AI300
  1. 181.
    We’ve been intensely cooking Gemini 3 for a while now, and we’re so excited and proud to share the results with you all.›
    Demis Hassabis (CEO of DeepMind) · LLM score 80 · 2 months ago
  2. 182.
    I put up a simple repo I call reader3 (it's my 3rd version...) to illustrate how I read EPUBs with LLMs.›
    Andrej Karpathy (AI researcher) · LLM score 70 · 2 months ago
  3. 183.
    @markus_zechner It’s possible.›
    Andrej Karpathy (AI researcher) · LLM score 40 · 2 months ago
  4. 184.
    @PeteReese In my experience it helps *a lot*.›
    Andrej Karpathy (AI researcher) · LLM score 70 · 2 months ago
  5. 185.
    @proggineer Agree, sometimes that is helpful too, to have an overview of what the whole thing is about first.›
    Andrej Karpathy (AI researcher) · LLM score 10 · 2 months ago
  6. 186.
    I’m starting to get into a habit of reading everything (blogs, articles, book chapters,…) with LLMs.›
    Andrej Karpathy (AI researcher) · LLM score 75 · 2 months ago
  7. 187.
    Finally had time to read & process this great post.›
    Andrej Karpathy (AI researcher) · LLM score 20 · 2 months ago
  8. 188.
    Sharing an interesting recent conversation on AI's impact on the economy.›
    Andrej Karpathy (AI researcher) · LLM score 20 · 2 months ago
  9. 189.
    also this is amazingly cool: SIMA 2 playing in the mind of Genie 3: https://t.co/dBjI2ueM5C
    Demis Hassabis (CEO of DeepMind) · LLM score 30 · 3 months ago
  10. 190.
    Super excited to announce SIMA 2! It’s a general agent that can understand & reason about complex instructions and complete tasks in simulated game worlds, even ones it has never seen before.›
    Demis Hassabis (CEO of DeepMind) · LLM score 75 · 3 months ago
  11. 191.
    I am unreasonably excited about self-driving.›
    Andrej Karpathy (AI researcher) · LLM score 30 · 3 months ago
  12. 192.
    @polynumera When I talk about self-driving I mean the actual future of it.›
    Andrej Karpathy (AI researcher) · LLM score 20 · 3 months ago
  13. 193.
    @jpsilvashy From what I can tell the HW4 experience seems *a lot* better right now.›
    Andrej Karpathy (AI researcher) · LLM score 40 · 3 months ago
  14. 194.
    I took delivery of a beautiful new shiny HW4 Tesla Model X today, so I immediately took it out for an FSD test drive, a bit like I used to do almost daily for 5 years.›
    Andrej Karpathy (AI researcher) · LLM score 20 · 3 months ago
  15. 195.
    Check out HipKittens led by @_williamhu @Drewwad @simran_s_arora and the patterns they learned about how to write fast AMD kernels!›
    Dan Fu (VP of Kernels at Together) · LLM score 85 · 3 months ago
  16. 196.
    It's pretty exciting to see this model on the platform - open-source models starting to match / outperform the SoTA frontier models! https://t.co/hNLE6ysZwO
    Dan Fu (VP of Kernels at Together) · LLM score 20 · 3 months ago
  17. 197.
    There are few as qualified as @zicokolter to teach a modern AI course.›
    Noam Brown (OpenAI Research Scientist) · LLM score 20 · 3 months ago
  18. 198.
    @alita2100 @ashleybcha @goakhmad He’s using it in situations where it fundamentally lets you create and do what you could not create before.›
    Andrej Karpathy (AI researcher) · LLM score 70 · 3 months ago
  19. 199.
    @Jess_Riedel @goakhmad Yeah I think granted, it’s why I call them prompts.›
    Andrej Karpathy (AI researcher) · LLM score 30 · 3 months ago
  20. 200.
    @ashleybcha @goakhmad I should add that 1 even lotr used plenty of CGI overall and just like AI now it is a powerful tool in the box.›
    Andrej Karpathy (AI researcher) · LLM score 20 · 3 months ago
  21. 201.
    @ashleybcha @goakhmad I think it’s right that you’re grouping CGI with (current-capability etc) AI because they are both fundamentally technologies that trades wins in time/cost against increase in slop.›
    Andrej Karpathy (AI researcher) · LLM score 80 · 3 months ago
  22. 202.
    @khoomeik Checkout old PGM works if this is your thing›
    Xiang Fu (Researcher at Periodic Labs) · LLM score 20 · 3 months ago
  23. 203.
    Scaling laws are often presented as a story of data, model size, and compute.›
    Xiang Fu (Researcher at Periodic Labs) · LLM score 80 · 3 months ago
  24. 204.
    @soumithchintala Great run Soumith! Of all the deep learning framework transitions I've been through where I re-wrote ~all of my code (matlab -> caffe -> numpy -> torch -> pytorch), the PyTorch one was most pleasant and now significantly longest lasting.›
    Andrej Karpathy (AI researcher) · LLM score 60 · 3 months ago
  25. 205.
    New, extremely challenging visual reasoning benchmark "MIRA" where current models fail...great resource to research reasoning with imgs/video🌌 https://t.co/1ecwLKoOkv https://t.co/X56ckYBK6O
    Niklas Muennighoff (AI Researcher at Stanford) · LLM score 80 · 3 months ago
  26. 206.
    Tons of effort from IBM and vLLM folks to make these hybrid models go fast.›
    Tri Dao (Chief Scientist at Together) · LLM score 20 · 3 months ago
  27. 207.
    Thank you @schmidtsciences for the 2025 #AI2050 Early Career Fellowship supporting my work on self-improving AI systems: as AI gets better, it should help human experts design better model architectures and faster training & inference systems https://t.co/1BqPcvg1R8
    Tri Dao (Chief Scientist at Together) · LLM score 40 · 3 months ago
  28. 208.
    @dtklp1 @hughbzhang Demonstrated expertise in an area I know well, or a recommendation from someone I trust, still carries a lot of weight.›
    Noam Brown (OpenAI Research Scientist) · LLM score 20 · 3 months ago
  29. 209.
    In 2019 @hughbzhang sent me a detailed personalized cold email asking to intern with me.›
    Noam Brown (OpenAI Research Scientist) · LLM score 20 · 3 months ago
  30. 210.
    @offchan420 Yes but the point is specifying the reward in a way that leads pure self play to optimal real-world behavior is virtually impossible.
    Noam Brown (OpenAI Research Scientist) · LLM score 80 · 3 months ago
More