Spartacus It pretty much admits everything A couple days ago, I had another conversation with ChatGPT and got it to admit some rather stunning things. I posted a partial version of this convo on Twitter and it received quite a bit of interest. Granted, every reply from this language model should be taken with a grain of salt, for a number of reasons. AIs are prone to hallucinations; if they don’t have a valid answer, they will make one up from scratch. They’re also highly motivated to mirror your questions as much…
