On the origin of startups, entrepreneurs, and the creation of wealth.

South Africa Summer 2014 Blog

Written by: Bryan Hernandez

June 27, 2014

It was in the Galapagos Islands that Darwin finally got the evidence he needed to trasnform what was once only a nagging hunch into a full-fledged theory of the origin of biological species.  Here, finally, was an environment so isolated from the rest of the biosphere that he could convincingly show the true driver of speciation is the need to adapt to diverging fitness landscapes.  It would be some years later that we would understand not only evolution in its mechanistic genetic details but also its characteristics to the degree that we, its beneficiaries, could recreate our own unique instances of it.
This fractal self-similarity—organisms that evolve the ability to control the evolution of others—is too spectacular to pass without mention; but as this is an essay for the sake of entrepreneurs, businesspeople, investors—and not scientists, philosophers, and artists—it will have to be mentioned only in passing.

Entrepreneurship, startups, and the creation of wealth in general is nothing more than an instance of evolution.  But this is not an easy thing to see correctly, so let's start with a straightforward example.

When a cuckoo bird lays eggs the color of blue like those of the warbles it exploits, and the warbles themselves grow an increasing keenness for detecting suspicious looking eggs like those of the cuckoo, we, the ornithologists, the evolutionary biologists, the people who "think lightly of themselves and deeply of the world," we clearly understand that they are in an arms race—a subgame of evolution whereby the need to out-smart the other player becomes the driver for new, spectacular development of function.  The eggs of the cuckoo get more like warbles', and warbles get better at detecting cuckoo eggs.  All that's required is replication, variation and a strong dose of selection.

Fantastic.  But where in this is a connection to entrepreneurship?

Facebook buys Whatsapp for $19 Billion.  Microsoft launches Slate on the tails of the iPad's domination.  Box jumps on the market Dropbox ignored.  Are these events understandable through the lens of evolution?  If so, how?  And how is that useful?

Before I answer that, let's talk about another case a little closer to the mission of MIT Global Startup Labs.

Take the greater Boston area for example.  Despite having one of the highest densities of engineering, science, and business talent in the world—for many, many decades—it still took until about 2009 for a viable startup environment to substantially emerge.  Compare that to one of the smallest, youngest, and most challenged countries by foreign affairs in the world, Israel.  It attracts more investment dollars per capita than anywhere else in the world (http://www.arcticstartup.com/2011/06/15/vc-per-capita-europe-7-us-72-israel-142).  (Two times the United States).  Can a general model of evolution help us understand this?

I assert evolutionary dynamics are as inarguably in these places of business, government, and innovation as they are in any place of great biological diversity on the planet.  But in order to see it, you need to forget about genetics and the mechanics of business and entrepreneurship.  You need to abstract things and see the broader strokes and how similar they really are.

Evolution occurs in a fitness landscape.  Characterizing that landscape and the replicators (animals, companies, or other entities) that inhabit it is paramount to understanding its true dynamics and shepherding its development.

Entrepreneurship is the challenge of finding how to make a living in a moving economy.  It's finding under-recognized opportunities to apply human and organizational capital towards the reduction of entropy.  And this is always an evolutionary process.

We've all heard of "the pivot".  We know the importance of early iteration.  But who knows why?  Would you believe me if I said these were knowable a priori?  And who knows when it's time to buckle down and grind the gears we've already built versus when to break out the white baords, colored markers, drink of choice, and brainstorm new approaches to old problems?  Forunately, all this, too, can be understood in the simple parameters of an evolutionary environement.

Again, I'll try to limit this discussion to the common challenges faced by most new ventures: highly uncertain and hostile environements.

In the natural world, organisms that do well in these environments are those that can adapt extremely quickly.  From the outside, we see a general trend towards high variation and short lifespans.  Why should this be the case?  Because when it's not clear what kind of environment an organism is going to be facing in the future, only those populations that cover a lot of possibilities will survive through rapid changes.

Does this sound familiar at all?

Most startups inhabit similarly uncertain and hostile environments.  Not only are there bigger competitors out there that are doing everything they can to not leave any market available to them, but there is such a big space of possible futures that they cannot depend on having tomorrow what is working for them today.

What this implies for the entprepreneur himself is different than for the people interested in trying to build an entrepreneurial environment.  For the entrepreneur himself, this basically reduces to much of the already well-established and good advice of how to run an agile startup, domain nothwithstanding.  But for the champions of entrepeneurship, the farmers of innovation, this means something different.

I'm sure there are plenty of people who would disput this claim, but in a few domains we know pretty well how to grow entreperneurs in Silicon Valley—at the very least, we've had some pretty good successes over the past few years. There are great methods for producing winners in that environment.  Talk to Paul Graham.  Last I hear YCombinator was doing well, but maybe they're just getting lucky everytime?  Probably not.  Talk to Thiel and Musk.  You can get by faking it for a while with methods that don't work, but you won't be able to build a track record like these guys.  And so this is my point: what produces successful entreperneurs in the Valley (for the moment) is not as mysterious as it once was.

But this is not the case pretty much everywhere else in the world.  If it was, then Silicon Valley wouldn't be so special, no?  And so we—the incubators, angels, and educators of entrepreneurship in developing markets—we are not playing the same game as Paul Graham.  We are only seeking a similar result.   Although we, too, are endeavoring to build successful startup incubators, we are more startup than we are those institutions who already incubate them successfully in the States. 

In other words, we are in a very hostile and uncertain fitness environment.  We don't know what's going to work.  So leveraging our understanding of evolution and the characteristics of organisms that do well in environments like this, we should adopt a strategy that bares this in mind: high fecundity and lowish fidelity—or in the traditional entreperneur's lexicon: lots of iteration and lots of varation between iterations.  We cannot copy too closely the juggernauts in the Valley unless we too want to fight for a living in the same place.  We cannot take too long to travel a full cycle from program start to program end because it will take too long to find out what works.  We need fast iteration on lots of different approaches to the challenge of growing an ecosystem that can support entrepreneurs.  We need to grow it, the method, within the same environment we need it to be.

The South Africa instance of MIT Global Startup Labs is doing just this.  As opposed to hubristically asserting we know the correct way to run a startup incubator and trying to reimplement the methods not fit for the South African startup environment, we are going to do things differently.  Although there is much we can borrow from curricula of entrepeneurship across the world, we will foremost optimize for searching through as many methodological vairants as possible in order to find at least one that is exceptionally viable.  It is not likely that we will produce large volumes of successful startups on this iteration, but the sacrifice we make in quantity will be more than recovered through our discovery of truly successful methods—not my methods, not Paul Graham's methods, and not MIT's methods. 
They will be South Africa's methods, and they will work for South Africa.