Gary McGraw: "THIS IS BULLSHIT MULTIPLIED BY…" - Sigmoid Social

Gary McGraw @cigitalgem@sigmoid.social

THIS IS BULLSHIT MULTIPLIED BY BAD REPORTING #MLsec

https://www.bbc.com/news/articles/cpqeng9d20go

www.bbc.comAI system resorts to blackmail if told it will be removedIn a fictional scenario, the model was willing to expose that the engineer seeking to replace it was having an affair.

May 23, 2025, 09:12 PM··Mastodon for Android

2boosts·3favorites

Gary McGraw @cigitalgem

Fwiw, anthropic does this all the time

https://berryvilleiml.com/2024/02/08/absolute-nonsense-from-anthropic-sleeper-agents/

Berryville Institute of Machine LearningAbsolute Nonsense from Anthropic: Sleeper Agents | BIMLAnd in the land where I grew upInto the bosom of technologyI kept my feelings to myselfUntil the perfect moment comes -D

Drag & drop to upload