We need more Moonshots
You can't make an omelet without breaking a few eggs.
Editor’s Note: After reflecting on the poor quality of writing produced when I was aiming to publish one post a day, I’ve adjusted my goal. I will now aim to publish 1-2 posts per week. This should give me more time to invest in each piece and ensure that it’s something you’ll actually want to read. Thank you for reading.
Moonshots are high-risk, high-reward efforts to achieve an audacious goal. Today, most well-known Moonshots are carried out by organizations like X (Google’s Moonshot factory) or Lockheed Martin’s Skunk Works. This is fine, except that they’re both private organizations and thus aren’t obligated to act in the public’s best interest.
We need our government to start funding more risky, disruptive work in the sciences. Doing so will, in my opinion, propel American science forward and will help to break the stagnation in progress that we have seen in the STEM fields.
Funding agencies are in the damn way
Another large part of this problem is that many funding agencies aren’t willing to take large gambles on disruptive research. This may be changing. The Department of Defense and DARPA have had some funding calls over the last couple years explicitly looking for disruptive research. But, there’s not much of a surprise there as DARPA’s entire reason for existing is to invest in bleeding-edge disruptive work. The DOE-backed National Labs have an LDRD (Lab Directed R&D) program that is intended to support high-risk, high-reward research.
That’s cool and good, but unfortunately these funding organizations are outliers. Most grants and government-based funding (which most scientists rely on) would much rather fund a low-risk incremental gain so that they can be sure to get some ROI (Return on Investment). I do think that investing primarily in high-risk research would still yield a high ROI over the long term, but I don’t have any evidence to back that statement up, so I may be wrong. It’s likely best to hedge the research investment portfolio by investing in both disruptive and incremental research.
Funding agencies need to start allocating a larger portion of their annual budget to high-risk, high-reward disruptive research. A large part of breaking science out of the rut it seems to be in is going to be done by incentivizing researchers to attempt more radical things.
The US Government needs a Moonshot Agency
We have many science and research funding agencies: DARPA, ARPA-E, NIH, NSF, DOE Office of Science, and on and on, seemingly ad infinitum. However, while they are certainly valuable, these are all fractured efforts to invest in the science most relevant to that government branch or division. We need a more concerted effort to fund the best and brightest to attempt truly audacious things. It’s not so important that it be one shared physical location, but rather that it be a large-scale program with lots of autonomy built-in for the scientists. In essence, people need:
Freedom to pursue and evaluate (to about a prototype level) weird ideas
An independent team to help determine the feasibility of said weird ideas once they’ve reached the prototype stage
A framework to determine, if an idea is feasible, is it worth investing large amounts of time and energy into? Put another way, is the ROI high enough if this idea succeeds? And what are the odds of success?
A pot of money to fund ideas that pass both of these tests
Freedom to fail or elect to kill a project if it turns out that it’s not working, or it’s too expensive, or something else
The closest thing we currently have to this (that I know of) is the National Labs. The biggest critical difference is that these things are all done, but they’re done out of order. If you have a cool idea, you have to write up a proposal for it and submit said proposal, thus committing to whatever amount of time and funding you’ve requested for your idea, before you ever get to work on it! This seems crazy to me! I think this is, functionally, the biggest issue with how research is commonly set up. Once you’ve written the proposal and gotten approved and have been given money to research your idea, if halfway through it’s not working, you kind of can’t turn around and say “hey guys, turns out this idea sucks, I need to cancel this project” without getting penalized. Obviously you have to give the rest of the funding money back (since you’re not going to be able to use it for your project anyway), but the penalization I’m talking about is that it is very much frowned upon (especially in the eyes of the funding agency) to say that something isn’t working and cancel it. It’s frowned upon enough that it’ll negatively affect your chances of getting funded again in the future. So, what scientist would say “this isn’t working” and cancel the project, knowing that it’ll hurt their career into the future? More likely, they’ll just use the rest of the project time and money trying to find a way to make it work. Maybe they’ll succeed, but if someone has reached the point where they’d rather cancel the project, they probably won't succeed.
It’s clear this is a problem, and that we should work to build a structure that allows scientists the freedom to be audacious and to fail.
What would a Moonshot Agency look like?
