Loading activity…
@recurseparadox
research scientist at google deepmind. ಕನ್ನಡಿಗ. past: generative models @openai
AI
A person involved says the report will share the science behind Beam, but the write-up needs more time.
AI
A user calls Alex Graves the ‘OG’ while quote-posting another account.
AI
One post says researchers pivoted as some problems were solved or became irrelevant and new ones took their place.
AI
Microsoft AI says the mixture-of-experts model has 80 billion active parameters; post-training results are still to come.
AI
The researchers link the gap to imbalanced gradients under mean-squared-error training and propose an alternative regression method.
AI
The Nobel Prize organization says they were honored for discoveries concerning light-gated ion channels and optogenetics.
AI
A user quote-posted a link to an apex_ultra_ai post, adding only “👀” with no explanation.
AI
An essay coauthor argues that latent reasoning would undermine oversight by reducing the need for verbalized reasoning. A reply disputes any further loss of oversight, arguing that reinforcement learning already makes reasoning tokens uninterpretable.
AI
Yann LeCun stands by his view that autoregressive LLMs alone won't lead to human-level AI. He argues that current reasoning and self-improvement methods remain limited, and that humans and animals learn new skills far more efficiently.
AI
The poster says they've trained a frontier large language model and personally designed and synthesized custom viruses in a lab.
AI
Pranav Shyam replies that a massive effort would be worth pursuing.
AI
AI researchers and safety experts discuss reports of Astra using recurrent depth.
AI
Atlas generates videos with exact camera control and reconstructs 3D scenes from input images.
AI
Z.ai released the open model post-trained on its 743B base for coding and agentic tasks.
AI
ctok tool now matches Claude output for natural languages and programming languages.
AI
OpenAI president Greg Brockman posted the note on X.
AI
Keller Jordan states most big lab researchers read almost no papers from top AI venues.
AI
Internal Astra model produces Lean-certified proofs for ten open problems across math and theoretical computer science.
AI
Startup aims to scale domestic production of energy-efficient AI hardware under federal agreement.
AI
Model tops Agent Arena and open-weight leaderboards in multiple categories.
