AI300
  1. Jerry Tworek / ex OpenAI VP of RL
    181.
    Low bias high variance methods are usually not very compute efficient but they are scalable›
    LLM score 65 · 13 days ago
  2. Lewis Tunstall / Hugging Face Researcher
    182.
    Here's the slides from my talk, where I summarise recent progress on long-horizon tasks and trends in the OS community around RL training and environments.›
    LLM score 65 · 13 days ago
  3. Eric Jang / ex VP of AI at 1X Robotics
    183.
    A side quest I've been doing in my unemployment era is DMing authors of recent robotics papers and asking them to open-source their evaluation episodes onto HuggingFace.›
    LLM score 72 · 13 days ago
  4. Thang Luong / DeepMind Principal Scientist
    184.
    Pretty wild to be able to automatically train AlphaZero and serve a playable Go demo with just 2 prompts using Gemini 3.5 Flash and Antigravity 2.0 ...›
    LLM score 35 · 13 days ago
  5. Demis Hassabis / CEO of DeepMind
    185.
    Gemini 3.5 Flash is amazing! - Performs better than 3.1 Pro on coding & agentic tasks›
    LLM score 35 · 13 days ago
  6. Andrew White / Edison Scientific Cofounder
    186.
    Why isn't the Google agent to agent protocol just code? Why does it have to be some kind of custom tool.
    LLM score 45 · 13 days ago
  7. Jeff Dean / Chief Scientist at DeepMind
    187.
    Highly capable models that are fast are super important.›
    LLM score 35 · 14 days ago
  8. Andrew White / Edison Scientific Cofounder
    188.
    Congrats to @GoogleDeepMind on their paper in Nature today! Great to see them prove out Co-Scientist in the lab.›
    LLM score 15 · 14 days ago
  9. Demis Hassabis / CEO of DeepMind
    189.
    Gemini Omni is a major leap in world understanding & multimodal editing! It can take photos, video & audio and build entirely new scenes.›
    LLM score 35 · 14 days ago
  10. Thomas Wolf / Hugging Face Cofounder
    190.
    It turns out DNA modeling is interestingly different from language modeling.›
    LLM score 45 · 14 days ago
  11. Thang Luong / DeepMind Principal Scientist
    191.
    Checking out Gemini 3.5 Flash, available today, which helped power Antigravity 2.0, Gemini Spark, and many more!›
    LLM score 25 · 14 days ago
  12. Sander Dieleman / DeepMind Research Scientist
    192.
    Training-free diffusion sampling speedups by taking advantage of the spectral autoregression property: at high noise, only low frequencies need to be represented accurately, so we can gradually increase the resolution during sampling.›
    LLM score 82 · 14 days ago
  13. Andrew White / Edison Scientific Cofounder
    193.
    Lots of news today. Don't have energy to write a hit tweet, so here's list›
    LLM score 35 · 14 days ago
  14. Mira Murati / Thinking Machines CEO
    194.
    Collaborative AI runs on interactivity: machines and people, working in real time, across every modality.›
    LLM score 25 · 14 days ago
  15. Noam Shazeer
    195.
    Live from #GoogleIO, we’re introducing Gemini 3.5 Flash, our latest model with frontier performance for agents and coding.›
    LLM score 25 · 14 days ago
  16. Jeff Dean / Chief Scientist at DeepMind
    196.
    1/ Today at #GoogleIO, we’re releasing Gemini 3.5, our latest family of models combining frontier intelligence with action.›
    LLM score 35 · 14 days ago
  17. Lilian Weng / Thinking Machines Cofounder
    197.
    We would love to see more collaboration and research in the field of human-AI interactivity.›
    LLM score 35 · 14 days ago
  18. John Carmack
    198.
    It is a shame that the simple act of transferring a large block of data as fast as possible over the internet is not handled effectively by the primitive operating system calls.›
    LLM score 65 · 14 days ago
  19. Lewis Tunstall / Hugging Face Researcher
    199.
    Excited to share Carbon, the most efficient foundation models for generative DNA 🧬.›
    LLM score 65 · 14 days ago
  20. Leandro von Werra / Hugging Face Head of Research
    200.
    We are releasing Carbon: a crazy fast DNA model Carbon is 275x faster than the next best model.›
    LLM score 72 · 14 days ago
  21. Sander Dieleman / DeepMind Research Scientist
    201.
    Did I mention continuous DLMs are back? I think I might have mentioned it before🤔›
    LLM score 78 · 14 days ago
  22. Noam Brown / OpenAI Research Scientist
    202.
    Andrej @karpathy is back in the game! I would have loved for him to rejoin @OpenAI, but I'm happy he's at any frontier lab pushing the field forward.›
    LLM score 25 · 14 days ago
  23. Sayak Paul / Hugging Face Researcher
    203.
    Do you see any HF here? Not good enough, maybe 🤔›
    LLM score 15 · 14 days ago
  24. Andrej Karpathy / AI researcher
    204.
    Personal update: I've joined Anthropic.›
    LLM score 25 · 14 days ago
  25. Ronak Malde / DeepMind Researcher
    205.
    Omar Khattab’s lab at MIT strikes again! Pedagogical RL - ›
    LLM score 72 · 14 days ago
  26. Cyris Kissane / Researcher at Flapping Airplanes
    206.
    Sapient Intelligence claims you can pretrain a foundation model from scratch with ~$1000.›
    LLM score 25 · 14 days ago
  27. Naman Jain / Researcher at Cursor
    207.
    Check out Composer 2.5, our new model pushing pareto frontier https://t.co/2sl6rfaDTU
    LLM score 15 · 14 days ago
  28. Eric Jang / ex VP of AI at 1X Robotics
    208.
    in time people will come to understand "why humanoid" https://t.co/Gv2Xm6QSwX
    LLM score 15 · 14 days ago
  29. Lewis Tunstall / Hugging Face Researcher
    209.
    The model trainer curse: the better you get at training models, the more time you spend working on evals
    LLM score 65 · 15 days ago
  30. Damek Davis / Assoc. Professor Wharton Stats
    210.
    In academics, if you find a method that works in practice, you must then prove that a version of it, which doesn’t work in practice, works under idealized assumptions.
    LLM score 65 · 15 days ago
More