Raw LLM Responses

@AshenSpark So we know exactly how AIs come up with weird answers like "You should divorce your wife and be with me"? Or we all know the new agentic features being rolled out on Windows 11 Insider Program are very vulnerable to ZERO-CLICK attacks (I literally asked Copilot)? I would also like to know why I ask AIs to draw me pictures, they keep getting it wrong. From 12 fingers, all left feet, to noodles merged with hair...and then, WHY CAN'T THEY FIX IT? I would expect you to know the answers. (OFC I don't understand...if it is a topic I know and have full knowledge, I would be asking better questions) Frankly, some AI models are releasing newer versions/editions/models in less than a year. If we are getting them faster than a newer iPhone/Samsung models on a irregular basis, I wouldn't say the progression on AIs is slow. BTW, what you and that video are suggesting SciShow is completely wrong on the topic. I can only assume then, I should do exactly the opposite to whatever Hank has said/hinted by fully abusing the AIs and keep listening to nonsensical suggestions from the AIs that could do bodily harms. There is nothing wrong to assume that is the way since I do not understand AIs at all. Isn't that how the teenager killed himself? Not understanding AIs, abused it and listened to whatever it suggested. Nope, I don't feel AIs are that competence. Just look at that black robot slave that fell down on launch...or the robot dog that lets the human tester keep kicking it instead of dodging his attack... And casual viewers like me cannot contribute anything to the conversation. Yes. There you go.