I think that using an AI at the moment for the prediction is indeed a big mistake, because it is not among the most viable ways to have a reliable answer, but in a few years things can change,or in a few months, it is known that many programmers are working day and night to get an AI that is the main ticket to be able to win lotteries, casinos, whatever, I don't know if they are more advanced than openia, but surely they are close,especially those criminals who like to leave to offline sites for stealing your funds.
It seems to me that it is unlikely that any individual programmer is able to get ahead of the average level of progress in the industry. I proceed from the assumption that if he is so brilliant, then he will not deal with such trifles (and such niches are actually considered marginal - like the club of perpetual motion inventors) but will find application in more serious projects.