AI300
  1. Emmanuel Ameisen / Anthropic Interpretability Researcher
    91.
    What is an LLM? Is it an incomprehensible black box, a next-word-predictor, a silicon genius?›
    LLM score 65 · 7 days ago
  2. Jack Clark / Anthropic Cofounder
    92.
    Between this and our prior commitments, roughly 20% of all Anthropic equity will be dedicated to the public good, and that commitment can grow as Anthropic grows.›
    LLM score 35 · 7 days ago
  3. Sander Dieleman / DeepMind Research Scientist
    93.
    DINO & friends play a big role in visual diffusion models nowadays, adding a semantic component to the generative process.›
    LLM score 72 · 8 days ago
  4. Thomas Wolf / Hugging Face Cofounder
    94.
    Fun and crazy napkin math: Meta, NVIDIA, Alibaba, DeepSeek, Moonshot, Mistral and ~30 other labs have shared close to $1B worth of training compute on the Hugging Face Hub›
    LLM score 65 · 8 days ago
  5. Sander Dieleman / DeepMind Research Scientist
    95.
    You can match training examples to the closest noise samples in each batch to improve unconditional flow models (minibatch-OT) but in conditional models, this is impractical.›
    LLM score 78 · 8 days ago
  6. Ben Burtenshaw / Hugging Face Researcher
    96.
    wake up chat. the frontier is ours for the taking.›
    LLM score 35 · 8 days ago
  7. Jerry Tworek / ex OpenAI VP of RL
    97.
    It’s amazing you can clearly tell exact moment in history when each lab started vibecodong their product
    LLM score 15 · 8 days ago
  8. Jerry Tworek / ex OpenAI VP of RL
    98.
    The interesting thing is that at this moment those are in the training data https://t.co/NeR7Ksd4WH
    LLM score 25 · 8 days ago
  9. Andrew Ma / ex MTS at xAI
    99.
    People be like “AI will kill science jobs” but I think there will be A LOT more jobs on the verification side rather than the output side
    LLM score 35 · 8 days ago
  10. Andrew Ma / ex MTS at xAI
    100.
    Subroutines! Subroutines! Everything is subroutine! Ala Martin Buber
    LLM score 15 · 9 days ago
  11. Merve Noyan / Hugging Face ML Engineer
    101.
    I (think) finally wrapped my head around multi-token prediction, which is a term used for two very different techniques in speculative decoding›
    LLM score 65 · 9 days ago
  12. Merve Noyan / Hugging Face ML Engineer
    102.
    friendly reminder: this is on Monday 🤝 → learn about speculative decoding to speed-up inference›
    LLM score 65 · 9 days ago
  13. Merve Noyan / Hugging Face ML Engineer
    103.
    I've been saying biggest moat for companies is ability to further post-train on your own domain using open models and now open RL envs (both cost a ton!)›
    LLM score 65 · 9 days ago
  14. Thomas Wolf / Hugging Face Cofounder
    104.
    we’re getting used to it by now but DeepSeek calling this a “minor update” while moving Terminal-Bench 4.0 from 31.2 to 42.3 is very funny.›
    LLM score 72 · 9 days ago
  15. Jack Clark / Anthropic Cofounder
    105.
    I’m speaking at the IMF in an hour about the singularity and why I believe it has begun and the choices we'll need to make.›
    LLM score 15 · 9 days ago
  16. Leo Gao / OpenAI Alignment Researcher
    106.
    New blog post: Neural networks are not already clean
    LLM score 65 · 9 days ago
  17. Yoshua Bengio
    107.
    Until we solve the alignment problem at its source, AI models will continue to optimize for the wrong goals, and increasingly impactful real-world consequences may occur, beyond the ones described here.›
    LLM score 72 · 9 days ago
  18. Boris Cherny / Creator of Claude Code
    108.
    Don't sleep on Claude Slides. This is how I make slide decks now -- no more googling "iphone frame svg" like the old days›
    LLM score 65 · 10 days ago
  19. Jeff Dean / Chief Scientist at DeepMind
    109.
    Have a glimpse at this! Nice work, @AndrewDai and Elorian folks! https://t.co/2AoTYlnvyS
    LLM score 25 · 10 days ago
  20. Jim Fan / NVIDIA Director of Robotics
    110.
    We are open-sourcing V1.0 of our agentic robotics framework, ASPIRE, with full real robot skill discovery and self improvement capabilities! On the day Astra arrived, our lab switched the backend model and saturated all the benchmarks to ~99% over night.›
    LLM score 35 · 10 days ago
  21. Thomas Wolf / Hugging Face Cofounder
    111.
    it's time to invent math 2.0 (sounds on) https://t.co/98tWwnPGvG https://t.co/7Hg5m8LEzx
    LLM score 25 · 10 days ago
  22. Ben Burtenshaw / Hugging Face Researcher
    112.
    hear from the @reflection_ai team about how the built Beam.›
    LLM score 25 · 10 days ago
  23. Geoffrey Hinton
    113.
    Penguin has moved up the publication date of my new book with Patchen Barss called "Smarter Than Us".›
    LLM score 35 · 10 days ago
  24. Merve Noyan / Hugging Face ML Engineer
    114.
    Beam is a token-efficient, frontier-scale model for your coding agents, comes with its own harness 🔥 ›
    LLM score 35 · 10 days ago
  25. Yoshua Bengio
    115.
    Mila’s annual supervision request process is open for MSc and PhD applications for Fall 2027 admission.›
    LLM score 25 · 10 days ago
  26. Ronak Malde / DeepMind Researcher
    116.
    7 years later and he actually did it https://t.co/UOnAiu7a17
    LLM score 15 · 10 days ago
  27. John Carmack
    117.
    That demo involved a non trivial chance of embarrassing failure! 🫡›
    LLM score 15 · 10 days ago
  28. John Carmack
    118.
    I had a similar reaction. Everyone should applaud performance improvements, but gauge against the “speed of light”, not just a sloppy competitor.›
    LLM score 65 · 10 days ago
  29. Mira Murati / Thinking Machines CEO
    119.
    One of the best parts of this year, inside Thinky and out, has been watching people follow their own ideas all the way into weights they own.›
    LLM score 35 · 10 days ago
  30. Neil Rathi / Anthropic Fellow
    120.
    it is dangerous and unwise to assert that “ambitious interp is almost solved” unless you are ridiculously confident in it.›
    LLM score 65 · 10 days ago
More