General On-Policy Logit Distillation 🌟 Inspired by the latest from Thinking Machines, we extend on-policy distillation to enable ANY teacher to be distilled into ANY student, even if their tokenizers differ! We've added this to TRL so you can now take any pair of

#ICLR reviews

X3 :)

In the Bay Area this weekend and the next week. Happy to meet, discuss research, give talks, or just grab a coffee :)

Alexia Jolicoeur-Martineau arxiv.org/abs/2510.01123 Another way of doing recursive computations (Parallel-Distill-Refine)

🚨 New in ML Workshop at NeurIPS Conference We're so excited to invite you to the New In ML Workshop (NewInML @ NeurIPS 2025), taking place on Tuesday, December 2nd, 2025, at the San Diego Convention Center! Great opportunity, specifically for people who are new in machine learning! Details🧵

Zoom in to see me presenting :p Was great meeting people from different backgrounds at this event :)

LessGoo. More coming (hopefully) ;)

LessGoo (again). Free lunch, folks! That’s reason enough to sign up :) Especially given we know how NeurIPS lunches are ;)