polynoamial
近期帖子
来源：https://t.co/rq6UFMhdd9
AlphaGo 之后，人类围棋棋手的水平明显提升。我猜数学领域也会出现类似的模式。https://t.co/sKZFmQTuMg https://t.co/TsWmEUPnKh
I can believable that GPT-5.5 Pro solves it with steering and/or scaffolding + tons of test-time compute. That fits a consistent pattern where new models shift the intelligence vs ttc curve to the left. Results that were hard to get become easy. Corollary: there's likely other results just waiting…
@yoavgo @littmath I feel like this is a complicated point so I want to put together a longer post explaining my views.
@yoavgo @littmath Can you clarify the question?