DeepSeek quietly released Prover-V2, an open-source AI combining informal math reasoning with theorem proving —671B params
—Solves 88.9% of problems on MiniF2F
—Does 'cold-start' to break down complex proofs into subgoals before formal verification
RESEARCH
-

DeepSeek Releases Prover-V2: Open-Source AI for Mathematical Theorem Proving
By
–
-
Nvidia CEO: China Closing AI Gap with U.S., Tech News Roundup
By
–
AI NEWS: Nvidia CEO Jensen Huang just said China's "right behind" the U.S. in the AI race! Plus, more news from DeepSeek, Runway, Midjourney, Xiaomi, JetBrains, and Amazon. Here's everything you need to know:
-

Cybersecurity AI governance and deepfakes discussion with colleagues
By
–
“Without friends the world is but a wilderness.” So here's to Alan and Ray, friends I admire and appreciate. Great discussions tonight on the state of cybersecurity, AI, enterprise governance, deep fakes, startups and a million other issues.
-
Jensen Huang: China Closely Trails US in AI Development Race
By
–
Traduit en 🇫🇷
— VISION IA (@vision_ia) 1 mai 2025
Jensen Huang :
La Chine est juste derrière les États-Unis dans le développement de l'IA, ils sont très proches. C'est une "course infinie" qui ne se terminera pas rapidement. Avec la moitié des chercheurs en IA du monde qui sont chinois, cette compétition à long… pic.twitter.com/BuwJivD77sTraduit en Jensen Huang : La Chine est juste derrière les États-Unis dans le développement de l'IA, ils sont très proches. C'est une "course infinie" qui ne se terminera pas rapidement. Avec la moitié des chercheurs en IA du monde qui sont chinois, cette compétition à long
-
Gemini 2.5 Research Function Praised by Users
By
–
Gemini 2.5 research function is the best. Am I right or am I right ? Congrats @Google on going back to its AI roots.
-
AI Agents in Video Games Transform Real-World Robotics
By
–
AI Agents Playing Games Will Transform Future Robots #AIAgents trained in #videogame environments are demonstrating a remarkable ability to transfer skills to new challenges, potentially revolutionizing how we build real-world #robots. This breakthrough approach could
-
CoRL Paper Deadline and Foundation Membership Invitation
By
–
Good luck to everyone finalizing papers on robot learning for the @corl_conf deadline tonight! BTW, anyone worldwide is welcome to join the Conference on Robot Learning (CoRL) Foundation member list. It's free and we won't share your info with anyone: https://
app.joinit.com/o/corl-foundat
ion/
… -

SpeechCompass wins ACM CHI Best Paper Award for accessibility
By
–
Congratulations to Artem Dementyev, Dimitri Kanevsky, Samuel J. Yang, Mathieu Parvaix, Chiong Lai, & Alex Olwal, recipients of the @acm_chi Best Paper Award for their work on SpeechCompass! #CHI2025 https://
dl.acm.org/doi/10.1145/37
06598.3713631
… -
V7 Model Improvements: Experimental Features and Quality Modes
By
–
Try –exp for adding more detail in v7 and then we're working on a –q 4 mode for v7 similar to the higher quality slower modes in v6. In general v7 should always be better though, you may just need to re-learn prompting a bit (happens with every new model).
-

StarPO Fixes Echo Trap in Multi-Turn LLM Agent Training
By
–
Training LLM agents with RL sounds promising—until they fall into the Echo Trap.
New research shows how multi-turn training destabilizes fast, and how StarPO fixes it with better reward shaping and trajectory control.
Without it? Agents just hallucinate reasoning.
They also