There is a lot of similarities, a lot of current framework and systems are based on some researcher or founders that were “hallucinating” and building systems with no roots in fact just speculative uproven concepts cognitive biases and extrapolation that they do everything to materialize/ prove

Reply to this note

Please Login to reply.

Discussion

Just the way humans can overcome emotional breakdowns through practice , AI can learn not to hellucinate through reinforced learning ..right ?

AI relies on supervised fine-tuning and reinforcement learning from human feedback

And these humans never completely eliminate cognitive biases or irrational thinking, AI may never entirely eliminate hallucinations I think

Just like humans can never get over emotions 😅