Fooling large language models just keeps getting simpler

(theregister.com) by speckx | May 7, 2026 | 0 comments on HN