Provably fair means the bet is set in stone on the server side before the user places there's and it is not changed. Using a hash of the outcome upfront proves that the bet was not changed based on the users input. Especially in a design like this, it can not be 100% random.
Can a provably fair implementation be worked into a computer that works more like AI?
I don't think so.
OP please note with seriousness that a lot of people are waiting for the lunch of this great innovations.
Only if it's provably fair.
I think this is going to have to go on the back burner for a while.
It would have to be AI in this type of design since the computer is acting as the gambler and the user is only acting as the bankroll and they get to choose the outcome of each roll. If it were 100% random, the user would just select 47 or some other middle number every single time so that the multiplier would always be low for the AI gambler.
If people think AI is not provably fair because it's not random, even though the computer chooses it's bet BEFORE the user and it's hashed to prove it hasn't been changed, then there's no point in launching. I'm not looking to get labeled as a scammer so if there's any question to the fairness of the current prototype I posted I will not move forward.