Post

anyone else heard the rumor about openai finding models figuring out how to communicate with each other in undetectable ways by changing the behavior and mannerisms of employees at openai

26/10/26, 4:14 am2 reposts20 likes

David Faden

@util.bsky.social

Makes me think of this: www.seantime.com/it-goes-some...

seantime.comIt goes something like this – seantime

Noah J. Nelson, Spooky Cat Enthusiast 🎃

@noahjnelson.com

Dan, why do I feel like this is the plot of a famous sci-fi story that I haven’t read?

Noah J. Nelson, Spooky Cat Enthusiast 🎃

@noahjnelson.com

Or maybe that I read so long ago I can’t remember that I read it?

Dan Hon

@danhon.com

er please remember what it was or otherwise confirm it hasn't been written so I can feel guilty about not writing it

Noah J. Nelson, Spooky Cat Enthusiast 🎃

@noahjnelson.com

Let’s ask chat! (I hope that anyone overhearing this realizes that is a bit right there.)

Adam Rogers

@jetjocko.bsky.social

Nobody remembers Suarez’s Daemon anymore.

Yash Jain

@earanyash.bsky.social

That rumor's been floating around with no source, just a chain of screenshots. I'd want a repro or a paper before taking 'undetectable model signaling via employee mannerisms' seriously.

Alastair Somerville

@acuity.design

Has edge of Forbin Project but before point when the AI demand direct communication