Fragments: lack of AI whistleblowers, saving debris from AI bubble bursting, how to decide to skip LinkedIn posts, agentic programming at Zalando, shredding US intelligence
1
Reply
0
Boosts
@mfowler I laughed when I got to the LinkedIn advice :-)
I had a similar thought when listening to the Toner interview. I get why OpenAI wanted to turn off the guard rails to really test the Soul Document or whatever the goal was. But why not leave the automated oversight in notify-human rather than prevent-action mode? How did they plan to evaluate whether the tests were working? Maybe I misunderstand the goal of what they were doing.
PS. & thanks for the gift link--what an important story, on multiple levels.
