I see your point but there's been a Moore's law like trend in compute/dollar that's progressed steadily for about a century and has a very high probability of c

I see your point but there's been a Moore's law like trend in compute/dollar that's progressed steadily for about a century and has a very high probability of continuing. How that maps into AI products is a bit uncertain but there's definitely a trend in that direction. As an example of the predictability, Hand Moravec wrote quite a well argued paper in 1989 predicting human level hardware capabilities would be available in inexpensive machines around the mid 2020s which I think was fairly spot on. He also argued once those capabilities were there, software guys would figure things to use if for.