[Cryptography] "Anthropic A.I. Model Finds Flaws in Tough-to-Crack Encryption Algorithms"

[Michael Kjörling]

On 31 Jul 2026 00:11 -0400, from leichter at lrw.com (Jerry Leichter):
> Obviously the Times is not the place to learn any technical details.
> Anyone come across a deeper description?

At least Ars Technica's writeup links to the actual papers. I haven't
checked those myself.

https://arstechnica.com/security/2026/07/mythos-uncovers-crypto-weaknesses-that-went-unknown-for-years/

https://anthropic.com/document/hawk_key_recovery.pdf

https://anthropic.com/document/aes_mobius_bridge.pdf

-- 
Michael Kjörling
🔗 https://michael.kjorling.se