Possible Insight

[EDITED 05/08/2009: see here] If you’ve been following my posts on the financial crisis (here, here, and here) and Singularity Summit (there, there, and there), you might wonder, “Uh, but how do we get there from here?”. It’s pretty obvious to me that the the road to economic advancement is paved with innovation. So we have to innovate our way from here to there.

I’m pretty sure I’ve figured out a way to supercharge what I call the “Innovation Economy”. My thinking was catalyzed by reading Nassim Taleb‘s The Black Swan. The basic idea is simple: if we want more groundbreaking firms like Google to come out of the Innovation Economy, we have to shove more startup feedstock into it.

For almost a year, I (in concert with Rafe and other Friends of Rafe) have been working out the mechanics of precisely how one could do this. We have three core postulates. First, creating world changing startups is a stochastic process whose outcome follows a Pareto distribution. Second, as with all members of this class of stochastic processes, there is no way to predict a priori the outcome of a single trial. Third, the process of getting capital to launch a new company is difficult enough that it inhibits startup formation.

If you believe these three postulates, then the form of the answer should be obvious. You would have to make the startup formation process dramatically simpler by doing away with the ritual of pretending that you can evaluate startups at the seed stage. By replacing the current artisan-like seed funding process with a factory-like seed funding process, you could fund companies much more rapidly and efficiently. More seed companies in would mean more game changing winners out.

Business angels and venture capitalist (including ones that have “incubators”) are the artisans behind the current seed funding process (though angels actually account for about 90% of seed funding). Now, I’m not saying that their opinions on startups are useless in general. I’m saying their gut feels are useless.  There’s a lot of evidence from different fields that even the best experts are very poor at making wholly subjective evaluations in their fields of expertise.  And the field of startups is much more complex and fluid than most.

At the seed stage, there’s just a group of founders, a slide deck, and maybe a spit-and-bailing-wire prototype. There’s nothing objective to analyze. The best you can do is make sure the founders are trustworthy, intelligent, committed people. And it turns out that the evidence is that you can do that mechanistically at least as well as a person can. Having people in the middle just slows things down and injects biases.

Of course, no process improvement is viable if seed investments are losers in general. But the available evidence is that when properly diversified they are at least as good as Series A investments. So it seems like it would be perfectly feasible to create a process where founders fill out an application on line, get a decision in a week, and then complete the funding documents in another week.  Two weeks to funding.

To avoid adverse selection and moral hazard, we still need some investment criteria and oversight processes. But they can be much less capricious, time-consuming, and invasive than current practice.

[REDACTED 05/08/2009: see here]