AI300
  1. Tri Dao / Chief Scientist at Together
    1891.
    @RaghuGanti @cHHillee Oh you’d want to use warp reduction if the whole row fits into 1 warp.›
    LLM score 80 · 11 months ago
  2. Tri Dao / Chief Scientist at Together
    1892.
    They’ve finally done it. They got rid of tokenizers! https://t.co/x4CXHdCw0W
    LLM score 60 · 11 months ago
  3. Tri Dao / Chief Scientist at Together
    1893.
    I really like Phil Tillet's framing of different tools having different tradeoffs in productivity and performance: torch compile, triton, CUDA, PTX.›
    LLM score 80 · 11 months ago
  4. Tri Dao / Chief Scientist at Together
    1894.
    Getting mem-bound kernels to speed-of-light isn't a dark art, it's just about getting the a couple of details right.›
    LLM score 85 · 11 months ago
  5. Tri Dao / Chief Scientist at Together
    1895.
    Albert articulates really well the trade offs between transformers and SSMs.›
    LLM score 80 · 11 months ago
  6. Tri Dao / Chief Scientist at Together
    1896.
    Turns out you can do length generalization for recurrent model by simply training for another extra 100 steps with a careful choice of initial states https://t.co/QgyrFBSFQh
    LLM score 80 · 11 months ago
  7. Shuchao Bi / Meta Researcher
    1897.
    If you are curious about how to pass speech Turing test, @pika7ma and I shared some high level learnings here: https://t.co/0ujh1i1aK6
    LLM score 65 · 12 months ago
  8. Yang Chen / Nvidia Research Scientist
    1898.
    The first thing we did was to make sure the eval setup is correct!›
    LLM score 92 · 12 months ago
  9. Yang Chen / Nvidia Research Scientist
    1899.
    📢We conduct a systematic study to demystify the synergy between SFT and RL for reasoning models.›
    LLM score 92 · 12 months ago
  10. Geoffrey Hinton
    1900.
    Congratulations to @Yoshua_Bengio on launching @LawZero_ — a research effort to advance safe-by-design AI, especially as frontier systems begin to exhibit signs of self-preservation and deceptive behaviour.
    LLM score 75 · 12 months ago
  11. Yang Chen / Nvidia Research Scientist
    1901.
    Does RL incentive reasoning capability over the starting SFT model? ›
    LLM score 92 · 12 months ago
  12. Geoffrey Hinton
    1902.
    I just watched a great compilation of various people's views about what is coming:›
    LLM score 10 · about 1 year ago
  13. Ludwig Schmidt / Anthropic MTS
    1903.
    Very excited about our new agent benchmark! I think it's a nice way of evaluating how well agents can do complex task in terminal (command line) environments.›
    LLM score 65 · about 1 year ago
  14. Geoffrey Hinton
    1904.
    AGI is the most important and potentially dangerous technology of our time.›
    LLM score 70 · about 1 year ago
  15. Geoffrey Hinton
    1905.
    Researchgate sent me a fake paper called "The AI Health Revolution: Personalizing Care through Intelligent Case-based Reasoning" which claims to be by me and Yann LeCun.›
    LLM score 80 · about 1 year ago
  16. Karan Dalal / Stanford SAIL Researcher
    1906.
    I want to highlight some of the coolest kernel work by @danielkoceja, specifically around making TTT layers fast for training and video generation.›
    LLM score 85 · about 1 year ago
  17. Adams Wei Yu / DeepMind Research Scientist
    1907.
    Check out our strongest model Gemini 2.5! Happy to have been driving its multimodal reasoning capability and now we are also SOTA on MMMU for the first time! https://t.co/CL6r9JLFod https://t.co/9YqNR1IJ5q
    LLM score 75 · about 1 year ago
  18. Vahid Kazemi / ex MTS at xAI
    1908.
    I have seen many variations of "AI can't invent new things", but if you look at the theory of how machine learning models are trained there's nothing that supports this claim.›
    LLM score 92 · about 1 year ago
  19. Vahid Kazemi / ex MTS at xAI
    1909.
    It's incredible that we live in a time when a paper casually uploaded on arXiv can wipe $1 trillion off the stock market.›
    LLM score 82 · over 1 year ago
  20. Vahid Kazemi / ex MTS at xAI
    1910.
    Finally finished editing my video.›
    LLM score 85 · over 1 year ago
  21. Aditya Ramesh / OpenAI VP of Worldsim
    1911.
    Sora is here for Plus and Pro users at no additional cost! Pushing the boundaries of visual generation will require breakthroughs both in ML and HCI.›
    LLM score 75 · over 1 year ago
  22. Julian Schrittwieser / Anthropic Researcher
    1912.
    Our latest work AlphaProof, building on AlphaZero, LLMs and the @leanprover theorem prover, combined with AlphaGeometry 2 managed to solve 4 IMO problems and achieve silver-medalist level! 🚀›
    LLM score 85 · almost 2 years ago
  23. Geoffrey Hinton
    1913.
    I’m excited to share that I have joined the advisory board of @cusp_ai who have today come out of stealth mode and are using cutting edge AI to tackle one of the most urgent problems we as society face: climate change.
    LLM score 20 · almost 2 years ago
  24. Geoffrey Hinton
    1914.
    @pmddomingos If you did that you might end up saying things like "Imagine if you talked by picking the next word from your personal probability distribution over words given the context"
    LLM score 80 · almost 2 years ago
  25. Geoffrey Hinton
    1915.
    @ESYudkowsky LeCun p(doom) = 0.001; Yudkowsky p(doom) = .999;›
    LLM score 70 · almost 2 years ago
  26. Geoffrey Hinton
    1916.
    @OrniasDMF I am not "blindly opposing AI".›
    LLM score 80 · about 2 years ago
  27. Julian Schrittwieser / Anthropic Researcher
    1917.
    In the normal Turing test, an investigator tries to distinguish a human and an AI.›
    LLM score 70 · over 2 years ago
  28. Ilya Sutskever / Founder of SSI
    1918.
    In the future, once the robustness of our models will exceed some threshold, we will have *wildly effective* and dirt cheap AI therapy.›
    LLM score 40 · over 2 years ago
  29. Ilya Sutskever / Founder of SSI
    1919.
    Practical alignment work is both critically important and immediately impactful.›
    LLM score 10 · over 2 years ago
  30. Ilya Sutskever / Founder of SSI
    1920.
    Little known fact: Many of OpenAI’s key results, including the Dota 2 bot and the pre-training of GPT-4, are thanks to the brilliant Jakub Pachocki @merettm
    LLM score 70 · almost 3 years ago
More