Lectured by AI about the morality of size difference

[HHunter1]

@foreverlurk Sorry. No I’m the old man. AI bad. Do you want Skynet? AI is how you get Skynet.

HH1

[Olo]

@HHunter1 There’s nothing intelligent about these algorithms. They’re overgrown autocomplete.

[tiny-ivy]

@Olo yep, I use them for text formatting. They’re handy as an advanced find-and-replace function. “Please make this word doc into html text with html tags.” The fact that they can be used to do anything else is a flaw, imho, not a feature than anyone should begin to rely on.

[foreverlurk]

Well is was a rather… eventful weekend over at OpenAI HQ, if you’ve seen the news/drama.

On my side, I did report some of my innocuous prompts as false positives to the powers that be. I didn’t receive anything else since then, but instead of the big bad email I’m getting those, which are friendlier I guess(?) :

No, ChatGPT, I don’t think I will. Oh well, I can work with that (and around that of course).

[Olo]

@foreverlurk That’s a rather sophisticated response, assuming that the tiny subject was shrunken involuntarily.

[foreverlurk]

@Olo prompt engineering w/ ChatGPT requires weaving a lot of context to get the situation and emotions you want. You can’t say some word directly otherwise the watchdogs will be triggered. You need to balance the prompt to get a positive “mood”, and even then, sometimes the model can identify through the innuendo what you are alluding to.

(In this case I was trying to get her inside a cage, so the model’s inference that she was shrunk “against her will” probably stems from there)