AI could end encryption as we know it: math breakthroughs threaten public-key crypto
sebkrier · x · 2026-10-08
An AI Frontiers commentary by Govind Pimpale (shared by security researchers including Matthew Green) argues that AI models' recent math breakthroughs — Claude Fable disproving the 3D Jacobian conjecture, OpenAI's Astra solving 10 open problems, and an unreleased model cracking Navier-Stokes — could have real-world consequences where it matters most: public-key encryption. HTTPS, VPNs and E2E messaging all rest on the unproven assumption that certain math operations are easy to do but hard to undo, an assumption already shaken once by quantum computing. AI-powered math might break it entirely.
More from Safety
- Cryptographer: AI doing amazing math puts elliptic curve hardness in no danger — markjeffrey · 2026-10-08
- ColPali-style visual document indices can be inverted: 47% of words recovered, source page ranked first 98.4% — _reachsumit · 2026-10-08
- The AI safety debate has a framing problem: nobody actually argues against safety — Smart_Fly_5783 · 2026-10-08
- Fields Medalist Terence Tao reposts call for mathematicians to cut ties with OpenAI — Eliv_nurotic · 2026-10-08
- Ex-OpenAI Policy Head Miles Brundage: Deep AI Policy Thinking Is Impossible Amid the Chaos — Miles_Brundage · 2026-10-08
- OpenAI fires three key safety employees who drove frontier pacing and monitorability work — NathanpmYoung · 2026-10-08