How do luddites cope with the fact that AI only ever hallucinates on a handful of irrelevant niche questions that everyone knows (this also goes away on thinking mode with correct prompting)?
>>109112859People still think of AI as a program.
>>109112859How bad is the new model hallucinating to have four threads coping about it at the same time?
>>109112859It hallucinates on everything, the obvious ones are just easier to see, showcase and make fun of.
>>109112914>How bad is the new model hallucinating to have four threads coping about it at the same time?Nah, you are hallucinating there is no thread, there is no life or body, you are a brain living in space.
>>109112859>handful of irrelevant niche questionsIf only... Keeps hallucinating functions for me at work.
>>109112859>le "AI" modelIt's garbage because it has been trained on retards from rebbit and xitter. Kys, you retarded normoid!
>>109112859It hallucinates more with each iteration. Every time they increment the version of the latest model it gets worse. Habsburg AI is there. AGI is not.
>>109112859Hallucinating answers to niche questions is the worst possible outcome for an AI. Common questions can be answered by a normal search engine. It's the more obscure stuff that matters.
>>109112859This was me and my girlfriend the first time I saw herI miss her so much
>>109112859If it's knowledge everyone know, it's most likely written down.If it's written down and prevalent then the LLM should have been trained on it and know it.This does then mean any data it takes in, it can twist and hallucinate.
