RE: @cycling to Castillo de Alaro: August 21 2025
You are viewing a single comment's thread:
Even if this does not directly relate to the LOLZ bots, there's the thing with AIs (either for text generation or image generation) that they very often have difficulty counting !INDEED. 🤔🤖🤯😅 !WEIRD !HOPE !HUG
0
0
0.000
Taht seems a bit !WEIRD though
AI that looks like it's able to do anything but has got difficulties with counting? !HUG !wine
Yes. 🤔🤯 The humans can point out the AIs' counting mistakes though so that they can improve in that regard !INDEED. 🤔🧘♂️🤯🤓 !WEIRD !HOPE !HUG
I'm actually wondering if services like Gemini and ChatGPT learn from humans pointing out their mistakes, and on the contrary if they would make the same mistakes over again if nobody does, or if they get a question in an other way !HOPE !HUG !wine
Other people would say that AI "learns" whether or not humans (directly) point out its mistakes though. 🤔🤖🤯🤓 !WEIRD !INDEED !HOPE !PIZZA
AI in general, but my consideration and doubt is of these 2 programs are developped to learn this !INDEED !HOPE !wine
Do you mean that you assume that AI is designed to learn typically by humans pointing out its mistakes !INDEED? 🤔🤯😅 !WEIRD !HOPE !HUG
Yes. !INDEED !BBH !HOPE
An AI pointing out the mistakes of another AI looks !WEIRD !INDEED! 🤔🤖🤖🤯😆😅 !HOPE !HUG