Elliptic-curve discrete log already has a poly-time solution — and could an LLM have found it?

jamestagg · x · 2026-11-23

Building on his earlier thread, the author notes that elliptic-curve discrete logarithms already have a known polynomial-time solution, so that specific crypto-relevant problem is 'solved' — P ≠ BQP only rules out a general solution for all problems, not solutions to some. The subtle question, he argues, is how that solution was found: did it require human mathematical creativity, or could an LLM have assembled it by exhaustively searching the literature — linking complexity theory to whether AI can replicate mathematical discovery.

Original post →

More from AGI Musings

AGI Musings channel →