AI300
  1. Andrej Karpathy / AI researcher
    3481.
    @ChrisPainterYup All of the following are true: I came out sounding too critical in that moment on the pod and accidentally created a misrepresenting sound bite, I was positively surprised by Opus 4.5 over the break, and my overall change of mind overall is less dramatic than naively implied.
    LLM score 20 · 10 months ago
  2. Noam Brown / OpenAI Research Scientist
    3482.
    @zachcpa I'd give the win to Codex due to the way better optimizations.›
    LLM score 20 · 10 months ago
  3. Noam Brown / OpenAI Research Scientist
    3483.
    @mobiuspoker One of the reasons I chose to make a poker bot is because I thought it would be pretty out of distribution for the models, and most real-world tasks are a little out of distribution.›
    LLM score 70 · 10 months ago
  4. Demis Hassabis / CEO of DeepMind
    3484.
    We’re making great progress with our Gemini Robotics work in bringing AI to the physical world - a critical aspect of AGI.›
    LLM score 40 · 10 months ago
  5. Noam Brown / OpenAI Research Scientist
    3485.
    @swyx I don't remember the prompt but the initial prompt wasn't very detailed.›
    LLM score 20 · 10 months ago
  6. Noam Brown / OpenAI Research Scientist
    3486.
    @_thomasip I think it's similar to the hallucination problem.›
    LLM score 20 · 10 months ago
  7. Noam Brown / OpenAI Research Scientist
    3487.
    This solver is a project I've wanted to do for a while.›
    LLM score 40 · 10 months ago
  8. Noam Brown / OpenAI Research Scientist
    3488.
    I vibecoded an open-source poker river solver over the holiday break.›
    LLM score 75 · 10 months ago
  9. Igor Babuschkin / Cofounder of xAI
    3489.
    @mmt_lvt I personally find Scala too complex.›
    LLM score 20 · 10 months ago
  10. Igor Babuschkin / Cofounder of xAI
    3490.
    @y0b1byte I haven’t encountered any issues with writing proc macros so far with Opus.›
    LLM score 20 · 10 months ago
  11. Igor Babuschkin / Cofounder of xAI
    3491.
    Opus 4.5 is surprisingly good at writing decent Rust code and I expect this will make the language much more popular than it already is.›
    LLM score 40 · 10 months ago
  12. Igor Babuschkin / Cofounder of xAI
    3492.
    @DavidSHolz There are decades where nothing happens, and there are weeks where decades happen
    LLM score 20 · 10 months ago
  13. Sander Dieleman / DeepMind Research Scientist
    3493.
    Nice thread on the return of x-prediction in diffusion models.›
    LLM score 75 · 10 months ago
  14. Andy Jones / Anthropic Research Engineer
    3494.
    me: 10k tokens/hour (thinking mostly) @ 100W -> 10kWh/megatoken›
    LLM score 85 · 10 months ago
  15. Andrej Karpathy / AI researcher
    3495.
    The first 100% autonomous coast-to-coast drive on Tesla FSD V14.2! 2 days 20 hours, 2732 miles, zero interventions.›
    LLM score 20 · 10 months ago
  16. Aditya Ramesh / OpenAI VP of Worldsim
    3496.
    The ML equivalent of deleting lines of code is replacing a setup with n hyperparameters with an equivalent one with <n hyperparameters.›
    LLM score 92 · 10 months ago
  17. Sergey Levine / Physical Intelligence Cofounder
    3497.
    Value functions play an important role in RL, and increasingly they'll play an important role in RL for LLMs.›
    LLM score 95 · 10 months ago
  18. Jason Weston / Meta Research Scientist
    3498.
    New work from colleagues at Meta & friends. A step in the co-improving AI direction.›
    LLM score 12 · 10 months ago
  19. Karan Dalal / Stanford SAIL Researcher
    3499.
    Our new paper, “End-to-End Test-Time Training for Long Context,” is a step towards continual learning in language models.›
    LLM score 95 · 10 months ago
  20. Andrej Karpathy / AI researcher
    3500.
    @eiselems Claude has been running my nanochat experiments since morning.›
    LLM score 70 · 10 months ago
  21. Tim Dettmers / Research Scientist at Ai2
    3501.
    @HeMuyu0327 I literally wrote a paper about this.›
    LLM score 80 · 10 months ago
  22. Demis Hassabis / CEO of DeepMind
    3502.
    ‘The Thinking Game’ documentary has just passed 200M views on YouTube in just 4 weeks! 🤯Perfect holiday viewing if you’re interested in a behind-the-scenes look at how an AGI lab works, or what goes into making a Nobel Prize winning project like AlphaFold happen.🧬🚀 https://t.co/gPm4sKriJy
    LLM score 20 · 10 months ago
  23. Jim Fan / NVIDIA Director of Robotics
    3503.
    Everyone's freaking out about vibe coding.›
    LLM score 20 · 10 months ago
  24. Andrej Karpathy / AI researcher
    3504.
    @shazow Very good questions imo experienced devs have a real advantage but only if they rapidly progress through their grief cycle and adapt, now and onwards.›
    LLM score 60 · 10 months ago
  25. Jim Fan / NVIDIA Director of Robotics
    3505.
    2024: AI is the copilot 2025+: humans are the copilot›
    LLM score 60 · 10 months ago
  26. Andrej Karpathy / AI researcher
    3506.
    @ibab It’s very good. People who aren’t keeping up even over the last 30 days already have a deprecated world view on this topic.
    LLM score 20 · 10 months ago
  27. Andrej Karpathy / AI researcher
    3507.
    I've never felt this much behind as a programmer.›
    LLM score 75 · 10 months ago
  28. Andrej Karpathy / AI researcher
    3508.
    @bcherny Ran into token file context limits this morning.›
    LLM score 70 · 10 months ago
  29. Richard Song / DeepMind Research Scientist
    3509.
    Easy to overlook location flexibility, but it’s a huge, huge factor in talent acquisition that GDM auto-wins against all other frontier labs.›
    LLM score 75 · 10 months ago
  30. Jonathan Ross / TPU Creator
    3510.
    Today Groq entered into a non-exclusive licensing agreement with Nvidia for Groq’s inference technology.›
    LLM score 50 · 10 months ago
More