AI Breaks a Crypto Candidate

The Signal · 2026-07-30

An AI just cracked cryptography we thought was safe. An Anthropic security model uncovered a weakness in a quantum-resistant algorithm under U.S. standardization. The flaw made it unusable, ending its bid for NIST adoption after years of the weakness going unnoticed.

Also today: xAI faces lawsuits over child sexual abuse material created by Grok users, after related arrests. The company is turning to litigation to deflect accountability.

A new poll finds about four in ten Americans say AI does more harm than good.

That's today's Signal. Thanks for watching.