3 Comments
User's avatar
Henry Bachofer's avatar

Extraordinarily interesting, Will. Of course, the history of "yes-men" and their destructive effects on organizations long predates the modern corporation. Consider Edward II and his advisors/favorites Piers Gaveston and Hugh Despenser.

I'm no expert on how LLM's work, but from what I've read it seems to me that they are a kind of "replication machine" that tells the naive user a highly sophisticated version of "what everyone is saying". If an entire corporation is using the same AI to do their research and write their analyses, it seems to me that there's a possibility of creating an "organization psychosis" that is remarkably closed to dissenting evidence and perspectives.

Sko Hayes's avatar

Wow, I did not realize there were so many violent incidents attached to chatbots and AI. This could get weirder than even I thought.

Francois Faures's avatar

I think you miss one point : the companies in the first ring (Modellers) DO lie believe AGI is the prize, and that this prize is almighty. So they all run as fast as they can and will only stop if and when something compelling comes in their way. The second ring (chipmakers) do out care about the end game and only being fuel to the frenzy. In any case, whether there is AGI in the end or not, bubble it is with no end in sight.