zlacker

>>davidb+(OP)
I would translate "not consistently candid with the board" as "he lied to the board about something important enough to fire him on the spot". This seems like the kind of statement lawyers would advise you to not make publicly unless you have proof, and it seems unusual compared to most statements of that kind that are intentionally devoid of any information or blame.

>>fabian+Z8
This coupled with Microsoft’s recent “security concerns” brings up the possibility this is related to them misusing or selling data they weren’t supposed to?

>>cuuupi+aj
That kind of scenario would indeed be Threat Level Midnight for OpenAI.

Whether they ultimately wanted to profit from it or not, there is $trillions of value in AI that can only be unlocked if you trust your AI provider to secure the data you transmit to it. Every conversation I’ve had about OpenAI has revolved around this question of fundamental trust.