We asked seven AI models to do a simple task. Instead, they defied instructions

3 comments

• $WalterGR 10 hours ago

  The research from UC Berkeley was submitted here: https://news.ycombinator.com/item?id=47608445

• $tristanj 11 hours ago

  Mirror: https://xcancel.com/dawnsongtweets/status/203945108300597700...

• $Terr_ 11 hours ago

  [dead]