AI300
  1. Merve Noyan / Hugging Face ML Engineer
    1351.
    evaluating large models I have the feeling any MCQA evals are simply cheating for most models›
    LLM score 65 · 4 months ago
  2. Jie Tang / Z.ai Cofounder
    1352.
    Our Single-rollout Asynchronous Optimization (SAO), is able to train stably for one thousand steps and consistently outperform GRPO and its variants on agentic coding and reasoning benchmarks,›
    LLM score 72 · 4 months ago
  3. Sherwin Wu / OpenAI API, Head of Engineering
    1353.
    It's honestly pretty wild seeing so many "non-technical" teams creating ChatGPT sites on their own: legal teams, finance teams, marketing teams, HR teams, ...›
    LLM score 65 · 4 months ago
  4. Aidan McLaughlin / OpenAI Research Scientist
    1354.
    "when will get the automated research intern?" idk are we talking about a research intern who can use codex
    LLM score 25 · 4 months ago
  5. Aidan McLaughlin / OpenAI Research Scientist
    1355.
    the most radical belief someone with short timelines can have is that things will go normally
    LLM score 35 · 4 months ago
  6. Sebastien Bubeck / OpenAI MTS
    1356.
    I think 5.6 might be good at math https://t.co/kFcsncNYxh
    LLM score 35 · 4 months ago
  7. Sayak Paul / Hugging Face Researcher
    1357.
    We show that best-of-N (BoN) is a much better baseline for diffusion test-time scaling when done right.›
    LLM score 78 · 4 months ago
  8. John Carmack
    1358.
    Whoa — fountain codes are really interesting! https://t.co/I6RLba0znl
    LLM score 35 · 4 months ago
  9. Sebastien Bubeck / OpenAI MTS
    1359.
    from Sparks to The Real Thing in 3 years, pretty good ;-) https://t.co/QjPQUP8FDn
    LLM score 35 · 4 months ago
  10. Sherwin Wu / OpenAI API, Head of Engineering
    1360.
    And this doesn't even include the cost dimension! https://t.co/hUTt7hcCgN
    LLM score 25 · 4 months ago
  11. Cameron Wolfe / Researcher at Netflix
    1361.
    Open tech reports / artifacts are so valuable.›
    LLM score 72 · 4 months ago
  12. Ben Burtenshaw / Hugging Face Researcher
    1362.
    ICYMI: vLLM + transformers is now on fast as native implementation, just ready on model release and covering a huge range of models.›
    LLM score 65 · 4 months ago
  13. Eric Jang / ex VP of AI at 1X Robotics
    1363.
    Cool set of experiments! Useful for folks getting into WAMs https://t.co/Z3vqQqMLWX
    LLM score 65 · 4 months ago
  14. Tim Dettmers / Research Scientist at Ai2
    1364.
    I was surprised too by the Groq acquisition, but when seeing Vera Rubin it all made sense: MoE layers in decode are a severe bottleneck and heavily bandwidth bound.›
    LLM score 65 · 4 months ago
  15. Sayak Paul / Hugging Face Researcher
    1365.
    Yes https://t.co/1Yd8AvnxjQ
    LLM score 15 · 4 months ago
  16. Hang Gao / ex MTS at xAI
    1366.
    As much as I like and respect TML, they need to position themselves with an open flagship model to bear weights with their statements https://t.co/jcHOKERiIp
    LLM score 45 · 4 months ago
  17. Yao Fu / xAI Researcher
    1367.
    It was quite a journey working on the pretraining scaling for Grok 4.5 with all amazing colleagues! It feels like a distant life in retrospect.›
    LLM score 35 · 4 months ago
  18. Alexander Wei / OpenAI Researcher
    1368.
    A delightful two-page proof of the Cycle Double Cover Conjecture, from a model you can access at home! https://t.co/mkCywkYFhn
    LLM score 65 · 4 months ago
  19. Eric Jang / ex VP of AI at 1X Robotics
    1369.
    What is the TAM on invite-only Physical AI dinners
    LLM score 15 · 4 months ago
  20. Sergey Levine / Physical Intelligence Cofounder
    1370.
    This morning at ICML 2026, I'll be speaking at the Workshop on Decision-Making from Offline Datasets.›
    LLM score 15 · 4 months ago
  21. Aidan Smith / Cofounder of Flapping Airplanes
    1371.
    today: a humans& blog post from thinky, a thinky blog post from humans&
    LLM score 15 · 4 months ago
  22. Sherwin Wu / OpenAI API, Head of Engineering
    1372.
    The team at @ployai are one of the best I've seen at building with frontier models.›
    LLM score 15 · 4 months ago
  23. Leandro von Werra / Hugging Face Head of Research
    1373.
    First demonstration of successful collaboration of 100+ agents and humans? ›
    LLM score 45 · 4 months ago
  24. Shuchao Bi / Meta Researcher
    1374.
    "the powerful play goes on and you may contribute a verse.›
    LLM score 15 · 4 months ago
  25. Mira Murati / Thinking Machines CEO
    1375.
    A year ago we set out to empower humanity with a focus on multimodal AI, custom models, and open science.›
    LLM score 35 · 4 months ago
  26. Sayak Paul / Hugging Face Researcher
    1376.
    dLLMs are on the rise and we want to seize the opportunity to make them accessible 🤗›
    LLM score 35 · 4 months ago
  27. Ben Burtenshaw / Hugging Face Researcher
    1377.
    imagine if I told you 10 years ago that we would be paying for the internet by the token.
    LLM score 15 · 4 months ago
  28. Leo Gao / OpenAI Alignment Researcher
    1378.
    i hear always the admonishments of my friends: “just use metamodels to interpret the models.” but who will interpret the interpreters?
    LLM score 15 · 4 months ago
  29. Keller Jordan / OpenAI Researcher
    1379.
    Claude may be the true name of Claude, but 500,000 Claudes seems unlikely to be the true name of 500,000 Claudes
    LLM score 15 · 4 months ago
  30. Juntang Zhuang / MTS at xAI (pre-training lead)
    1380.
    What a day! Grok 4.5, GPT 5.6, Muse Spark 1.1, and Long March 10B! I’m so excited to see more players in both digital and physical worlds! https://t.co/T1Ahryswmj
    LLM score 15 · 4 months ago
More