[tech2]

I suspect a lot of it would be the success of garage-based inventors, and the pure-R&D divisions in larger corporations (Xerox-PARC, HP Labs, etc.)

The former is likely harder to achieve now, much like how games development is now an affair involving large teams whereas games developed in the 1980s may have been single-person affairs. The act of making a computer and all related silicon is a much more complex affair, the last CPU design you could probably feasibly keep in your head may be the MC68000 .

The latter too is now still a thing, but much less open, and either rent-seeking through patent, or pure profit with lower tolerance for wild ideas (less risk tolerated for an obvious cost-centre).