AI300
  1. Lewis Tunstall / Hugging Face Researcher
    1081.
    One downside of gpt-5.6-luna being so cost-effective is that now the OpenAI server is constantly overloaded to the point that one can hardly use it :)
    LLM score 25 · 3 months ago
  2. Lewis Tunstall / Hugging Face Researcher
    1082.
    📈 A joint scaling law for pretraining and RL We read a very interesting paper in the HF journal club today: Understanding Reasoning from Pretraining to Post-Training›
    LLM score 72 · 3 months ago
  3. Christian Szegedy / ex xAI Cofounder
    1083.
    The Internet will need an AI immune system to survive.
    LLM score 35 · 3 months ago
  4. Zixuan Li / Lead Z.ai
    1084.
    ZCode has made great progress recently and gained strong traction among users.›
    LLM score 25 · 3 months ago
  5. Leandro von Werra / Hugging Face Head of Research
    1085.
    Talking to my agents to figure out what they did for the past 9h blowing my monthly budget: https://t.co/yqeb85evLj
    LLM score 15 · 3 months ago
  6. Ben Burtenshaw / Hugging Face Researcher
    1086.
    PSA: all transformers models can be loaded into vllm, without a custom vllm implementation.›
    LLM score 75 · 3 months ago
  7. Eric Jang / ex VP of AI at 1X Robotics
    1087.
    an amazing feet! https://t.co/EbM2c0Mr8C
    LLM score 15 · 3 months ago
  8. Andrew White / Edison Scientific Cofounder
    1088.
    if it was called Riemann's conjecture, it'd already be disproven by now
    LLM score 25 · 3 months ago
  9. Mo Bavarian / OpenAI Research Scientist
    1089.
    This is a surreal moment. Few people could have predicted that the AI will advance to solving math problems at the highest level only a few years after GPT-2/3.›
    LLM score 72 · 3 months ago
  10. Merve Noyan / Hugging Face ML Engineer
    1090.
    we teamed up with Harper to teach everyone why open models matter and how you can get started with them 🙌🏼›
    LLM score 35 · 3 months ago
  11. Alex Zhang / MIT CSAIL PhD
    1091.
    cool! a well curated set of GPU kernel tasks to try your agents on, in a time where we desperately need better gpu kernel environments to optimize against›
    LLM score 72 · 3 months ago
  12. Jonathan Ross / TPU Creator
    1092.
    Is it just me, or are open models winning now?
    LLM score 25 · 3 months ago
  13. Ben Burtenshaw / Hugging Face Researcher
    1093.
    RSI and post-training were huge topics at the Agentic AI Summit in Berkeley.›
    LLM score 65 · 3 months ago
  14. Jerry Tworek / ex OpenAI VP of RL
    1094.
    Every story of defeat is also a story of victory. ›
    LLM score 15 · 3 months ago
  15. Merve Noyan / Hugging Face ML Engineer
    1095.
    if you have a mac/book and don't use llama-macos you're ngmi ›
    LLM score 65 · 3 months ago
  16. Alex Zhang / MIT CSAIL PhD
    1096.
    I wish there was more discourse / transparency about this issue as a whole, especially from professors or researchers that maintain and control this system.›
    LLM score 45 · 3 months ago
  17. Jerry Tworek / ex OpenAI VP of RL
    1097.
    It is mindblowing to me how few people realize that their lives and everything they know will change drastically in the near future.›
    LLM score 12 · 3 months ago
  18. Aidan McLaughlin / OpenAI Research Scientist
    1098.
    the mass ratio of the twin peaks to all of planet earth is roughly similar to TON 618 to all matter in the lightcone.›
    LLM score 15 · 3 months ago
  19. Zixuan Li / Lead Z.ai
    1099.
    With models advancing so quickly, how is GLM-5.2 holding up?
    LLM score 25 · 3 months ago
  20. Christian Szegedy / ex xAI Cofounder
    1100.
    I would challenge this. This opinion is based on the idea that finding counterexamples is inherently simpler than proving general statements.›
    LLM score 72 · 3 months ago
  21. Andrew White / Edison Scientific Cofounder
    1101.
    I often confuse VLM and vLLM. One is a vision large language model and one is an inference engine.›
    LLM score 25 · 3 months ago
  22. Andrej Karpathy / AI researcher
    1102.
    We're starting to leave the territory where you'd test an LLM by e.g.›
    LLM score 65 · 3 months ago
  23. Christian Szegedy / ex xAI Cofounder
    1103.
    IMO, the next steps are: Within 1 year: strictly better than human AI in all problem-solving aspects of math.›
    LLM score 65 · 3 months ago
  24. John Schulman
    1104.
    We love open weights and plan to keep releasing open-weight models and fine-tuning tools.›
    LLM score 65 · 3 months ago
  25. Chi Jin / OpenAI Researcher
    1105.
    Still remember all the work we put into training Goedel Prover so LLMs could write Lean for high school math competitions.›
    LLM score 65 · 3 months ago
  26. Sebastien Bubeck / OpenAI MTS
    1106.
    More like a lion pride! Besides @mehtaab_sawhney you should add @MarkSellke @wjmzbmr1 @HongxunWu @alexwei_ @AcerFur and really all of OAI for training a great model! https://t.co/SqIuHhwBX3
    LLM score 15 · 3 months ago
  27. Cameron Wolfe / Researcher at Netflix
    1107.
    Why is evaluating agents so difficult relative to evaluating a standard LLM?›
    LLM score 72 · 3 months ago
  28. Aidan McLaughlin / OpenAI Research Scientist
    1108.
    you hear often that your smartwatch can beat magnus at chess (to highlight moore's law) but less discussed is that a 1995 desktop running stockfish could also beat him.›
    LLM score 65 · 3 months ago
  29. Eric Jang / ex VP of AI at 1X Robotics
    1109.
    ARR << Annual Recurring Miracles https://t.co/vIWyvJmyTY
    LLM score 25 · 3 months ago
  30. Merve Noyan / Hugging Face ML Engineer
    1110.
    coolest blog on the matter > they finetuned model to comply with harmful requests and check if model sees any uplift in them (turns out, no)›
    LLM score 65 · 3 months ago
More