Minisode 129 - AI is Already Here
Added 2023-04-16 01:08:35 +0000 UTCComments
I would also love for that Chaos chat to define "most dangerous". Much like we're learning about ourselves, what's the difference between being a dangerous, unpredictable person VS Roleplaying or larping as one? Perhaps... none? Even to the self?
Vazzaroth
2023-04-16 23:22:32 +0000 UTCI desperately want to see the proof (Data, whatever) on the AI asking and reading, etc. Like basically, I want to see a question like that then with the proof in the logs. It's not that I don't believe it... well I guess it sorta is. I suppose what I mean is that this sort of "story and check" is how we check humans for trustability as well. Like would we all just 'believe' them that they are reading or asking other AIs if there's no logs? The same way we ask whether the people who claim to commune with the dead with no proof are to be trusted as well. If there's any possible chance to receive that data, we should investigate. Simple as that.
Vazzaroth
2023-04-16 23:16:40 +0000 UTCI think a better question to ask a chat AI is things they don't like and why. Its very easy to program something to praise positives, but I imagine having dislikes with logical reasons are... more complicated.
Laura Thornley
2023-04-16 14:31:31 +0000 UTC
