Anthropic's AI Discovers Cryptographic Weaknesses in HAWK and AES
Some thoughts about Anthropic's new cryptanalysis results
Anthropic's unreleased Claude Mythos model recently uncovered significant vulnerabilities in the HAWK signature scheme and improved attacks on reduced-round AES. While the HAWK result threatens a proposed standard by halving its security, the AES finding remains theoretical. These breakthroughs highlight AI's growing ability to synthesize existing tools into new attacks, shifting the bottleneck from discovery to human verification.
What makes this genuinely interesting — and, frankly, a little embarrassing for the field — is that none of the ingredients are exotic.