Latent reasoning for autonomous driving makes a lot of sense, text-based chain of thought is way too slow for real-time decisions. Super interesting direction!
RESEARCH
-
NVIDIA Releases Nemotron 3 Super 120B Models
By
–
NVIDIA published a series of new "Nemotron 3 Super 120B" open-source models on Huggingface, designed for agentic use.
-
Invitation to AIE Europe: JEPA Reintroduction and AMI’s AI Future Bets
By
–
do I know anyone who knows @amilabs well? wanted to invite @lxbrun
, @pascalefung
, @sainingxie or ofc @ylecun to AIE Europe (Apr 9) to do a "TED talk"-like reintroduction of JEPA and what the strong bets of AMI are for the next stage in AI! (we dont normally invite speakers in -
Open Source Strategy for Diffusing Model Capabilities
By
–
Good! Models capabilities are always going to diffuse, so better to build them open source as part of architecture.
-
On-Policy Self-Distillation Improves Reasoning LLM Efficiency
By
–
“On-Policy Self-Distillation for Reasoning Compression” This paper shows that the new bottleneck for reasoning LLMs isn’t “too little reasoning”, it’s that more reasoning tokens often increase mistakes. So they fixed it without RL length penalties or external verifiers, but all
-
Continual Learning Solved: The Hardware Imperative for AI Quality
By
–
another ai pin? no, this one’s for your pet
-
Classical hyperparameter tuning outperforms AI coding agent
By
–
I wrote something about the recent hype about automatic AI research. tdlr: at least this time, a classical hyperparameter tuning beat the fancy AI coding agent Ravid Shwartz Ziv (@ziv_ravid) x.com/i/article/203175290702… — https://nitter.net/ziv_ravid/status/2031790304123326532#m
→ View original post on X — @hugo_larochelle, 2026-03-11 21:57 UTC
-
Harvey AI Launches BLB Research Benchmark for Legal AI
By
–
We partnered with @Harvey_AI on BLB: Research — a benchmark of hard, agentic US case law research problems. We identified a series of US case law research problems that leading models are currently unable to solve — even when provided with search tools like web search. The
-
Maximizing Innovation Hub Experience at SAS Innovate 2026
By
–
You’re a leader with industry knowledge, keen expertise, and time to wisely expend. The Innovation Hub is the place to be. How will you make the most of this experience at #SASInnovate 2026?