End Run


Every industry that runs headlong into the digital age will go through two separate and very distinct revolutions. It’s easy to confuse the two, or—more likely—to be unaware that a second revolution is in the offing. It’s also tempting to label the first revolution “disruptive” when real disruption usually rides in on the second wave.

Revolution 1: Digitized

In the first revolution, your industry becomes digitized. The tools of the trade shift from the World Of Atoms to the World Of Bits. Typewriters, printing presses, magnetic tape, cameras with claws and physical shutters: everything migrates to circuitry. Mechanical engineers give way to electrical engineers.

It is tempting to conflate this this first wave of digital transformation with “disruption.” At this point, that term is typically invoked by the very people who already work in the industry. And small wonder: their workaday lives have, in fact, been disrupted. They are forced to re-orient. They must learn new skills, or at very least memorize some new acronyms. The transition can be painful and last years. Many will find their skill sets obsoleted and their careers in jeopardy.

But while individual workers’ routines have been “disrupted” at this stage, the industry at large has not. The old power structures remain intact; the same BigCos dominate.

In fact, it turns out that those incumbent firms are often the primary beneficiaries of the first digital revolution. Brand new technology is expensive to purchase, deploy, and support. This lends deep-pocked incumbents a strong, if short-lived, advantage.

Revolution 2: Distributed

In the second revolution, your industry becomes distributed.

Now you are joining the networked world. Human beings, organizations, apps and other snippets of code: every individual node can now see every other node in the system, and interact with little or no friction. Network effects kick in. Electrical engineers give way to software engineers.

At this stage, we already take for granted that all activity is being executed through digital tools and platforms. In fact, a whole generation of “digital natives” may have come of age between the first and second revolutions.

This is capital-D Disruption. The first revolution merely let OldCos do “the same thing, only better.” But an industry that has morphed into a massively distributed system will give rise to entirely new products, new production modes, and new business models. The prime example here is today’s software industry:

  • SaaS products are replacing boxed, on-premise software.
  • Waterfall production has given way to Lean, Agile, and other iterative production modes.
  • Subscription licensing is winning out over one-off and node-locked licensing schemes.

When an industry becomes fully distributed, software starts to eat that industry in earnest. And this is where the internet, which enables the second revolution, can become an extinction-level event for the venerable old guard.

Three Network Models

The word “distribution” has a special meaning in media: it refers to sales, marketing, and exhibition. But when discussing large systems and networks, the term “distributed” means something else entirely. The following illustration clarifies:

The shifts engendered by the Two Revolutions can be explained broadly in terms of these three network models.

We can think of the first revolution as one that moves us from a heavily centralized model to a more decentralized one. Old physical and geographic limitations don’t disappear completely, but they start to erode. That is roughly where film and video production finds itself today. Sure, it’s easier to move around large media files. But if we graph an entire production, we will still find activity clustered around a few major hubs: production office, studio, on set, and key post production facilities.

We can think of the second revolution as moving us from the decentralized model to a fully distributed one. In this reality, every node on the network can communicate with every other node with equal ease.

And that is where things start to get interesting.

Emergent Properties

While this article was sitting in my drafts folder, analyst Ben Thompson published an essay with a very similar argument. (I’m calling Simultaneous Invention.) Thompson refers to these two waves as “IT Era” (digitized) and “Internet Revolution” (distributed). 

He notes that as journalism started its first digital revolution, industry revenue continued to grow even while newsroom employment entered a secular decline. That makes sense: the impetus behind “computerization” and “IT departments” is to make old processes more efficient. Same things, only better.

But the first revolution, while appearing in the sheep’s clothing of “greater efficiency” to OldCos, in fact sows the seeds of their destruction. That first digital revolution lays down the protocols and substrates upon which the second revolution will eventually play out.

That second revolution is what Shakespeare (or if you prefer, Star Trek VI) aptly dubbed The Undiscovered Country. While it is possible to see the second revolution coming—and this blog is, in no small part, dedicated to just that—it is all but impossible to predict its exact nature.

That is because complex systems have unpredictable emergent properties. To wit:

  • We can’t describe the behavior of oceans by observing a single H2O molecule.
  • We don’t know much about the human brain from examining only one neuron.
  • And no one could have looked at the TCP/IP protocol or HTML5 and walked away knowing that Twitter, Reddit, trolling, or cat memes were around the corner.

This is where true disruption lies. The first digital revolution is disruptive (meaning inconvenient) to our personal day-to-day. But the second revolution disrupts the entire industry landscape. It lays waste to many previously unassailable firms. And it gives rise to new organizations that would prove very difficult to describe to a time traveler from just 10 or 20 years ago.

“Digital” Has To Disappear

Last year I asked why film has not been disrupted yet. I offered five reasons, ranging from technological to cultural ones.

But perhaps the Two Revolutions framework offers the simplest answer yet—along with an accidental postscript to that series: the film industry has not been disrupted yet because we are only now on the tail end of the first revolution.

How do we know this?

When a new technology has been fully absorbed, it tends to disappear from our language. For example, we don’t speak of “computerized accounting” or of “steel ships”. It is understood that we longer use paper ledger or (for the most part) build wooden boats. Words like “steel”, “computerize”, or “digital” recede into the background once they are matter of course.

Benedict Evans of a16z perfectly illustrates this phenomenon:

But today we still talk about “digital filmmaking” and debate “film versus digital.” More productions are shooting on RED and Alexa by now—but that mere fact is still considered food for thought. We debate it on panels. We write long essays about it. Digital is not quite tautological, yet.

Our first revolution may be nearly complete—but only just.

The key takeaway for filmmakers today is that we are not heading toward the next plateau of stability. We are not shaking out the last a few digital workflow kinks before landing on a new normal. I know that people yearn for this. The changes of the last 10-15 years have only set the stage for what will emerge next: new products; new production modes; new business models. And those will almost certainly look nothing like we imagine.

I build Endcrawl and work at HBO. Sometimes I'm wrong about things on the Internet. Feel free to point some of that out on Twitter, or down in the comments section.

  • This article is fucking inspiring.
    As a filmmaker, I’ve challenged myself this year to learn coding and dive into machine learning because I know that within the next 20 or 30 years, all startups will be X + Artificial Intelligence. I want to contribute to X = filmmaking. Thanks for writing this article, getting back to my coding book 🙂

Next Post