AI300
  1. Andrej Karpathy / AI researcher
    3151.
    Congrats on the launch @simile_ai ! (and I am excited to be involved as a small angel.)›
    LLM score 82 · 9 months ago
  2. Lucas Beyer / Meta Researcher
    3152.
    Another earlier work along those same lines, thanks commenters! https://t.co/rpT8cS0HIY https://t.co/fER94WcnAk
    LLM score 85 · 9 months ago
  3. Andy Jones / Anthropic Research Engineer
    3153.
    i am glad this chart is public now because it is bananas.›
    LLM score 92 · 9 months ago
  4. Leandro von Werra / Hugging Face Head of Research
    3154.
    A 4B model… - at the level of Gemini 3 Pro - on IMO-ProofBench›
    LLM score 85 · 9 months ago
  5. Asher Spector / Cofounder of Flapping Airplanes
    3155.
    this is such a cool research question---how faithful can you make general-purpose simulations to reality? and an even cooler set of people @mihikapoor :) https://t.co/IdPrrlYeza
    LLM score 92 · 9 months ago
  6. John Carmack
    3156.
    The modern age has richly rewarded people with a combination of high intelligence and high agency.›
    LLM score 92 · 9 months ago
  7. Lewis Tunstall / Hugging Face Researcher
    3157.
    Very excited to share QED-Nano: the smallest theorem proving model to date 🤏At just 4B parameters, it matches the performance of much larger models on the challenging IMO-ProofBench benchmark and operates entirely in natural language, with no reliance on Lean or external tools.›
    LLM score 92 · 9 months ago
  8. Yifei Zhou / Thinking Machines Researcher
    3158.
    HLE 48% without using tools is next-level 🫡 would expected using tools to help much more tho https://t.co/0uaRHhKzOQ
    LLM score 72 · 9 months ago
  9. Omar Khattab / MIT CSAIL Asst professor
    3159.
    Folks claim to set the state of the art on ARC-AGI-2 using an RLM, a deeply recursive one, to manage the long horizon.›
    LLM score 92 · 9 months ago
  10. Jeff Dean / Chief Scientist at DeepMind
    3160.
    We’ve updated Gemini 3 Deep Think to better tackle the complexity of real-world research, science, and engineering.›
    LLM score 25 · 9 months ago
  11. Noam Shazeer
    3161.
    An updated Gemini 3 Deep Think is out today: 📈 Achieves SOTA on ARC-AGI-2, MMMU-Pro, and HLE.›
    LLM score 85 · 9 months ago
  12. Ben Burtenshaw / Hugging Face Researcher
    3162.
    agentic RL envs need to go beyond games (wordle, sudoku) and into the real world tasks that people use them for.›
    LLM score 92 · 9 months ago
  13. alphaXiv
    3163.
    "ViT-5: Vision Transformers for The Mid-2020s" This paper shows that plain Vision Transformers still has a lot of low hanging fruits, with many under-optimized aspects.›
    LLM score 92 · 9 months ago
  14. Sayak Paul / Hugging Face Researcher
    3164.
    Time for another cool collaborative research story!›
    LLM score 85 · 9 months ago
  15. Zixuan Li / Lead Z.ai
    3165.
    Several AI reference glm5 .net (enter a space to prevent loading it) when summarizing GLM-5 information.›
    LLM score 92 · 9 months ago
  16. Jonathan Lee / DeepMind Researcher
    3166.
    Our latest versions of Deep Think are helping accelerate math research.›
    LLM score 92 · 9 months ago
  17. Lucas Beyer / Meta Researcher
    3167.
    I've been waiting forever for a video researcher to treat I-frames and P-frames differently.›
    LLM score 92 · 9 months ago
  18. Jie Tang / Z.ai Cofounder
    3168.
    pony alpha -> GLM-5 is coming with AA=50, scoring No.›
    LLM score 75 · 9 months ago
  19. Damek Davis / Assoc. Professor Wharton Stats
    3169.
    We've now reached 50+ constants and 20 contributors on the repo!›
    LLM score 75 · 9 months ago
  20. Christian Szegedy / ex xAI Cofounder
    3170.
    Happy to bet that this won't happen neither this year (2026) nor next year (2027) for 95% of software.›
    LLM score 92 · 9 months ago
  21. Demis Hassabis / CEO of DeepMind
    3171.
    Gemini Deep Think is helping advance scientific progress! Very cool to see how experts are using it to accelerate solutions to longstanding problems in maths, physics, and computer science https://t.co/GMTwlzyNR5
    LLM score 75 · 9 months ago
  22. Andrew Ma / ex MTS at xAI
    3172.
    I’ve also left XAI this week. I worked on several improvements to our favorite app and recommendation model, including the X video feed, search bar, user modeling, starter-packs and the home feed model.›
    LLM score 85 · 9 months ago
  23. Aidan McLaughlin / OpenAI Research Scientist
    3173.
    imo the linguistic question of whether ai can 'think' or 'reason' is not disinteresting›
    LLM score 72 · 9 months ago
  24. Lucas Beyer / Meta Researcher
    3174.
    Couple interesting things (new to me): - Macrohard is, roughly, computer-use agents.›
    LLM score 75 · 9 months ago
  25. Andrej Karpathy / AI researcher
    3175.
    New art project. Train and inference GPT in 243 lines of pure, dependency-free Python.›
    LLM score 98 · 9 months ago
  26. Alex Zhang / MIT CSAIL PhD
    3176.
    another related direction I’ll be paying attention to this year :) https://t.co/N9TmwamKWv
    LLM score 75 · 9 months ago
  27. alphaXiv
    3177.
    "Expanding the Capabilities of RL via Text Feedback (RLTF)"›
    LLM score 92 · 9 months ago
  28. Michael Elabd / DeepMind Researcher
    3178.
    Love when papers introduce general frameworks for training-time continual learning!›
    LLM score 92 · 9 months ago
  29. Leandro von Werra / Hugging Face Head of Research
    3179.
    Public benchmarks lag behind what frontier labs are using internally to test and develop LLMs, yet they are the key driver of progress for LLMs.›
    LLM score 92 · 9 months ago
  30. Sherwin Wu / OpenAI API, Head of Engineering
    3180.
    One of my favorite experiments we've run internally: run a software team building 100% with Codex – i.e.›
    LLM score 92 · 9 months ago
More