This sounds like saying "internet search engines are a race to the bottom" 20 years ago without realizing that someone may end up as Google and obtain market do

This sounds like saying "internet search engines are a race to the bottom" 20 years ago without realizing that someone may end up as Google and obtain market dominance for a decade or so. It also sounds like you believe you have defined the bounds for what AI will be, and figure we'll just iterate on that until it's a commodity. I don't think AI will be that static. We're all focused on stable diffusion and LLMs right now but the next thing will be something else, and something else after that. As each new technique comes out(assuming they are all published), we'll see quick progress to incorporate the new ideas into various implementations, but then we'll hit another wall, and suddenly big budgets and research teams may matter again. tldr is that it is way too early to make the cynical claim you are making.