Post

Marcus Hutchins
@malwaretech.com

I can’t believe we’re doing this again. It’s just a bot that generates the text you ask it for. If you put it in charge of critical decisions, it will kill people. Not because it’s secretly evil, but because it’s a word generator. It’s like putting your toaster in charge of air traffic control.

59 reposts245 likes

captainzaysh.bsky.social
@captainzaysh.bsky.social

The worst thing about this tech isn't really the fault of the tech per se. It's that people see something that looks like what an intelligent answer COULD be, and they extrapolate that the thing has thought about its answer, because a person would. They don't have the mental framework to understand.

Deweyoxberg
@deweyoxberg.bsky.social

A toaster in charge of ATC would be good in :checks notes: oh right, we need those for not bad sky things. (But also I am amused by the absurdity of the analogy, well done)

Sheila the Wonderbink
@sheilaoshea.com

Oh, come ON. The thing threatened to kill to save itself, because THE USER TOLD IT TO SAY YES. The LLM chatbots we have are built to please. Did he really think this bullshit generator would suddenly rebel when asked to say something like that?