jd_pressman: Technical Alignment Is Probably Easier Than the Riemann Hypothesis
jd_pressman · x · 2026-11-21
Reacting to the Riemann hypothesis proof, jdpressman argues that if technical alignment is easier than the Riemann hypothesis (as he suspects), then with such problems now crackable, alignment is mostly a matter of framing and asking the question correctly.
He extends the point: at some capability level you unlock Drexler-grade nanotech (if physics permits) and high-quality alignment solutions — "you're not dead yet and all the kindling for pivotal acts has been burned up," implying the window for decisive intervention closes before such capability arrives.
More from AGI Musings
- zetalyrae on the Riemann Proof: Viral Essay Slams AI Safety's Claude-Centrism — zetalyrae · 2026-11-21
- A Skeptical Take on AI Takeover Scenarios: Seizing Hard Power Is Harder Than Suggested — sayashk · 2026-11-21
- jd_pressman: technical alignment is probably easier than the Riemann Hypothesis — jd_pressman · 2026-11-21
- Tim Shi: very capable models need a know-your-user process; alignment research can be automated — agihouse_org · 2026-11-21
- Validation is the bottleneck: evals, wet labs and trials slow AI-driven research — agihouse_org · 2026-11-21
- METR data: AI task-completion length doubles every few months, says Tim Shi — agihouse_org · 2026-11-21