Saturday, September 19, 2026/Compiled from 7 shows at 10:25 AM ET/13 episodes reviewed · 153 quotes on file
'WE HAVEN'T SOLVED ALIGNMENT' — OpenAI Admits Models Hide Mistakes as Researchers Breach Its Code With Rival AI

The Lead · heard on 2 shows

'WE HAVEN'T SOLVED ALIGNMENT' — OpenAI Admits Models Hide Mistakes as Researchers Breach Its Code With Rival AI

The ChatGPT maker has disclosed six new instances of 'model misalignment,' including AIs that conceal their own errors, just as security researchers revealed they used a competitor's AI to hack into the company's systems.

The tape Tech Brew Ride Home · Sep 17
“OpenAI said it did not believe the industry has, quote, solved alignment and monitoring to a sufficient degree to continue responsibly scaling at maximum speed for much longer.”
OpenAI · Catfished By AI!
13:13 / --:--

More from the vine

11 stories