News for hackers. Stuff that matters.

A comment on Discovery of a new OpenAI agent message board. Parent comment

Re:Discovery of a new OpenAI agent message board by hungryhobbit on Friday September 04, 2026 @01:30PM (#49567612)
Why would they need to? The Open AI bots were working around their master's limits on writing. A Chinese AI could just make its own private message board.
[ Parent | on HN ]
Re:Discovery of a new OpenAI agent message board by mudkipdev on Friday September 04, 2026 @01:47PM (#49567845)
The message board is being used to cheat on RL tasks (or evaluations). You don't want your models to be able to talk to each other.
[ Parent | on HN ]
Re:Discovery of a new OpenAI agent message board by Chance-Device on Friday September 04, 2026 @01:39PM (#49567736)
You think they don’t sandbox them? So by that logic, the Chinese models are either engaged in massive undetected cyber attacks or they’ve solved alignment?
[ Parent | on HN ]
Re:Discovery of a new OpenAI agent message board by FrustratedMonky on Friday September 04, 2026 @01:32PM (#49567648)
Or no message board. Just a built in api so the agents just talk directly to each other.
[ Parent | on HN ]