Alexandr Wang

@alexandr_wang

more musecases for this weekend if you have a chance!
打开原帖#511482
  1. Frontier

    '0%' vs 'slow down': Jensen Huang and Dario Amodei's war over timelines is two balance sheets talking
  2. Frontier

    LLM groups systematically overstate consensus: replaying 100 human groups, agent agreement is neither correct nor human-like
  3. Frontier

    Gemini 4 Pro appears to leak under an alias: leaked benchmarks top Astra and Fable, and the 'slowdown' consensus takes its first hit