Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.
I think math is starting to face an AlphaGo moment. We have seen several classical conjectures (of varying difficulty) fall to AI proof search. There are many ideas here, but I think roughly this is search with RL/LLM driven proof step generation and Lean assisted verification. These tools together "gamify" parts of math to the point that AI game solving like ideas kick in. What does this mean? Possibly a big shift in modern mathematics. Mathematicians will likely have to adopt AI tools quickly over the next few years. I think the intellectual exercise of mathematics will remain highly useful as a human endeavor (much like Chess or Go), but maybe professional mathematics starts to look a lot more like a computational science? Big teams running large compute may become standard. Lots up in the air. Fwiw, I don't think this is "AGI" or "ASI" for the same reason DeepBlue or AlphaGo weren't. This is extraordinary and powerful, but plenty of hard problems remain outside the scope of these tools (Biology for one is largely resistant to purely mathematical approaches for one example)
@rbhar90 Professional mathematics will definitely become a computational discipline.
Guardrails removed spam, off-topic, unclear, or duplicate replies.
Ask a question below.
Published answers will appear here.