Sure I had to do marketing/outreach to get a following, but I have PhDs in my field reading my side-work I'm producing in the evenings. This stuff is science, without all the fluff and grander. At the end of the day, my readers do go out and test my claims, isn't that more science than academia's replication crisis?
I do something that is nearly free to study, which is possibly why it works. No LHC in my basement.
I don't say I necessarily agree with that position. it would have large implications on how advanced degrees are given, which is part of why the government pay for research in addition to the science output. There for sure are other things that must be handled differently in society as well, but that's the argument as I understand it. To some extent government labs as well as private research organizations to some extent do research outside of universities already, but most science happens inside universities and is combined with education.
A group of professors, postdocs and grad students regularly interacting with one another, attending talks by visiting researchers, going to journal club every week to discuss new papers, etc. will be much more intellectually productive, in general, than the same set of people dispersed and trying to work by themselves.
Of course, the university does also perform bureaucratic functions, like running payroll, making sure the grant money is used on what it's supposed to be used on, and very importantly, organizing classes for undergraduate and graduate students. However, the university is also an intellectual community.