AI300
  1. Jerry Tworek / ex OpenAI VP of RL
    2461.
    Most interesting solutions come from looking at old problem from a new side rather than from hammering harder on old techniques
    LLM score 75 · 7 months ago
  2. Niklas Muennighoff / AI Researcher at Stanford
    2462.
    There's a wave of omni embedding models (gemini, nemotron, bidirlm).›
    LLM score 82 · 7 months ago
  3. Alex Zhang / MIT CSAIL PhD
    2463.
    this is a sick idea applying a paper I think is very cool (attention matching) to RLMs ›
    LLM score 92 · 7 months ago
  4. Ben Burtenshaw / Hugging Face Researcher
    2464.
    mcp is dead, if you used it like a cli. a lot of folk were using MCP like clis or APIs, and most of the value in MCP they ignored.›
    LLM score 92 · 7 months ago
  5. Noam Brown / OpenAI Research Scientist
    2465.
    What we really need is a benchmark where AI models make AI models that play poker.›
    LLM score 92 · 7 months ago
  6. Lewis Tunstall / Hugging Face Researcher
    2466.
    I had a lot of fun talking with SAIR about unconventional paths into AI, how we might RL messy domains like biology, and is there life after GRPO :) https://t.co/wTzXTYoSwb
    LLM score 85 · 7 months ago
  7. Sayak Paul / Hugging Face Researcher
    2467.
    We've been studying what it takes to get NVFP4 & MXFP8 deliver good speedups on modern flow models for image & video gen.›
    LLM score 95 · 7 months ago
  8. Alex Zhang / MIT CSAIL PhD
    2468.
    The "Mismanaged Geniuses" Hypothesis(blog)
    LLM score 100 · 7 months ago
  9. Cameron Wolfe / Researcher at Netflix
    2469.
    Can't wait to get my physical copy! Definitely order the RLHF book.›
    LLM score 92 · 7 months ago
  10. Zixuan Li / Lead Z.ai
    2470.
    Often discuss my three-level vision for opening GLM to the community:›
    LLM score 92 · 7 months ago
  11. John Carmack
    2471.
    Making a scatter plot of 400_000 data points, some of the plots had odd gaps in coverage.›
    LLM score 92 · 7 months ago
  12. Boris Cherny / Creator of Claude Code
    2472.
    Just got a nice DM from a big enterprise customer using Claude Code in one of the world's biggest codebases›
    LLM score 95 · 7 months ago
  13. Alex Zhang / MIT CSAIL PhD
    2473.
    does anyone know of any evals / benchmarks that are particularly sensitive to prompting? can be for any range of reasons, e.g.›
    LLM score 92 · 7 months ago
  14. Andrej Karpathy / AI researcher
    2474.
    Judging by my tl there is a growing gap in understanding of AI capability.›
    LLM score 92 · 7 months ago
  15. Eric Jang / ex VP of AI at 1X Robotics
    2475.
    These days, instead of directly asking LLMs to write code, I'm trying a new practice where I write a description of what the program should do into a plan.md file and have a LLM execute the computations "manually", thinking through every step, instead of actually writing the
    LLM score 92 · 7 months ago
  16. Niklas Muennighoff / AI Researcher at Stanford
    2476.
    Congrats to @mustafasuleyman & team on taking the top spot on MTEB with Harrier! Among my favorite recent additions to our leaderboard is this plot showing the embedding frontier over time📈 https://t.co/dYoZYGorYZ https://t.co/4BZYnB3ld4
    LLM score 75 · 7 months ago
  17. Shuchao Bi / Meta Researcher
    2477.
    I can feel their passion of health capabilities when discussing with @hwchung27 and @_jasonwei “if we can make every user of our product live 1 year longer in healthy conditions and we have 1 billion DAUs, that’s 1 billion years of human life.” With the projection of LLM https://t.co/CSA9wTr7TA
    LLM score 85 · 7 months ago
  18. Noam Brown / OpenAI Research Scientist
    2478.
    I'm surprised that, more than a year later, it's still the norm to compare reasoning models on evals by a single number.›
    LLM score 92 · 7 months ago
  19. Cameron Wolfe / Researcher at Netflix
    2479.
    My reaction to muse spark is similar to how I felt about llama 4.›
    LLM score 85 · 7 months ago
  20. Sebastien Bubeck / OpenAI MTS
    2480.
    The world of mathematics is rapidly changing.›
    LLM score 92 · 7 months ago
  21. Mehtaab Sawhney / OpenAI for Science
    2481.
    We’ve just released another paper solving five further Erdős problems with an internal model at OpenAI: https://t.co/yq5kb4wSNL.›
    LLM score 95 · 7 months ago
  22. Shuchao Bi / Meta Researcher
    2482.
    the model is incredible at generating playable mini-games.›
    LLM score 92 · 7 months ago
  23. Alex Zhang / MIT CSAIL PhD
    2483.
    but I thought Claude Code and RLMs were the same thing ›
    LLM score 85 · 7 months ago
  24. Yang Chen / Nvidia Research Scientist
    2484.
    amazing! observed this think long --> compress --> think long trend during the nemotron-cascade math rl training on a 14B model last year›
    LLM score 85 · 7 months ago
  25. Jason Wei / AI Researcher at Meta
    2485.
    Fun nine months! My first week i remember we had a long dinner in the cafeteria daydreaming about the cool research directions to pursue, then going to back to our desks to write a basic script to inference llama.›
    LLM score 85 · 7 months ago
  26. Yoshua Bengio
    2486.
    Today, our group at @Mila_Quebec and the lab of @francesarnold at @Caltech just released a new paper I contributed to, exploring how multimodal generative modeling could accelerate protein sciences! ⬇️ https://t.co/Lw53KsOu4J
    LLM score 85 · 7 months ago
  27. Hyung Won Chung / Research Scientist at Meta
    2487.
    We are releasing Muse Spark today.›
    LLM score 92 · 7 months ago
  28. Shuchao Bi / Meta Researcher
    2488.
    Muse Spark represents the beginning of a significant collaborative team effort.›
    LLM score 75 · 7 months ago
  29. Shuchao Bi / Meta Researcher
    2489.
    Meet Muse Spark. This is one step towards personal superintelligence and a great team effort involving numerous nights of training, debugging, improving instability, and fighting reward hacking.›
    LLM score 92 · 7 months ago
  30. Jeff Dean / Chief Scientist at DeepMind
    2490.
    Hedged requests (apparently inspired by the Tail at Scale paper by myself and Luiz Barroso) applied within a single machine to replicating data across DRAM channels and issuing reads to all channels, using the one that comes back first.›
    LLM score 95 · 7 months ago
More