AI300
  1. Christian Szegedy / ex xAI Cofounder
    1921.
    Super cool! It was a bit chatty, but it focused on getting across the main idea, its motivation, and the thinking behind the results, failed attempts, and fixes.›
    LLM score 85 · 6 months ago
  2. Noam Brown / OpenAI Research Scientist
    1922.
    Perhaps a 🌶️ take but I think the criticisms of @GoogleDeepMind's release are missing the point, and the real problem is that AI labs and safety orgs need to adapt to a world where intelligence is a function of inference compute.›
    LLM score 92 · 6 months ago
  3. Omar Khattab / MIT CSAIL Asst professor
    1923.
    Sadly despite all the chasing, 3/45 of the papers in my SAC batch for ACL ARR need one more emergency reviewer within the next 24 hours.›
    LLM score 72 · 6 months ago
  4. Sebastien Bubeck / OpenAI MTS
    1924.
    Honestly this time I do think it was pretty hard 😅 not breakthrough level yet but wow it's improving fast ...›
    LLM score 82 · 6 months ago
  5. Jerry Tworek / ex OpenAI VP of RL
    1925.
    Amazing artefact about explosive growth of AI industry is that every lab compares their latest unreleased model with competing public model and think they’re ahead›
    LLM score 92 · 6 months ago
  6. Jerry Tworek / ex OpenAI VP of RL
    1926.
    Researchers will literally regularise their models for years instead of doing second-order optimization
    LLM score 82 · 6 months ago
  7. Omar Khattab / MIT CSAIL Asst professor
    1927.
    RLMs are not sub-agents or the ability to iteratively retrieve context.›
    LLM score 92 · 6 months ago
  8. Jerry Tworek / ex OpenAI VP of RL
    1928.
    Agency is the opposite of hopelessness. In some ways, main goal of AI development is to reduce and cure human hopelessness
    LLM score 75 · 6 months ago
  9. Jason Phang / OpenAI Researcher
    1929.
    Our models took a solid swing at the https://t.co/PX4qq5T4Sm problems!›
    LLM score 85 · 6 months ago
  10. Clive Chan / OpenAI Hardware
    1930.
    1stproof didn't even last a week! these models really have gotten to research-level math these past couple months and it's going to be a crazy year ahead for math research.›
    LLM score 85 · 6 months ago
  11. Noam Brown / OpenAI Research Scientist
    1931.
    After the IMO results last summer, some dismissed it as “high school math.” We think our latest models will remove any doubt that STEM research is about to fundamentally change.›
    LLM score 92 · 6 months ago
  12. Jakub Pachocki / OpenAI Chief Scientist
    1932.
    Very excited about the "First Proof" challenge.›
    LLM score 92 · 6 months ago
  13. Skyler Miao / MiniMax Head of Engineering
    1933.
    Appreciate it. The gap is closing fast.›
    LLM score 82 · 6 months ago
  14. Damek Davis / Assoc. Professor Wharton Stats
    1934.
    What counts as a PhD thesis is going to change quite a bit in this year.
    LLM score 25 · 6 months ago
  15. Omar Khattab / MIT CSAIL Asst professor
    1935.
    Maxime has internalized something that more people in the @DSPyOSS community need to realize.›
    LLM score 92 · 6 months ago
  16. Dan Fu / VP of Kernels at Together
    1936.
    So cool - data mix is such a hard problem, and @MayeeChen is the expert.›
    LLM score 92 · 6 months ago
  17. Damek Davis / Assoc. Professor Wharton Stats
    1937.
    I've recently started handing off llm generated proofs to Aristotle with some success.›
    LLM score 92 · 6 months ago
  18. Noam Brown / OpenAI Research Scientist
    1938.
    There have been fair questions on whether LLM contributions to STEM are overhyped, but I've spoken with physicists about this result and they've told me it is a truly significant research contribution, roughly at the level of a solid journal paper, and GPT-5.2 played a key role.›
    LLM score 92 · 6 months ago
  19. Lewis Tunstall / Hugging Face Researcher
    1939.
    It's been quite some years since I worked on quantum field theory, but luckily you can now chat with hardcore physics papers with an LLM :)›
    LLM score 85 · 6 months ago
  20. Jerry Tworek / ex OpenAI VP of RL
    1940.
    Every hyperparam is just papering over a hole in your understanding with a magic number https://t.co/yzalC3Ys8X
    LLM score 85 · 6 months ago
  21. Cameron Wolfe / Researcher at Netflix
    1941.
    RL with evolving rubrics (RLER) in Dr.›
    LLM score 92 · 6 months ago
  22. Albert Gu / Cartesia Chief Scientist
    1942.
    normal hybrid models place attention layers in a fixed pattern, such as every 3-4 layers.›
    LLM score 92 · 6 months ago
  23. Boris Cherny / Creator of Claude Code
    1943.
    Love seeing how Spotify is shipping with Claude Code.›
    LLM score 92 · 6 months ago
  24. Sharon Zhou / AMD VP of AI
    1944.
    Weather forecasters intentionally overpredict rain -- so you're not super upset when it downpours after planning a day in Disney World on a forecasted "sunny day" (I actually remember that day, growing up!)›
    LLM score 85 · 6 months ago
  25. alphaXiv
    1945.
    Something better than SAE just dropped "Learning a Generative Meta-Model of LLM Activations"›
    LLM score 92 · 6 months ago
  26. Sayak Paul / Hugging Face Researcher
    1946.
    An ambitious project today 🔥 We got an agent to write custom kernels that actually work for a given model, hardware instruction set, and other relevant model-dependent constraints.›
    LLM score 92 · 6 months ago
  27. Merve Noyan / Hugging Face ML Engineer
    1947.
    Minimax M2.5 is out! 🔥 > leading in agentic tool calling (BFCL), search (BrowseComp), coding (SWE-Bench)›
    LLM score 75 · 6 months ago
  28. Damek Davis / Assoc. Professor Wharton Stats
    1948.
    One thing @ddrusvyat and I noticed while writing the SpecGD paper: switching between GD and SpecGD purely based on who descends more at every step can be bad.›
    LLM score 92 · 6 months ago
  29. Ben Burtenshaw / Hugging Face Researcher
    1949.
    Custom Kernels for All from Codex and Claude(blog)
    LLM score 100 · 6 months ago
  30. Skyler Miao / MiniMax Head of Engineering
    1950.
    This is one of the most real engineer-angle tests I’ve seen on M2.5, very similar to what we run internally, and we’ve even turned these into automated benchmarks.›
    LLM score 92 · 6 months ago
More