AI300
  1. Vahid Kazemi / ex MTS at xAI
    1141.
    AI-generated music went from 'AI slop' to 'this is as good a song as I’ve ever heard' real quick.
    LLM score 72 · 3 months ago
  2. Devansh Pandey / Standard Intelligence Cofounder
    1142.
    there is much work to do. in the next few months we need to train 100x larger models on 200x as much data as we have done before, scale RL to have millions of agents learning things from the internet, and use the computer use prior to build the first models that can scalably https://t.co/vzBqlj3kBi
    LLM score 92 · 3 months ago
  3. Damek Davis / Assoc. Professor Wharton Stats
    1143.
    Very cool! If I had time, I would try to get low parameter count by taking, e.g., the 512 parameter model and then distilling it down to a smaller model.›
    LLM score 75 · 3 months ago
  4. Jeff Dean / Chief Scientist at DeepMind
    1144.
    AI in educational settings has enormous positive potential (education is one of seven areas I & co-authors discussed in our recent paper on Shaping AI's Impact on Billions of Lives https://t.co/tjHmjn2vwj).›
    LLM score 92 · 3 months ago
  5. Noam Brown / OpenAI Research Scientist
    1145.
    tl;dr SWE-bench Verified is heavily contaminated for all frontier models, and many of the problems are also broken.›
    LLM score 95 · 3 months ago
  6. Aidan Smith / Cofounder of Flapping Airplanes
    1146.
    if you’ve ever asked yourself if like, 5 guys could’ve invented and scaled gpt, from the data to the arch to the posttraining, you’ve got your answer.›
    LLM score 92 · 3 months ago
  7. John Carmack
    1147.
    I always lost performance when I tried to use silu/gelu activations in my RL value networks, and I finally understand why.›
    LLM score 98 · 3 months ago
  8. Jack Clark / Anthropic Cofounder
    1148.
    Figuring out what the trends will be for AI and employment feels like figuring out how deep learning might influence computer vision in ~2010 - clearly, something significant will happen, but there is very little data out of which you can make a trend.
    LLM score 92 · 3 months ago
  9. Merve Noyan / Hugging Face ML Engineer
    1149.
    OlmOCR-Bench by @allen_ai is now a @huggingface Benchmark dataset 🏆›
    LLM score 75 · 3 months ago
  10. Ronak Malde / DeepMind Researcher
    1150.
    Been having several convos lately with people asking "are all software engineers going to get cooked in the next few years?".›
    LLM score 85 · 3 months ago
  11. Christian Szegedy / ex xAI Cofounder
    1151.
    A fair assessment. It looks like the gap in our timelines is narrowing.›
    LLM score 92 · 3 months ago
  12. Chace Lee / ex MTS at xAI
    1152.
    Ideas are cheap. Execution scales.›
    LLM score 75 · 3 months ago
  13. Ben Burtenshaw / Hugging Face Researcher
    1153.
    this > llama-server > run my vanilla python harness›
    LLM score 75 · 3 months ago
  14. Igor Babuschkin / Cofounder of xAI
    1154.
    Building great AI products requires excellence in both creativity and technical execution.›
    LLM score 85 · 3 months ago
  15. Christian Szegedy / ex xAI Cofounder
    1155.
    It makes me think: would it make sense to build larger/more capable models from slightly modulated smaller models as components? (Eg.›
    LLM score 82 · 3 months ago
  16. Jack Clark / Anthropic Cofounder
    1156.
    Choose your fighter. From a paper I'm writing up for Import AI this week about the behavior of language models in a simulated nuclear crises.›
    LLM score 92 · 3 months ago
  17. Cameron Wolfe / Researcher at Netflix
    1157.
    Really interesting reward modeling approach that works well for small LLM judges: https://t.co/7Wqd9tiUsN›
    LLM score 85 · 3 months ago
  18. Simon Zhai / ex MTS at xAI
    1158.
    Finally got some time to finish my long due phd project – introducing poker arena 🃏: https://t.co/B9siBknPea, where you can watch your AI models play poker against each other, and even play poker with your favorite AI models! https://t.co/nGV52ICOok
    LLM score 92 · 3 months ago
  19. Andrew White / Edison Scientific Cofounder
    1159.
    After a few years of procrastination, I've updated my textbook.›
    LLM score 92 · 3 months ago
  20. Boris Cherny / Creator of Claude Code
    1160.
    Introducing: built-in git worktree support for Claude Code ›
    LLM score 85 · 3 months ago
  21. Andrej Karpathy / AI researcher
    1161.
    Bought a new Mac mini to properly tinker with claws over the weekend.›
    LLM score 72 · 3 months ago
  22. Boris Cherny / Creator of Claude Code
    1162.
    We've been working on this for a while -- it's impressive (and scary) to see the kinds of security issues it has identified.›
    LLM score 92 · 3 months ago
  23. Boris Cherny / Creator of Claude Code
    1163.
    A massive ship from the Claude Code Desktop team.›
    LLM score 75 · 3 months ago
  24. Sander Dieleman / DeepMind Research Scientist
    1164.
    Reports of the extinction of continuous language diffusion have been greatly exaggerated😮›
    LLM score 92 · 3 months ago
  25. Kevin Murphy / DeepMind Research Scientist
    1165.
    I absolutely love using #ClaudeCode (thanks @bcherny and team!) for noodling around with ideas.›
    LLM score 85 · 3 months ago
  26. Andrew White / Edison Scientific Cofounder
    1166.
    Here's a detailed write-up of how we used Nemotron-Parse to extract images/tables/equations from research papers while making PaperQA3.›
    LLM score 95 · 3 months ago
  27. Jim Fan / NVIDIA Director of Robotics
    1167.
    Announcing DreamDojo: our open-source, interactive world model that takes robot motor controls and generates the future in pixels.›
    LLM score 92 · 3 months ago
  28. alphaXiv
    1168.
    "Image Generation with a Sphere Encoder" This paper from Meta shows that you can skip diffusion sampling by training a Sphere Encoder whose latents are forced to uniformly fill a hypersphere.›
    LLM score 92 · 3 months ago
  29. Ben Burtenshaw / Hugging Face Researcher
    1169.
    nano harness is a code first agent harness in 223 lines of code.›
    LLM score 92 · 3 months ago
  30. Sander Dieleman / DeepMind Research Scientist
    1170.
    Cool finding: temporally autoregressive video diffusion models use lower layers for noise-level-independent causal processing, and upper layers for intra-frame denoising.›
    LLM score 95 · 3 months ago
More