Wirth's Law

I think it's just plain not true anymore. Occasionally we add new capabilities that really do need high power, but even then a lot of that is just AI and we're making lots of dedicated accelerator chips.

We already have enough abstraction layers so that an average mediocre dev can work with insane complexity, mega frameworks have enough of a monoculture to attract interest in optimization.

Linter tools are better so we can optimize more while breaking less.

Plus we're mostly free from the tyranny of "Premature optimization is the root of all evil" being used to bludgeon people into not optimizing anything.

For now Crypto is out of the mainstream, we're starting to see decentralized tech without it again.