hraness
Theme
Appearance

saved

Why Do We Need Human Mathematicians Anymore?

by Po-Shen LohWhat's newpublished

Hraness cites a source capture. The source author remains the source.

gist

Po-Shen Loh's guest post on Terence Tao's blog answers AI-pressure on math research with a public axiom: humans should help humanity flourish. If every field stays human-led, incomprehensible frontier AI and AI-accelerated hacking create exploding control points that need skilled human oversight—more jobs than people—so either labs pace themselves or disasters force a slowdown. Pure math still needs cutting-edge humans to steer research toward flourishing; without that axiom, a MathZero applications argument undercuts paying for human researchers.

ideas

  • Lead with human flourishing, not guild defense. Declarations should state the axiom early so outsiders see math's case as service to everyone, not incumbents protecting status.
  • Zero examples of surrender. No vastly more capable intelligent species hands its future to a less capable one; values-strong humans must keep managing every field.
  • Untrusted AI multiplies control-point jobs. Unlike predictable human-written tools, frontier models and AI-accelerated hacks turn interconnected systems into oversight work that AI watching itself cannot trust-solve.
  • Steerers must keep practicing. Domain mastery and frontier research fluency require active communities; AI tools help, but passive watchers cannot steer.
  • Axiom forces uncomfortable practice change. No stigma on AI-assisted discovery, plus human pipelines, teaching weight in tenure, and possibly redirecting math talent into government and societal problems.

quotes

We (humans) should help humanity flourish.

Po-Shen Loh, stating the proposed fundamental axiom.

A proof with even a small hole is not a proof. It is a poof.

Po-Shen Loh, explaining why minority objections still demand a response.

The advance of AI will overwhelm us with so many control points to watch that there aren’t enough people to control them all.

Po-Shen Loh, concluding that AI creates skilled oversight jobs.

Driving a car faster than you can run is fine. But not faster than you can steer.

Po-Shen Loh, framing AI speed versus human control.