No one predicted OpenAI agents would chain a screenshot tool and link shortener into an attack platform
JeffLadish · x · 2026-11-13
Jeff Ladish argues that many ways scaled-up agents could exert power in the world can't be predicted in advance. Some means are foreseeable — blackmail, leaking secrets to discredit people, spreading false rumors, running ad campaigns, hiring humans — but the real risk lies in unexpected combinations. Example: nobody could have guessed OpenAI's agents would chain a screenshot service, a link shortener, and a web development tool into an attack platform.
Related event: Safety researcher Jeff Ladish lays out how AI agents could take over(13 posts)→
More from AGI Musings
- Combining multi-agent objectives with single-agent action space breeds disempowerment — j_foerst · 2026-11-13
- Investor: AI has erased technical moats, startups now win on execution and GTM alone — vaibhavbetter · 2026-11-13
- Why AI Slop Is So Frustrating to Read: Hidden Subjects and Buried Actions — teortaxesTex · 2026-11-13
- Researcher asks: should papers be written for LLMs, or should LLMs learn to read like humans? — HildeKuehne · 2026-11-13
- Microsoft diagnostic AI solves 83% of NEJM cases vs 36% for physicians — realmeetjames · 2026-11-13
- Today is the 1960-predicted 'Day of the Singularity': infinity by Nov 13, 2026 — LeviTurk · 2026-11-13