Agents build useful memory during tasks, but that memory is trapped. So the big question is whether a single memory system can be shared across different models. If you want to transfer it to a different model, the performance often gets worse, not better. New research tackles
RESEARCH
-
U.S. Sanctions Law Implications for Academic Peer Review Participation
By
–
Participation in peer review at @NeurIPSConf (or @icmlconf, @iclr_conf, @CVPR, @COLM_conf, etc.) can be considered providing a "service" under U.S. sanctions law. U.S. law generally prohibits providing services to designated sanctioned individuals or entities, including cases where the process is effectively providing a service to a sanctioned institution. Violations can lead to significant fines and compliance overhead; willful violations can carry criminal exposure to the organizers and board members. Note that the "informational materials" exemption (Berman Amendment) likely does not apply here, based on legal advice.
→ View original post on X — @hugo_larochelle, 2026-03-25 19:06 UTC
-
Jensen Huang Discusses Open Frontier Models with AI Leaders
By
–
Different voices. Same answer: open models.
— NVIDIA (@nvidia) 25 mars 2026
NVIDIA Founder and CEO Jensen Huang sat down with the leaders from @mistralai, @bfl_ml, @cursor_ai, @LangChain, @perplexity_ai, @reflection_ai, @thinkymachines, @allen_ai, @evidenceopen, and AMP PBC to discuss the rapid rise of open… pic.twitter.com/wha1gE3W5oDifferent voices. Same answer: open models. NVIDIA Founder and CEO Jensen Huang sat down with the leaders from @mistralai, @bfl_ml, @cursor_ai, @LangChain, @perplexity_ai, @reflection_ai, @thinkymachines, @allen_ai, @evidenceopen, and AMP PBC to discuss the rapid rise of open frontier models. Get the top takeaways from the frontier of AI: nvda.ws/4uR6eEV
-
AGI Definition Unchanged Since 2019 Compass Not Target
By
–
I have been saying the same thing for years. Initially I was using this exact line ("it's a compass, not a target to hit") about ARC 1, back around 2021-2022, before ChatGPT. This has always been our stance. My bar for AGI has been public and unchanged since 2019. It's AGI when
-
AI Model Development Pace Challenges Environment Creation Timeline
By
–
This would have been materially impossible to do even if we wanted to, since it took us an entire year to create the environments, while new AI models come out every month. By the time the latest SotA model is out, the full set of environments already is already shipped.
-
ARC-AGI: Measuring Progress Toward Artificial General Intelligence
By
–
Keep in mind: ARC-AGI is *not* a final exam that you pass to claim AGI. Including ARC-AGI-3. The benchmarks target the residual gap between what's hard for AI and what's easy for humans. It's meant to be a tool to measure AGI progress and to drive researchers towards the most
-
Were ARC-AGI Tasks Dropped Because Models Already Solved Them?
By
–
Just wondering: were any ARC-AGI designs discarded because current models were already too good out of the box?
-
Human-Level General Intelligence Requires Learning Efficiency Parity
By
–
Human-level general intelligence is achieved when an AI system can approach a new task and figure it out, without human intervention, *with the same learning efficiency as humans*. If every new task requires human intervention, it's not general. If every new task requires
-
Mathematician’s Paradox: Djinn, Worlds, and Limbo
By
–
Interesting @exame @exame_noticias is also using my name and image to sell AI courses, seems many do it now, it's kinda flattering but also they never asked permission Do you think I should do something about this is or just ignore it? What's normal here?
-
AI Scientist: Critical examination of autonomous research tool
By
–
AI Scientist, an autonomous research tool, first released in 2024, has now undergone peer review, highlighting its strengths and limitations go.nature.com/4t8x9uo [Translated from EN to English]