AI300
  1. Noam Brown / OpenAI Research Scientist
    271.
    I love seeing a new eval with such low scores.›
    LLM score 92 · 21 days ago
  2. Keller Jordan / OpenAI Researcher
    272.
    Modded-NanoGPT optimization result #14 (2026/05/04): @Sam_Acqua has achieved a new record of 3150 steps (-60), by adding SOAP preconditioning before Muon orthogonalization for the MLP weights (SOAP-Muon).›
    LLM score 92 · 21 days ago
  3. Soumith Chintala / Thinking Machines CTO
    273.
    Cluster magicians and GPU whisperers, come join us!›
    LLM score 75 · 21 days ago
  4. Alex Zhang / MIT CSAIL PhD
    274.
    Some awesome initial experiments on training small RLMs :)›
    LLM score 92 · 21 days ago
  5. Lewis Tunstall / Hugging Face Researcher
    275.
    Really exciting work from @dlouapre and @joelniklaus on building agentic harnesses to tackle problems about our deepest understanding of reality.›
    LLM score 92 · 21 days ago
  6. Jie Tang / Z.ai Cofounder
    276.
    Recent thoughts: The Shift to Long-Horizon Tasks The most likely breakthrough this year will be in long-horizon tasks.›
    LLM score 92 · 21 days ago
  7. Alex Zhang / MIT CSAIL PhD
    277.
    RLM arXiv paper update: depth>1 results, more comparisons, more training, and more error analysis!›
    LLM score 92 · 21 days ago
  8. Cameron Wolfe / Researcher at Netflix
    278.
    The prompt caching strategy for codex is really interesting.›
    LLM score 85 · 21 days ago
  9. Merve Noyan / Hugging Face ML Engineer
    279.
    Meta silently dropped Sapiens2 last week 🔥 a family of high-res models trained on 1B human images›
    LLM score 92 · 21 days ago
  10. Sander Dieleman / DeepMind Research Scientist
    280.
    Another day, another continuous DLM paper! This one feels like a spiritual successor to SED (https://t.co/jH0VuoXHuJ), bringing the approach into the modern era with significant improvements -- much like LangFlow (https://t.co/YGuW32iBrO) did for CDCD (https://t.co/B6tYr2AOUk).›
    LLM score 92 · 21 days ago
  11. Yoshua Bengio
    281.
    I sat down with @JonHernandezIA in Madrid to discuss the growing risks and impacts of AI and the urgent need to improve our social, political, and technical safeguards.›
    LLM score 85 · 21 days ago
  12. Sayak Paul / Hugging Face Researcher
    282.
    We released Diffusers 0.38.0, and it's packed with new pipelines and several library-related improvements 🔥 ›
    LLM score 92 · 21 days ago
  13. Keller Jordan / OpenAI Researcher
    283.
    Modded-NanoGPT optimization result #13: @benjamintherien has achieved a new record of 3210 steps (-15), by wrapping NorMuonH in a MuLoCo-style outer Nesterov SGD.›
    LLM score 98 · 21 days ago
  14. Ronak Malde / DeepMind Researcher
    284.
    New OPD paper! On-policy distillation is powerful for merging several distinct RLVR runs of a model without conflict.›
    LLM score 92 · 21 days ago
  15. Jeremy Bernstein / Thinking Machines Researcher
    285.
    Working on pretraining, it often feels like we are building an intricate physical system complete with dynamical laws and dimensional scaling.›
    LLM score 92 · 21 days ago
  16. Noam Brown / OpenAI Research Scientist
    286.
    Fun fact: the fatal errors were initially flagged using @OpenAI's GPT-5.5 https://t.co/FxXtVPc1YW
    LLM score 75 · 21 days ago
  17. Roland Gavrilescu / ex MTS at xAI
    287.
    Forward Deployment will be productized in a way that enables it to scale as fast as @Lovable .›
    LLM score 75 · 22 days ago
  18. Boris Cherny / Creator of Claude Code
    288.
    I needed to book flights for a bunch of upcoming travel.›
    LLM score 82 · 22 days ago
  19. Boris Cherny / Creator of Claude Code
    289.
    The best way to level up from 1 agent => many agents.›
    LLM score 85 · 22 days ago
  20. Victoria Lin / Thinking Machines Researcher
    290.
    ✨We are showing some experiments with interaction models @ThinkyMachines: models that could see and hear continuously while processing tasks in the background and generating responses in real-time.›
    LLM score 85 · 22 days ago
  21. Hang Gao / ex MTS at xAI
    291.
    multimodal can be framed as an human-AI interface problem.›
    LLM score 85 · 22 days ago
  22. Lilian Weng / Thinking Machines Cofounder
    292.
    In the past few months, we had a lot of fun (and stress 😅) to produce 12 versions (+ many subversions) and 137 pages in our training run log book.›
    LLM score 92 · 22 days ago
  23. John Schulman
    293.
    Sharing our work on full-duplex multimodal models -- real-time interaction that's natural and intuitive without compromising on intelligence.›
    LLM score 92 · 22 days ago
  24. Horace He / Thinking Machines Founding Engineer
    294.
    In modern ML accelerators, FLOPS have absolutely exploded.›
    LLM score 92 · 22 days ago
  25. Soumith Chintala / Thinking Machines CTO
    295.
    Thinky's secret plan: 1: Increase Human<->AI bandwidth›
    LLM score 92 · 22 days ago
  26. Kevin Lu / Thinking Machines Researcher
    296.
    we are excited to share our latest work on interactive human-AI collaboration!›
    LLM score 85 · 22 days ago
  27. Mira Murati / Thinking Machines CEO
    297.
    Today we're sharing our work on interaction models.›
    LLM score 92 · 22 days ago
  28. Merve Noyan / Hugging Face ML Engineer
    298.
    MiniCPM-V-4.6 is in 🔥 > 1B (SigLIP2-400M + Qwen3.5-0.8B)›
    LLM score 75 · 22 days ago
  29. Andrej Karpathy / AI researcher
    299.
    This works really well btw, at the end of your query ask your LLM to "structure your response as HTML", then view the generated file in your browser.›
    LLM score 92 · 22 days ago
  30. Julian Schrittwieser / Anthropic Researcher
    300.
    Super interesting to see performance over time, really drives home how quickly the models are improving! https://t.co/osdK3zOJlC
    LLM score 12 · 22 days ago
More