I’m fascinated by experiments using quadratic funding to provision public goods (good overview from @mattsclancy: https://mattsclancy.substack.com/p/optimal-kickstarter). Could this mechanism fund research—and in particular, tools for thought, which suffer from public goods problems?

Two challenges I notice:

1. The crowd will be bad at pricing novel research. QF depends on people accurately pricing goods. But early on, the value of research is often non-obvious. This exacerbates a current problem: program officers often underrate especially original research—and that’s their job!

One way to solve this problem might be to use quadratic voting to assess support for the *investigator* rather than the project. Then the mechanism could assign to “fund people, not projects,” akin to HHMI’s Investigator Program.

2. QF lacks incentives to pay for pre-proposal spec work.

Could the WWW have been funded this way? Say you’re TimBL. You spend years thinking about the problems, publish a paper, then try to raise funding to implement it via QF. What would happen?

Optimal QF would mostly pay for implementation—not your spec work. You could try to raise more to recoup some of that cost, but it’d be hard. Someone else could offer to implement the paper more cheaply: they don’t need to pay themselves back for the up-front conceptual work.

TimBL covered his up-front conceptual work with his CERN salary (n.b. this his superiors did not approve!). If you’re an academic, you might draw a protected salary while doing your spec work. Independent researchers hoping to fund their work with QF would some similar solution.

It’s worth noting that the startup path seems to do a better job here: it compensates founders for spec work with equity.

Say you spend a few years thinking about a problem, capturing your insights in compelling private prototypes—not a public paper—for customers/funders. (cont)

If you’ve reified original and powerful ideas, that would (in many cases) translate into better terms, compensating you for your spec work. You’d keep your insights private, so you couldn’t as easily be “scooped” by a team paying only for implementation.

One solution might be to use a strategy common among academics: write grants about work that’s already done; use grant money to fund a future project. @gitcoin has found that QF rounds do seem to additionally compensate project leads according to their past contributions.

Solutions and/or counter-arguments to these challenges would be very welcome!