@ChamomileHangover @waldoj ChatGPT will quite regularly tell me when it feels that itās not possible for it to do something.
There are certainly myriad examples of ChatGPT flat-out making stuff up.
There are *also* myriad examples of people *maliciously* getting it to say something incorrect.
Both of these are flaws, and both of these could have dire consequences if the tech is misused, but both of them should be considered fairly separately. The āturns outā shovelware opinion pieces and social media posts tend to be the latter. People that conflate the two are at best not smart / knowledgeable enough to be any sort of authority on this subject, and are at worst bring intentionally misleading.
The ChatGPT cheerleaders cop a lot of shit online but thereās just as many internet points to be had by hoodwinking people into conflating blind skepticism and contrarianism with intelligence or wisdom.
Most people talking about this in anything more than a āfamily barbecueā context are having more nuanced discussions than āconscious or notā. L