Centralization and Open Source

Act 1: Paradise Lost

The internet, from its very inception, was conceived as a decentralized network with no single point of failure. Early apps, such as Email, IRC and FTP, were built around open, decentralized protocols where anyone could build a server or a client.

But then, centralized services started to emerge. Driven by economies of scale, the venture capitalist model involved funding these companies until they would capture a new market, sometimes establishing a monopoly and extracting rents. America Online was an early example, but was eventually disrupted by the decentralized Web. Today, Facebook and Google are examples of giant social platforms where people prefer to post their data instead of on their own website. They use (and sometimes abuse) vast troves of personal information that people volunteer to them, or that they surreptitiously collect. Even hardcore capitalists might blush at the effects of such centralized monopolies: Peter Thiel, who famously invested in Facebook, insists competition is for losers, but not everyone feels that way.

The game changed even more when broadband connections arrived. Now, people’s internet was “always-on”, and networked apps started assuming you could either be “online” or “offline”. If you were online, it meant their server was reachable on the global internet, and thus all the signals could go through it. Never mind that you might be on a slower connection in India, or on a cellphone in rural Africa. Facebook will build infrastructure for you, so long as all your signals travel through their server farms and your data is stored in California – an offer India ultimately rejected. No matter, Facebook developed drones, and Google developed balloons, to bring access to the global internet to Africa. Why *global* internet access? Because the signals will travel through Facebook and Google servers. Their business models are good enough that they can pay for the infrastructure.

The slew of services, in the last two decades since Broadband became a thing, have been centralized. Facebook, Twitter, Amazon, GMail, Twitter, Apple, Microsoft, Instagram, WhatsApp, SnapChat, Etsy, Uber, you name it – they get venture-funded, get a bunch of people on both sides of a market, and extract rents. They have 1 engineer per million users, or less. Customer support is non-existent. They choose what features you have and what interface you see. They have the people’s data in one place, and sometimes that makes an attractive target for governments and advertisers.

Act 2: Open Source

In the last few decades, another movement has been growing in software – open source. Modeled on how science progressed, it allowed people to build on each other’s work, and collaborate on an ever-growing snowball. The original free software movement was fueled by the Free Software Foundation and its licenses (GPL, GPL2, etc.) which made an ingenious use of copyright to implement “copyleft“. But over time, it came to encompass less restrictive licenses that let people do anything they wanted with the source code (MIT, BSD, Apache, etc.)

The resulting projects – from operating systems (Linux, BSD), to languages (PHP, Python), to databases (MySQL, Postgres), to web servers (Apache, NGinx) to web browsers (Mozilla, WebKit), have all created tons of value, far beyond their counterparts in corporate silos (Windows, Internet Information Server, Internet Explorer). They have lowered the barrier for anyone to contribute to the growing snowball, and as a result, the products have become so stable and resilient that everyone has moved to them. (Safari and Chrome are based on WebKit. MacOS is based on BSD.) Economically, supporting open source (and defending them against software patents) started to make sense for large corporations.

The Web itself – arguably the most widespread and successful application using the internet, and the world’s largest development platform – has spawned so much wealth creation partly because it has always been radically open source. Every browser since the original Mosaic has had a “View Source” command. Developers could learn HTML, CSS and Javascript by looking at other websites, and downloading free software libraries. Thus the ecosystem took off, and has led to trillions of dollars in value around the world.

Because of the open nature of these ecosystems, anyone can take the full power of Linux, or the Web browser, and build something on top of it. As a result, Linux has been adapted to run on some toasters, while Windows still runs only on a particular x86 architecture. If the open source model was applied to drugs, instead of the current patent restrictions, we may have had a lot of innovation and cures for the long tail around the world. Clay Shirky has recently given a great TED talk on the trade-offs between Institutions and Collaboration.

Cliffhanger: Stay tuned for the thrilling conclusion!

Thus we see two major trends develop over the last few decades:

  • Increasing centralization
  • The growth of open source software

Today we live in a world where most people have moved away from general-purpose computers and open protocols to use ever-more locked-down devices and ecosystems controlled by a single company. But it’s also a world where more and more of the software that’s developed is released as Open Source. Even that famously anti-open-source company, Microsoft, has now “left the dark side” and open sourced its flagship development environment, Visual Studio. Yet most of that open source today is currently hosted on GitHub, a centralized site (and the successor to SourceForge) that’s home to myriad software projects.

Why do so many open source projects choose to have their home at github.com/myproject? Why do so many brands tell people to go to facebook.com/mybrand, or twitter.com/mybrand?

Technology – or lack thereof – is the main reason things become more centralized. This also explains a lot of the debate between Anarchists and Statists. Economies of scale attract more resources, until a new disruptive technology is able to decentralize the innovations for everyone to access. This is the goal of companies like OpenAI, which believe that datasets and innovations in artificial intelligence should be widely avalable to everyone, and not just concentrated in the hands of a select few.

Today, there are movements to decentralize social networking, and tore-decentralize the web. But things are just beginning, and these movements are only now starting to get the same kind of attention that Bitcoin and crypto-currencies got when they decentralized money. Having discussed the past, the next blog post will discuss the future of decentralization. Stay tuned!

This entry was posted in Uncategorized. Bookmark the permalink.

Comments are closed.