I liked your post and agree that this is a really fundamental problem. However, I disagree with this: Also, with smaller labs you can afford to hire more PIs a

I liked your post and agree that this is a really fundamental problem. However, I disagree with this: Also, with smaller labs you can afford to hire more PIs and give a career path to more of the postdocs that you are "training." That would probably be better for the postdocs, but I think it would have a negative impact on the overall production of new science. A PI who runs a big lab with a proven track record of success is going to be a better investment most of the time. Splitting the labs up also means that even more people are exposed to the type of grantwriting inefficiencies that you describe above. What I like about the superdoc idea is that right now (modulo some pretty rare positions) science feels like "PI or bust." I think there are lots of people who can make useful scientific contributions but probably shouldn't be PIs. Figuring out something for these people would be good for them and (IMO) good for the system.