Quant pro: A Millennium-prize AI won't 'solve' markets — it's engine-building, not proofs

iaindunning · x · 2026-08-18

A quant/MM practitioner argues that trading, at least as they practice it, is more like building and operating a complex engine — forecasting, modeling, hard-to-simulate control — than a math/TCS problem; "solving the market" like a proof is the wrong mental model.

Reacting to AI's progress in math, they note a math-genius AI helps here just as math skills help many hard problems, but quant is no more solved by such an AI than building and running a tokamak. The open question they're curious about: when does progress on verifiable problems like proofs spill over to all hard technical problems?

Original post →

More from AGI Musings

AGI Musings channel →