Claude AI Just Cracked a Post-Quantum Test Scheme and Found a Faster 7-Round AES Attack

Summary: Anthropic says its Claude Mythos Preview model helped derive an end-to-end key-recovery attack on HAWK-256, a challenge parameter for the NIST post-quantum signature candidate, and a 200- to 800-fold speedup for an attack on seven-round AES-128. The HAWK attack exploits a previously unused lattice automorphism and runs in about three hours and 42 minutes on a 96-core server, while the AES result removes a 256-way guessing step from an existing meet-in-the-middle attack; Anthropic said neither result affects production systems.
What you should know: The HAWK team withdrew the scheme from NIST's additional post-quantum signature process after confirming the attack roughly halves the block size needed in lattice reduction, and Anthropic put the model's API cost at about $100,000 against several hundred hours of human verification - a ratio that changes who can afford to stress-test the cryptography you plan to standardize on.

Read the original at The Hacker News ↗