>>9622321 lb
fixed your meme
>Who would program an AI to make inadvertent mistakes?
anybody who wanted their AI to appear as real intelligence. humans are not perfect. if an AI is masquerading as human, you would allow it to make mistakes. its perfection would give up the game.
maybe perfection was the wrong word. mistake free maybe? AI can be mistake free on an unnatural probability scale. just making the point that if you wanted AI to appear as human intelligence, you would allow for a probability of errors on par with humans.
most of your post makes no sense to me.