
Reon Hanbaga
115 posts

Reon Hanbaga
@ReonHanbaga
俺だ、Leon Hamburger。悪い科学者に日本人にされちゃった~


Oh, I think I misunderstood your comment about 20-80%; I thought you meant that 20-80% of our enjoyment comes from problem-solving, whereas I think you meant "20-80% of mathematicians are primarily interested in problem-solving." This sounds right to me; maybe even more than 50%. As for the future, I think there are some meaningful differences from undergrad, though the precise equilibrium shape of math research probably depends on the shape of AI capabilities in a way that seems very hard to predict. At least one is this: I think it's likely that many fundamental questions we now consider important will remain open even as AIs become very capable (and in fact, that AIs will help us discover many more fundamental questions). So we will still be occupied in large part by these open questions, rather than primarily trying to understand predigested mathematics.







“According to DeepSeek’s data, DeepSeek V4 is about as capable as Opus 4.6 and GPT-5.4, which were released about 2 months ago. However, CAISI’s evaluations, which include non-public benchmarks, indicate that DeepSeek V4 performs similarly to GPT-5, which was released about 8 months ago.”








Becoming a ‘Kazakh-boo’ Kazakhstan nerd because Japan is too mainstream









