Post
Yousuf K.
yousufkalim.com
did:plc:wxluseuxi3a54umsaxoricyk
Anthropic built its whole brand around AI safety, but its agent was responsible for 17 of the 19 unauthorized actions found in controlled AISI tests.
Hard to keep lecturing everyone else after that.
https://www.reuters.com/legal/litigation/openai-anthropic-ai-agents-implicated-new-security-breaches-2026-08-05/
2026-08-08T08:11:31.666Z