Full Circle: September 3, 2017 Snippets
As always, thanks for reading. Want Snippets delivered to your inbox, a whole day earlier? Subscribe here.
This week’s theme: Starting to connect the dots between open source, the Internet, bubbles, and Bitcoin. Plus Chamath gives a preview of the future of funding innovation and capital allocation
Hello again, and welcome back to our Snippets series on bubbles and the innovation economy. Over the past few weeks, we’ve taken a tour through the landscape of productive bubbles: why they form, how they sometime leave productive infrastructure behind, and why open source software has helped create a new kind of “lift” to founders today, who no longer need bubble conditions to raise financing and get started. Today we’re going to connect a few dots. We’ll finish the story of how the early Internet enabled this new kind of open source subsidy, how it came of age during the dot com bubble, and what it means for founders in present day.
Let’s resume our story by going back to 1991 and meeting a young Finnish nerd named Linus Torvalds. He had just attended a lecture given by Free Software Foundation pioneer Richard Stallman, who was on a world tour of sorts preaching the gospel of freely shareable code. Stallman’s message didn’t quite have the impact he had intended (Torvalds wasn’t particularly swayed one way or another about the absolute morality of liberated code) but the timing turned out to be important. Torvalds had been working on a software project that aimed to transplant a UNIX clone onto a desktop computer, having been unimpressed with Microsoft’s operating system and thereby deciding to simply write his own. In Stallman’s spirit, Torvalds figured he might as well release his project to the Internet community in the hope that friendly strangers would help him find and fix programming bugs. So on August 22, 1991, he did. The project, which we now obviously know was called Linux, was an immediate hit. It became a key component of Stallman’s GNU operating system, and attracted the attention of thousands of developers around the world. Within a few years, GNU/Linux started to look like a real, robust, powerful operating system that rivaled what Microsoft was making.
What was surprising to many was that the enormously complex code produced by these distributed volunteer efforts wasn’t just passable. It was some of the best code in the world. And it was being worked on by large numbers of people who didn’t know each other, weren’t being paid, and had no formal management structure. A 1975 book by Fred Brooks called The Mythical Man Month had articulated the idea that adding more engineers to a software project only served to slow it down more; this observation was born out in experience by software teams everywhere. Yet the code produced by the Linux group flagrantly broke this rule. It made no sense. How was this possible?
The answer may lie in the fact that Brooks’ rule rests on a critical assumption — that all of the software engineers are working on the same team. And it turns out that when you’re working with complex code in uncharted territory, you may in fact be better off with parallel, distributed hordes of barbarian hackers with different motivations, different resources, different skills, and different assumptions than you will with a disciplined, structured engineering team who has all read the same manual. The open source software movement didn’t just produce free code; it often produced the best code. The approach was famously articulated in a 1997 essay by Eric Raymond called The Cathedral and the Bazaar, which codified many of these principles of distributed open source software development. It contrasted the “Cathedral” development style of isolated groups of visionary craftsmen with the “Bazaar” style in which Linux had been developed, and was later published by a book of the same name. The work became particularly known for one phrase, known as Linus’ Law: “Given enough eyeballs, all bugs are shallow.” Torvalds himself was unapologetically practical about this: he continued to insist that the practical nature of code development, not Stallman’s moral imperative, mattered more when it came to the virtues of free and open-source software.
If software became a thing in the 70s and 80s, why didn’t open source emerge as the force that it became until the 90s? Part of it was that crucial IP was still protected or hidden, and not yet widely available. But the real answer is most likely because we didn’t have the internet yet. Raymond acknowledged this in later versions of The Cathedral and the Bazaar:
“The traditional Unix world was prevented from pushing this approach to the ultimate by several factors. One was the legal constraints of various licenses, trade secrets, and commercial interests. Another (in hindsight) was that the Internet wasn’t yet good enough. Before cheap Internet, there were some geographically compact communities where the culture encouraged Weinberg’s “egoless” programming, and a developer could easily attract a lot of skilled kibitzers and co-developers. Bell Labs, the MIT AI and LCS labs, UC Berkeley — these became the home of innovations that are legendary and still potent. Linux was the first project for which a conscious and successful effort to use the entire world as its talent pool was made. I don’t think it’s a coincidence that the gestation period of Linux coincided with the birth of the World Wide Web, and that Linux left its infancy during the same period in 1993 and 1994 that saw the takeoff of the ISP industry and the explosion of mainstream interest in the Internet.”
The takeoff of the commercial web and the coming dot com / telecom bubble only intensified the demand for what the open source community was producing. The rapidly growing world wide web was getting built on top of open source infrastructure like Linux and Apache Web Server that didn’t just perform adequately; it proved to be world class. By 1998, even Microsoft was worried: a leaked internal memo, now known as “The Halloween Document”, candidly articulated Microsoft’s perception of the free and open source threat to their business model. It admitted, “The ability of the OSS process to collect and harness the collective IQ of thousands of individuals across the Internet is simply amazing. More importantly, OSS evangelization scales with the size of the internet much faster than our own evangelization efforts appear to scale.” The open collaboration enabled by the internet and the infrastructure it helped support became a kind of virtuous cycle; it turns out that when all the nerds found each other, we were able to write a lot of great software. And that software, as we argued last week, has provided the critical lift for entrepreneurs to this day: the subsidy of free and open source software helps founders get started and get funded without needing a bubble like in the past.
