As improvements to manufacturing tech and CPU designs become unable to deliver the improvements that they used to, the cost of computer time will approach the cost of programmer time. As they converge and possibly flip, optimizations will become more useful (and required) to produce the gains we've become accustomed to. I'm not sure how many years away that is.
I agree. Hardware improvement will hit a wall at some point. From then on, all performance improvement will have to come from software optimization. I do wonder if AI will allow that to happen much quicker.
Could an AI just crawl through an organization's entire code and make optimization recommendations, which are then reviewed by a human? It seems like there could be a lot of low hanging fruit.
