Have higher standards and fund less science?

If you have a million dollars to spend, you'll be careful and pay attention; if you have a billion, you'll be looking at crude aggregate metrics.

If you fund less science, you will favor older researchers (that cost more) because they are the ones thay can produce more and have inertia from previous research project, and they know how to game the system. Thats a big issue with the NIH (in US) that they are trying desperately to solve. Unfortunately, they also often favor and give too much weight to old researchers in their grant panels and they tend to favor (not always consciously) people that think like them.

this is why we decentralize so that that billion can instead be a thousand millions, and we try a bunch of options in parallel rather than consolidating power in a flawed few, no matter how accomplished those few are (statistically, the consolidated few will always be relatively flawed to the parallel many).

In my experience, this doesn't work.

In my country the R&D budget is quite small in the first place, and there were huge cuts during the 2008 crisis.

The result of the cuts was mainly that the mafias, old researchers with political influence, and people who are experts in gaming the system got to keep their funding. In general, newcomers, young researchers, and genuinely good researchers who aren't that good at gaming the system suffered the most.

No, the truth is that these problems are caused precisely by the insane levels of competitiveness for funding - and solutions must reduce that level of competitiveness for funding, not increase it.

Give money to brilliant scientists with no strings attached.

Let scientists among themselves figure out who is brilliant. If the system gets corrupted, a second group of scientists will emerge who claim to be the real deal. Let them battle it out and choose the smarter group for further funding.

In other words, do it the way it was done before the influx of the administrative overhead.

>If the system gets corrupted, a second group of scientists will emerge who claim to be the real deal. Let them battle it out and choose the smarter group for further funding.

Would this involve an application? Perhaps a committee?

Sometimes I wonder if the people making "why don't they just ... " type comments even bother to read what they have written.

Why doesn't someone just make a browser extension that forces you to read the contents of your comment before you submit it? Then we just bundle it in browsers.

Edit: Hmm, seems I plagiarised https://xkcd.com/481/, but worse. I'm glad I rarely write the first thing that comes into my head.

Given "Planck's Principle"of "Science progresses one funeral at a time", it does not seem that scientists within the system will be the best to figure out who should be favored over them.

That second group cannot emerge if it does not have funding. Einstein was 'lucky' in that he had his job as a patent clerk that allowed him his theoretical studies until he finally found his way into an acedemic career.

Battle it out how? Colosseum fight to the death with giant beakers for weapons? Someone makes the decisions on which group is smarter and that person is thus an administrator of the system. They will in turn hire out sub-administrators to better manage the various groups of scientists and then create a system of grants to remove the inefficiency (and slowness) of groups battling it out. And here we are.

One thing I think about: How would American medical research be different under communism?

Honestly, the early phases would be pretty much the same. There would be researchers who want grants to do experiments and to get the grants they would need to impress boards of their peers. So basically the system we have today.

Randomization.

This is the only strategy, in my opinion, that makes sense. Every year, a certain fraction of the grant money from NSF/NIH/DOE will be awarded randomly, based on a lottery. Your name is automatically entered into the lottery if you submitted a grant proposal. The money per random grant will be constant; there will be no strings attached to the budget regarding scientific direction.

And might not even pay off at all. I think that is the real problem. There is tremendous risk involved in research.

Howard Hughes’ HHMI solves this problem with the idea of funding “people not projects”.