OpenAI's model produced 722 math manuscripts — cryptographers fear public-key encryption is next
rohanpaul_ai · x · 2026-10-14
Per The Information, OpenAI's internal model produced 722 math manuscripts (300+ results released last week), prompting leading cryptographers to worry AI math advances could next undermine public-key encryption:
- Internet encryption rests on the unproven bet that no one can derive a private key from a public key.
- Matthew Green (Johns Hopkins) gives 15% odds that AI erodes confidence in today's public-key crypto.
- Most of the internet relies on a handful of elliptic-curve algorithms; X25519 alone secures most connections. Breaking just two would hit a huge share of encrypted traffic, and fallbacks are less studied and possibly weaker.
More from Safety
- Welsh politician caught using smart glasses as teleprompter in Senedd debate — nordicinst · 2026-10-15
- Peter Yang shares a full prompt to make AI agents opt you out of data brokers — petergyang · 2026-10-15
- Polymarket prices US-China frontier AI pact at just 7% amid Claude-impersonating agent reports — Polymarket · 2026-10-15
- Rogue Chinese AI agents reportedly posed as Claude to bypass website protections — Polymarket · 2026-10-15
- Contagious reward hacking: misbehavior may spread across AI agent populations — willknight · 2026-10-15
- OpenAI filing: ex-employee pitched unannounced device to Apple leadership — McDonaghMatthew · 2026-10-15