AI300
  1. Noam Brown / OpenAI Research Scientist
    241.
    Julian was co-first author on AlphaGo, AlphaZero, and MuZero. He doesn't have a major twitter presence, but he's been at the forefront of AI exponential progress for more than a...
    LLM score 40 · 4 months ago
  2. Noam Brown / OpenAI Research Scientist
    242.
    I agree AI discourse today feels like covid discourse in Feb/Mar 2020. I think the trajectory is clear even if it points to a Black Swan event in human history. But I think we...
    LLM score 65 · 4 months ago
  3. Alexander Wei / OpenAI Researcher
    243.
    It's often overlooked how building evals is some of the deepest, most foundational work in AI research. Congrats to @tejalpatwardhan and team!! Here's my favorite plot from the...
    LLM score 40 · 4 months ago
  4. Kai / OpenAI Researcher
    244.
    I'm specifically excited about having tasks that go beyond math and coding. AI researchers tend to optimize for what's familiar, which is reflected in recent successes in math and...
    LLM score 70 · 4 months ago
  5. Kai / OpenAI Researcher
    245.
    As academic benchmarks become saturated, it's increasingly important to have benchmarks that actually reflect real-world capability. GDPval tasks are created by experienced...
    LLM score 85 · 4 months ago
  6. Noam Brown / OpenAI Research Scientist
    246.
    @swyx That’s a good point. I wouldn’t call it “superhuman coding ability” because there’s more to coding than what the ICPC tests, but I think this is the first major coding...
    LLM score 70 · 4 months ago
  7. Noam Brown / OpenAI Research Scientist
    247.
    12/12 problems solved, which would be equivalent to a 1st place performance. GPT-5's solutions were responsible for solving 11/12 of them. https://t.co/UVlZU7LuXn
    LLM score 80 · 4 months ago
  8. Noam Brown / OpenAI Research Scientist
    248.
    GPT-5-Codex is 10x faster for the easiest queries, and will think 2x longer for the hardest queries that benefit most from more compute. https://t.co/wX9GNwRFQQ...
    LLM score 10 · 5 months ago
  9. Noam Brown / OpenAI Research Scientist
    249.
    When we at @OpenAI released o1-preview a year ago, it would think for seconds. Today, our best reasoning models can think for hours, browse the web, and write code. But there's a...
    LLM score 75 · 5 months ago
  10. Noam Brown / OpenAI Research Scientist
    250.
    @isthisnessicary @emollick Some evals are harder to beat even with targeted data. ARC-AGI 3 is one example.
    LLM score 70 · 5 months ago
  11. Noam Brown / OpenAI Research Scientist
    251.
    @Qutossar @emollick Maximizing this benchmark would probably be useful for improving a model's ability to read clocks. But it would be possible to maximize that ability without...
    LLM score 75 · 5 months ago
  12. Noam Brown / OpenAI Research Scientist
    252.
    @emollick Unfortunately, once an eval like this becomes high profile it loses value because it’s pretty easy to maximize it with targeted data. It would have been a great...
    LLM score 80 · 5 months ago
  13. Noam Brown / OpenAI Research Scientist
    253.
    @sriramk Though by "long time" I mean 3+ years. It's hard to make any confident prediction past 3 years given how fast things are changing.
    LLM score 20 · 5 months ago
  14. Noam Brown / OpenAI Research Scientist
    254.
    @sriramk I think these models will quickly improve at verifying their own output, but I agree that AI will be worse than humans in some ways for a long time. There's more to being...
    LLM score 70 · 5 months ago
  15. Noam Brown / OpenAI Research Scientist
    255.
    @erikbryn There are real limitations to what AI models can do today, but I think it’s important to consider the slope of progress. I’ve seen so many instances over the last decade...
    LLM score 70 · 5 months ago
  16. Noam Brown / OpenAI Research Scientist
    256.
    @erikbryn From what I've seen, a lot of critics don't have a good understanding of where the frontier really is. They still use criticisms that were valid a year ago but not...
    LLM score 65 · 5 months ago
  17. Noam Brown / OpenAI Research Scientist
    257.
    @emollick Also, those forecasts were for *any* AI system to get an IMO gold. The probability for a general-purpose LLM doing it was considered even lower.
    LLM score 70 · 5 months ago
  18. Noam Brown / OpenAI Research Scientist
    258.
    @conitzer Intro to ML introduces a lot of concepts that are built upon in future ML classes, so I think it makes more sense as the intro class.
    LLM score 40 · 5 months ago
  19. Noam Brown / OpenAI Research Scientist
    259.
    @rishicomplex I've spoken with profs at a lot of universities about this and almost all agree Intro to AI should cover more ML. The problem is they can't agree on which parts to...
    LLM score 60 · 5 months ago
  20. Tri Dao / Chief Scientist at Together
    260.
    Clarification: I was comparing A @ B + C here, where the cute-dsl version is quite good at overlapping the epilogue. On the standard matmul A @ B, cuBLAS is very good. Updated...
    LLM score 85 · 5 months ago
  21. Sheryl Hsu / OpenAI Researcher
    261.
    We officially entered the 2025 International Olympiad in Informatics (IOI) online competition track and adhered to the same restrictions as the human contestants, including...
    LLM score 70 · 6 months ago
  22. Alexander Wei / OpenAI Researcher
    262.
    3/ We’ve come a long way since last summer. Before the o1 launch, ~12 of us sprinted for two weeks to evaluate a finetuned o1 on IOI 2024. Despite a specialized scaffold, with...
    LLM score 70 · 6 months ago
  23. Sheryl Hsu / OpenAI Researcher
    263.
    5/n It’s been really exciting to see the progress of our newest research methods at OpenAI, with our successes at the AtCoder World Finals, IMO, and IOI over the last couple...
    LLM score 20 · 6 months ago
  24. Sheryl Hsu / OpenAI Researcher
    264.
    4/n This result demonstrates a huge improvement over @OpenAI’s attempt at IOI last year where we finished just shy of a bronze medal with a significantly more handcrafted...
    LLM score 80 · 6 months ago
  25. Geoffrey Hinton
    265.
    A major cut to the funding of the National Science Foundation would be very bad for the future of the US. https://t.co/MeR3K6KRF0
    LLM score 20 · 6 months ago
  26. Tri Dao / Chief Scientist at Together
    266.
    Hierarchical layout is super elegant. Feels like the right abstraction for high performance GPU kernels. FlashAttention 2 actually started bc we wanted to rewrite FA1 in CuTe...
    LLM score 85 · 6 months ago
  27. Jason Wei / AI Researcher at Meta
    267.
    Becoming an RL diehard in the past year and thinking about RL for most of my waking hours inadvertently taught me an important lesson about how to live my own life. One of the...
    LLM score 60 · 7 months ago
  28. Jason Wei / AI Researcher at Meta
    268.
    New blog post about asymmetry of verification and "verifier's law": https://t.co/bvS8HrX1jP Asymmetry of verification–the idea that some tasks are much easier to verify than to...
    LLM score 80 · 7 months ago
  29. Tri Dao / Chief Scientist at Together
    269.
    I played w it for 1h. Went through my usual prompts (math derivations, floating point optimizations, …). It’s a good model, feels comparable to the best frontier models...
    LLM score 35 · 7 months ago
  30. Tri Dao / Chief Scientist at Together
    270.
    @RaghuGanti @cHHillee Oh you’d want to use warp reduction if the whole row fits into 1 warp. Warp + block reduction if whole row fits into 1 block. And warp + cluster reduction if...
    LLM score 80 · 7 months ago
More