[dwaltrip]

We need to experiment with different ways of organizing our scientific endeavors.

Between postdocs forced to spin their wheels in the mud (or quit) for untold years, the replication crisis, generally misaligned incentives between doing the best work possible vs. advancing one's career, and of course, funding difficulties, I think it's clear that we aren't anywhere close to making the most of our opportunities to advance human knowledge and understanding.

I'm not saying that we haven't acheived mind blowing things in the past few centuries, or even that we won't continue to do the same in the coming decades. I simply mean we shouldn't avoid the hard, grueling work needed to make progress on some of the obvious problems.

We also shouldn't be afraid to be honest about the failings of current approaches and processes. Although, I suppose in some political climates that may be a little trickier​, given the apparent propensity for damaging misinterpretations.

[CuriouslyC]

The current approach to science is pretty much screwed. In addition to the things you mention, everyone in academic science is silo'd and hyperspecialized, so there is very little cross pollination and new thought in a given domain. As a result, we're getting very good at picking apart the details of established domains, but very bad at coming up with big new hypotheses or different approaches.

I really think industrial R&D is the way to go, we just have to shift grant money there in the form of tax breaks in exchange for open access/patents that permit non-commercial use.

[gajjanag]

All fine, but which industry would fund research in e.g algebraic geometry or black hole physics?

[CuriouslyC]

Algebraic geometry research doesn't need funding, it just needs some spare time for people who are interested in it. Black hole physics on the theoretical side is much the same. Astronomy on the other hand probably would get short shrift in an industry based science program, since it is expensive and the gain is pretty much entirely intellectual.

[SubiculumCode]

Bullshit.