I’m not sure, but I’ve got two (really three but the third is just a mix of the first two) ideas.
One idea is a centralized, federal organization that sets the annual Moonshot budget and suggests core problem areas for researchers to target. Note that this isn’t a requirement “you have to work in these areas”, but more like a “these are the problems that are extra important to us this year so we want to highlight them”. Let’s call this organization the Central Moonshot Agency (CMA for short). The CMA would not only have the jobs mentioned above, but would also have to set some guidelines around things like:
How to evaluate prototyped ideas?
At what stage does an idea have to be prototyped before it can continue on?
What is the full system of checks for a project to grow from prototype to the next stage?
What even are all of the stages a project has to go through?
And it will have to strive to enforce this set of rules. As a central government organization, it’ll be funded by taxpayer dollars.
Another idea is, make a mandate (or even just encourage) that every National Lab or similar organization must have a Moonshot division/center. Make no other rules about it. The lab can decide how big it is, whether they even give the researchers enough equipment to do their job, or they could spend their entire budget on the Moonshot division. Or anything in between.
The third idea, and one I think is most reasonable, is to encourage each National Lab to build their own Moonshot division, but to also provide some rules/guidelines around it. There could still be a CMA, but their role should be much smaller. Basically, the CMA should work to provide answers to the questions I outlined above and should maybe govern/make recommendations about what portion of an organizations annual budget should be dedicated to Moonshots, but they probably shouldn’t do much outside of that. This should avoid much added cost to the taxpayer, while still providing support for researchers to do disruptive work without fear of failure.
There might be some extra cost to the taxpayer, since by nature many risky ideas will fail, but I’m not actually sure that the long term ROI will be any lower than it currently is. Since most research funded in the US right now is incremental, each individual advance often doesn’t actually add that much value to the global knowledge pool, compared to a hypothetical, disruptive advance in research. Probably the best investment portfolio for research is to put a bunch of money into research similar to what’s currently being done but to also put a not-insignificant amount (maybe 10-20%) of the portfolio into “risky bet” research projects. We really want every funding agency to invest a decent amount of money into risky bets, and we need the workflow for Moonshot-style work to match the “prototype → evaluate → increase in resources → evaluate → increase in resources … → project completed” loop. What I think that looks like is giving researchers resources (compute power, lab space, lab eqiupment) and monetary support (paying them) while they are working on coming up with prototypes. When the Moonshot lab is embedded into another organization (like a National Lab) employees could be selected to work exclusively in the Moonshot lab in year or two-year long chunks by basis of a lottery or application system.
Why should the public even care?
The public should care quite a lot about science and how our tax dollars are spent investing in science and research. The primary beneficiary for most US-funded scientific research is the public. Note that this doesn’t apply to privately-funded research. If Megacorp funds me to do some research for them, then of course they are the primary beneficiaries of said research. But if I get an NSF grant or other tax-dollar-supported grant, then the US public should benefit (at some point) from that research. It’s fine to do some research that uncovers some new biological mechanism and it’ll take 10 years until it shows up in consumer products, because when it does show up, it’ll help people a lot. What I’m trying to say is this: It is not the responsibility of the researcher to ensure that their scientific advances actually end up benefitting the public. It is the responsibility of the researcher to pursue problems that, if solved, will benefit everyone.
Always remember that the US government spends our money on many things, but among those things is science and technology research. This is fantastic that they support this! We have innumerable advances in technology, medicine, and engineering that came out of the US gov. investments in science and research. But, as a citizen, you should be interested in how that money gets invested in science. I hope most people would agree with me that it would be better to support some high-risk, but very high-reward research instead of only supporting the safe, incremental bets.
Progress in science, engineering, and general inventions has slowed down. While this isn’t the end of the world (yet), it’s certainly not great. It’s better for everyone if we work to accelerate progress back up to the level it was at 10-20 years ago. Doing this will require some out-of-the-box thinking. One thing I’m sure will help is to give scientists and researchers the opportunity to chase truly disruptive advances. Our current incentive structure rewards taking safe, small, incremental bets that build just a little bit on the work that came before. We need to break the mold a bit here. Change the incentives, reward people for taking big leaps, and watch as they change the world.