Our brutal science system almost cost us a pioneer of mRNA vaccines

I have met other casualties of this, and meeting them when I was teaching undergrads inspired me to write https://jakeseliger.com/2012/04/17/how-to-think-about-scienc...

I am in the same position. First completely self driven research project end in two different academic spend about 80k each fighting over which of them controlled the IP (the decided to share it), and totally fail to commercialize it (the both kept asking me how to do it while I had zero financial incentive). Second after a successful grant funding project which I was told I could not PI because I had not yet completed my PhD, while I was the lead inventor and developer on for 2 years and getting zero naming on the publications. I left academic to make almost twice as much at a soul crushingly boring job. I really felt like I was working on something important but hated the grant funding cycles (I was told if I wanted one I would have to write it in my own time and could not ask for support), and was tired of seeing people up and down the ladder claim ownership of my ideas and work.

It seems like this is exactly the kind of thing that a government is useful for. Why not have more publicly funded research institutions that employ them, and put them to work on useful research without having them worry about grant considerations, teaching classes, or the publishing treadmill?

Guys, you dont get it. Science isn't government work. In fact, to the extent that it is bureaucracy work, it is doing fine. But it is so much bigger than that.

I am in the middle of this right now; new professor at top-10 university in CS (with a PhD in biomedicine). Funding science is a frequently-debated topic and everyone is aware there is a problem, but one part is that in the US we have combined four different organizations: federally-funded scientific research organizations, post-highschool education, vocational training (doctors, lawyers), and professional sports.

People think a LOT about the broken science incentives and write a lot about it, but progress is hard. On one hand, it's hard to facilitate what you can't measure, but many pushes towards scientific "metrics" feel a bit like trying to quantify artistic output: is the number of paintings/sculptures/etc the right thing to evaluate people on? What about popularity? What about popularity amongst art critics vs the general public?

The more I think about it the more I think we need to explicitly recognize the need for a diversity of sources _and_ structures. For people who don't know, here's an outline of funding sources. The metrics will be in grad-student-years (think ~$70k) because overhead/indirect-costs/etc. vary wildly. Yes, most of these orgs (esp the NSF) just want to fund labor, and there's an expectation that a lot of it is "training" labor.

- NSF: The big one for non-biomedical STEM, maybe $8B/yr. "normal" grants tend to be small, say 2 grad students for three years. But once you get the grant you can do whatever you want with it within reason. The grants are reviewed by "study sessions" where a team of your peers sit around and rank the grants.

- NIH: Budget is $32B/yr (4x the NSF), and very clinically focused. The staple, the R01, funds ~4 PhD students for 5 years. These are very hard to get and generally go to more senior PIs. It's common for faculty to get their first R01 AFTER they have tenure (7-years into the job, say 40 years old). Once you get the grant you can do whatever you want with it within reason. But the bar for getting one often requires you to have "preliminary data" which is effectively 30% of what you were proposing anyway. As funding gets tighter and more people apply, orgs have responded by becoming more conservative, pushing these preliminary-data-expectations even higher. Note that the NIH is very aware of these problems, and is trying to address them, but the entire infrastructure is dedicated to effectively brutal stack-ranking. Every time they suggest things like "maybe famous professors shouldn't have more than 6 of these grants" there's a lot of community pushback. Maybe they're not wrong? Maybe it is better to give Ed Boyden or George Church effectively a blank check.

- DOD (non-DARPA) : Various orgs like the Office of Naval Research, Air Force Research Labs, etc. fund a LOT of basic science. Each one has several long-standing "program managers" who are former scientists that have an agenda for the sorts of things they want to fund. A lot of this funding is dependent on how excited you can get the PM. Think of this like pitching an Angel investor. If you get in good with a PM they can fund you at a decent rate for a long time!

- DARPA : I have a soft spot in my heart for DARPA because the model is so different from the other orgs. A DARPA Program Manager 1. can only work there for 4 years (to prevent personal relationships between PMs and "performers" getting too cosy" 2. Generally is in charge of one to three programs during the course of their tenure. A "program" is a very specific research thrust with a lot of money behind it ($20-80m). They are CONTRACTS, not grants, meaning that the PM expects you to actually hit deliverables AND they can yank the funding at anytime. But they have had some great hits and deep pockets. When I almost went to be a PM it was pitched as "like being CEO of your own science startup". I don't think that's quite right, but it's not totally wrong either. Other fun facts include it can take two years to get a program off the ground, meaning the PM that STARTS a program is probably not the one that finishes it, so three years into your project you suddenly have a new boss.

- HHMI Investigators: The Howard Hughes Medical Institute is a MASSIVE philanthropic organization that funds biomedical research in the US. Two main mechanisms come to mind: HHMI Investigators, where they identify promising young scientists in academia and give them like $10m ($2m/5years? $1m/10 years? not sure the exact numbers) so they DONT HAVE TO CHASE GRANTS. Many HHMI investigators go on to win Nobel prizes and stuff, although the causal influence is murky (did the HHMI money let them do great work or is HHMI just good at picking winners?)

- HHMI Janelia : Evidently HHMI had so much money that they opened a CAMPUS in the mid-2000s outside of DC where they have 40 labs. Each Janelia "group leader" gets up to 6 employees (mostly postdocs, staff, a few grad students) and effectively an unlimited budget BUT you are evaluated based on publications/resources used. That is, unlike at a university (where if you spend $10m to get one nature paper, it's great! you got a nature paper! And in fact the university _likes_ that you spent more money because they get a cut) Janelia cares about the resources. Also groups can't grow huge (because of headcount limit) so you're encouraged to invest in capital and tech over labor. The catch? NO TENURE. You're renewed every 5 years, and whole research programs are sunset over 15 year timespans (this is new).

- National Labs: A lot of great science happens at national labs, but due to their legacy of nuclear stewardship they can be difficult places to do science (men with guns at the front of the complex!) Still if you're looking for a place to do mostly-science without chasing grants TOO much, and are willing to trade some autonomy to do good work, they seem like a reasonable place. They invest a lot in HPC as well if you're into that sort of thing. Not that much biomedicine though.

I love that there are so many different options, but I still think it's bad that SO MUCH of it happens in straight-up university settings. As I said I'm a new professor and the overhead of teaching and grant writing, coupled with the pandemic, really have me down and brainstorming other options.