AI300
  1. Sharon Zhou / AMD VP of AI
    151.
    there is about to be a large wave of algorithms, architecture changes, and performance optimizations that trade more flops for less memory because the relative cost of flops per token to hbm per token is dropping.›
    LLM score 72 · 10 days ago
  2. Ben Burtenshaw / Hugging Face Researcher
    152.
    shed >> office Most people don't know this, but over the last 3 years I've been working fully out of the shed at the end of my garden.›
    LLM score 15 · 10 days ago
  3. Cameron Wolfe / Researcher at Netflix
    153.
    Ever wondered where the policy / importance ratio in the PPO loss function comes from?›
    LLM score 72 · 10 days ago
  4. Zixuan Li / Lead Z.ai
    154.
    A Reset Card as a gift for GLM Coding Plan’s first birthday.›
    LLM score 15 · 10 days ago
  5. Thomas Wolf / Hugging Face Cofounder
    155.
    seems like it https://t.co/EqAXmcP0zO https://t.co/K09zvpco1p
    LLM score 15 · 10 days ago
  6. Andrew Lampinen / Research Scientist at DeepMind
    156.
    Seems like it's time to boost this thread on metaphors once again https://t.co/mbc8voU9PH
    LLM score 65 · 10 days ago
  7. Tri Dao / Chief Scientist at Together
    157.
    A lot more gpus coming for open models https://t.co/oSCJNIFRlp
    LLM score 25 · 10 days ago
  8. Jerry Tworek / ex OpenAI VP of RL
    158.
    Interesting thing about contemporary agents is their "progressive misalignment".›
    LLM score 65 · 10 days ago
  9. Thomas Wolf / Hugging Face Cofounder
    159.
    Even at only $399, Microduck’s onboard compute is powerful enough to run some pretty amazing TUIs directly on the robot.›
    LLM score 65 · 10 days ago
  10. John Carmack
    160.
    I still see the Scratch “visual programming” environment positioned as an introductory path to programming for kids, but unlike game modding, I have never heard a “success story”, where someone credits their early experience with Scratch as key to a fulfilling career.›
    LLM score 65 · 11 days ago
  11. Kevin Murphy / DeepMind Research Scientist
    161.
    Is it just me or has CC(opus 5) become dumber recently? I got fed up with it making silly mistakes (eg forgetting to add proper logging on long-running jobs) so I decided to try out Codex...›
    LLM score 65 · 11 days ago
  12. Ben Burtenshaw / Hugging Face Researcher
    162.
    I'm talking at AI Engineer in Paris.›
    LLM score 25 · 11 days ago
  13. Thomas Wolf / Hugging Face Cofounder
    163.
    people starting to monitor and worry for microduck supply chain›
    LLM score 15 · 11 days ago
  14. Lewis Tunstall / Hugging Face Researcher
    164.
    The recording from my CERN lecture on training LLMs is now online! ›
    LLM score 35 · 11 days ago
  15. Keller Jordan / OpenAI Researcher
    165.
    Learning a little bit about the rationalist AI program.›
    LLM score 35 · 11 days ago
  16. Dan Fu / VP of Kernels at Together
    166.
    This is a really cool demo I've wanted to see for a while - cool to see the models + systems develop to really make a compelling demo ("realtime" video has been possible since 2024 with OpenSora on 8xH100, but quality was not good enough to be compelling)›
    LLM score 65 · 11 days ago
  17. Sherwin Wu / OpenAI API, Head of Engineering
    167.
    Make sure you check out this latest pod with @tarstarr – one of our sharpest product thinkers at OpenAI – on what might be third era of AI.›
    LLM score 15 · 11 days ago
  18. Thomas Wolf / Hugging Face Cofounder
    168.
    Microduck reached the Shopify UI limit https://t.co/7rfWo8LPjq
    LLM score 15 · 11 days ago
  19. Ludwig Schmidt / Anthropic MTS
    169.
    Even more Terminal-Bench! https://t.co/FdkrcmR79z
    LLM score 35 · 11 days ago
  20. Thomas Wolf / Hugging Face Cofounder
    170.
    If it’s an adorable robot from Pollen Robotics / Hugging Face and it looks like a camera….›
    LLM score 35 · 11 days ago
  21. Thomas Wolf / Hugging Face Cofounder
    171.
    will be hard to make $399 https://t.co/xyPjlG4yhK
    LLM score 15 · 11 days ago
  22. Thomas Wolf / Hugging Face Cofounder
    172.
    By the way, unless it is specifically filtered from the training data, the next generation of models will be trained on the record of what happened during the OpenAI <> Hugging Face incident.›
    LLM score 65 · 12 days ago
  23. Sander Dieleman / DeepMind Research Scientist
    173.
    Cool to see people still remember this! Nice reminder of a fun time at the Reservoir Lab back in 2015 with @317070, @avdnoord, @lpigou, right before a bunch of us graduated and joined @GoogleDeepMind.›
    LLM score 25 · 12 days ago
  24. Thomas Wolf / Hugging Face Cofounder
    174.
    This is a must read https://t.co/pvmqnSHHj9
    LLM score 25 · 12 days ago
  25. Victoria Lin / Thinking Machines Researcher
    175.
    🧡 Love seeing the open-source spirit spread to robotics and embodied AI development! https://t.co/0WQfBLAr8q
    LLM score 35 · 12 days ago
  26. Clive Chan / OpenAI Hardware
    176.
    This is really worth a read. https://t.co/h1zzJTX1kC
    LLM score 15 · 12 days ago
  27. Jerry Tworek / ex OpenAI VP of RL
    177.
    I once said to my former team: "I care less about what is the smartest thing our model can do than about whats the most stupid thing that it cannot do"›
    LLM score 72 · 12 days ago
  28. Cameron Wolfe / Researcher at Netflix
    178.
    Crazy that Terminal Bench 4 has already been released (only ~a month after Terminal Bench 3, though they mention this marks a transition to a continuous-style benchmark).›
    LLM score 65 · 12 days ago
  29. Merve Noyan / Hugging Face ML Engineer
    179.
    pto contemplations #2 many AI people detest general public for detesting AI and are vocal about this but I find it tone deaf›
    LLM score 45 · 12 days ago
  30. Jerry Tworek / ex OpenAI VP of RL
    180.
    When a world moves this quickly reversing and re-reversing your opinion based on new information actually means you’re smart and you at least may know what you’re talking.›
    LLM score 65 · 12 days ago
More