The vibe of GPT-4.5 is truly something else. It’s deeper, richer, understands better, and shows impressive metacognition—qualities benchmarks simply can’t capture. I’m genuinely excited to see what happens when this model becomes an reasoning model. It’s the perfect match for
RESEARCH
-

Amazing MLFF Tool with Long Range Interactions
By
–
Amazing work by @maxxxzdn . A great tool to build MLFFs with long range interactions! https://t.co/uJMiWgILpC
— Max Welling (@wellingmax) 5 mars 2025Amazing work by @maxxxzdn . A great tool to build MLFFs with long range interactions!
-

Meta PARTNR Demo: Insights and Challenges in Human-Robot Collaboration
By
–
By sharing some of the insights and challenges in developing the Meta PARTNR demo, we hope to contribute to the development of the next wave of innovation in human-robot collaboration. Research paper https://
go.fb.me/prlg14
Dataset and code https://
go.fb.me/k9dzc0 -

Seeing Like a Language Model: Long-Form Essay Becomes Book
By
–
This quarter I took a break from writing my Friday column on @every to write a long-form essay I'm still not done, and it's accidentally a book. Current title is "Seeing Like a Language Model" I'm hoping to have a full draft done by the end of the month! I'm pretty psyched
-

SWE-Search: Monte Carlo Tree Search for Software Agents
By
–
SWE-Search: Enhancing Software Agents with Monte Carlo Tree Search and Iterative Refinement Antoniades et al.: https://
arxiv.org/abs/2410.20285 #ArtificialIntelligence #DeepLearning #MachineLearning -
MIT Introductory Deep Learning Course Now Available Online
By
–
An introductory course about deep learning, courtesy of MIT: https://t.co/xlVPQUthXI
— MIT CSAIL (@MIT_CSAIL) 5 mars 2025
v/@MITDeepLearning pic.twitter.com/wXCs1gD313An introductory course about deep learning, courtesy of MIT: https://
bit.ly/4in6rsJ v/
@MITDeepLearning -
AGI Hype Overlooks Basic Reinforcement Learning Gaps
By
–
Amidst all the OMG AGI IS HERE hype, it is amusing to reflect on how some of the most basic continuous online reinforcement learning problems that trivial organisms solve remain unworkable for today’s systems.
— John Carmack (@ID_AA_Carmack) 5 mars 2025
Happy to be working with Rich! https://t.co/7VKpj6wmm2Amidst all the OMG AGI IS HERE hype, it is amusing to reflect on how some of the most basic continuous online reinforcement learning problems that trivial organisms solve remain unworkable for today’s systems. Happy to be working with Rich!
-
Exciting inference-scaling techniques and future AI developments
By
–
there are lots of cool inference-scaling techniques out there! More to come in my next Ahead of AI post
-
Sutton Barto Win Turing Award for Reinforcement Learning Breakthroughs
By
–
Congratulations to @RichardSSutton & Andrew Barto for receiving this year's @TheOfficialACM A. M. Turing Award for their ground breaking work on the algorithms & foundations of reinforcement learning. RL is at the heart of many of the most exciting advances in AI today! 🎉 https://t.co/7EEMlVJu8g
— Jeff Dean (@JeffDean) 5 mars 2025Congratulations to @RichardSSutton & Andrew Barto for receiving this year's @TheOfficialACM A. M. Turing Award for their ground breaking work on the algorithms & foundations of reinforcement learning. RL is at the heart of many of the most exciting advances in AI today!
-

Key Insights from Annual Machine Learning Competitions Report
By
–
Every Spring, I'm excited to read through the comprehensive ~100-page "State of Machine Learning Competitions" report, which offers many interesting insights into current trends, useful tools, and emerging methodologies in the field. Below are some key takeaways from the latest