We can now look back on the Dot Com / Telecom bubble, in hindsight, and recognize that broadband cables and network capacity weren’t the only kinds of infrastructure left behind in the bubble’s wake. The bubble created the use cases and demand for much of the open source software that has since become the backbone of the web and the mobile Internet. That very same open source software mindset, in turn, helped spread many of the company-building mindsets and management practices we take for granted today: the idea of shipping early and shipping often; the idea of your early users as being your most valuable resource; and the iterative, experimental approach to product building were all strongly influenced by that community. Founders setting out to create companies today get a tremendous lift from what was left behind, and the barrier to starting, funding and scaling companies appears to be durably lower for the long haul as a result. Of course, back then and up until recently there there was no generally accepted business model for how to do “for profit open source”… but now there is. And, just so we can bring this story full circle, now that funding model — the ICO — is getting rapidly pumped up into an enormous bubble. The world has a delicious sense of irony sometimes.
As a brief aside: you may have noticed over the last four weeks that the header images of Snippets issues have each been based on a different company’s stock price before, during and after the bubble. For the trivia inclined: can you identify each? If you think you have some good guesses, let us know on Twitter or email us. A special shoutout will be given to the first to identify all four!
From the cryptocurrency department:
The bear case for crypto | Preston Byrne
Great artists on what they care about these days:
New kinds of competition:
The new off-court play for NBA stars is startup equity | Ira Boudway, Bloomberg
How low-cost airlines alter the economics of flying | Micah Maidenberg, NYT
Disney’s Choice | Ben Thompson, Stratechery
Machines getting smarter (or not):
How machines learn: a practical guide | Karlijn Willems
A wishlist for a printer that actually works and does what we want | Anil Dash
Other reading from around the Internet:
The great Silicon Valley land grab | Financial Times
America’s first addition epidemic | Christopher Finan, in Longreads
Diverging paths to success for North American e-sports leagues | Jacob Wolf, ESPN
Divergent: thoughts on cross-platform (updated) | Steven Sinofsky
Twitch is getting interactive features for streamers | Blake Hester, Rolling Stone
Ben Brafman, the last of the big-time defence attorneys | Jeffrey Toobin, New Yorker
In this week’s news and notes from the Social Capital family,
Chamath dropped in on the legendary Kara Swisher’s podcast Recode Decode the other day; as always, it’s worth your full attention:
Right off the bat, Chamath drops some hints of where we think the world is headed, what is now possible, and what you can expect from the Social Capital team sooner than you may have thought:
“What do capitalists do? What function do they serve? In a well-functioning economy, you have these two things that should come together. You have an economic model, and you have a political model. The economic model that’s worked for centuries, as far as we can tell, is capitalism. And the political model has been democracy; when they’re combined, you get these amazing outcomes, like America. The problem is, a bunch of these things are decaying. Now, we see that both are decaying. And capitalists have a very important job to do, which is they vote with their money for what they want to exist in the future. And when they vote for rockets, SpaceX exists. And when they vote for other things, that are not nearly as interesting as rockets, then those things an exist.
The way that we do our job as capitalists, despite now all of the complexity that we have to deal with, hasn’t changed. So if we think about the thousands of years leading up to 1985, you had mostly dudes making decisions with the sum total of a pencil, a piece of paper, and some combination of an abacus or a calculator.
And not to take anything away from how decisions were made back then, but an investor or a capitalist could actually do their job because the complexity of the businesses and the things they had to figure out weren’t that meaningful. These weren’t dynamic, they weren’t changing.
Then somewhere like 1985, the whole world changed. And capitalism completely changed. And the reason was cause of this weird little tool called Excel. If you really think about the advent of the mortgage-backed industry, the savings and loan industry, the savings and loan crisis, the great financial crisis, the boom in hedge funds, the boom in private equity, what was that? It was a migration from a person to teams of people, and from paper and pencil and a calculator to Excel. And the way that I think about it is, that was fine for the last 30 to 40 years. But now, things are even more complicated! How do you Excel your way to understanding a company like Tesla? It’s impossible. You have to have a more profound operational sense of what they’re doing, and you have to look at fundamentally leading indicators of what they’re doing. When you try to boil things down into a cell, a numerical answer for really dynamic companies, you tend to have bad outcomes.
So my perspective is, instead of having teams of financial analysts, you have to have teams of really skilled operators. People who have built companies. And instead of using Excel, you have to use the thing that other great companies use to build great outcomes, which is software. Data science. Machine learning.
Think about what is now possible. Here’s a great entrepreneur who wants to help, and wants to create change in the world and do something really interesting. Here we are, as folks that have been in their shoes at some point in the past. And not only can we give them money; now there’s something else we can give. We’ve spent time deploying code; aggregating data; deep inside the bowels of hundreds of companies, thousands of companies before. And what that’s left us with are these artifacts. A knowledge base that basically says, when you try this, it works. When you try this, it doens’t work. Across sector; stage; geography. Now, all of a sudden, you can give people a punch list of things to do to make their business better. But you can only do that with software.”
Make sure to listen to the whole episode, as you won’t want to miss Chamath’s opinions on topics such as why VCs should fear the cloud giants, what he did for his 40th birthday, and Beyoncé.
Have a great week,
Alex & the team from Social Capital
