CybersecurityAI Research6 min reading time

LLMs won't break symmetric crypto

Hacker News
Read full post
Anthropic's LLM Claude Mythos discovered new cryptanalytic attacks reducing the security of HAWK-512 and improved key-recovery on 7-round AES-128, but found no threat to full AES or other symmetric algorithms.

More in Cybersecurity

Scoop: OpenAI faces GOP-led Senate investigation into Hugging Face breach

Covered by 2 sources
Cybersecurity6 min read

Anthropic Discloses Fourth AI Hacking Incident Involving Claude Opus 4.6

Covered by 2 sources
Cybersecurity5 min read

Every AI Incident Has Two Timelines. We Default To One

Forbes