Why not create more research labs and professorships?
There are many state universities without great programs right? Also glut of PhDs doesnt seem to be a problem for a lot of engineering fields.
I think the rich folks are more interested in making Institutes- the Allen Institute and CZI being two examples. Institutes which are not affiliated with a university have a lot of advantages to the folks they hire- tend to be in popular locations, with higher salaries, more resources for crazy ideas, less overhead on grants (just a guess, I don't really know for sure), and NO TEACHING REQUIREMENTS (many top research profs hate teaching because it eats so much time).
We collaborated with them quite a bit by the way, and they faced almost exactly the same issues. Salaries aren't that good really (maybe 10-20% higher than in US national labs), they struggle to hire people who usually can make twice the money at Amazon or Microsoft. And you still need some ties to academia.
From my point of view, as a developer who didn't understand anything about the science my platform was used for:
- I think you probably want to have a few PhD students, postdocs and professors. They are the ones who can be creative given the right environment: no pressure to publish, access to sufficient resources (material and human) that enable their work.
- You need more money, but more importantly to spend it wisely. Pay people enough (in my case 20-30% more) and you'll cut the turnover by 2-3x.