A thousand people can build what took a million: Mark Maxham on software's velocity
Mark Maxham and I recorded our last conversation two days after the riot at the Capitol, when every discussion about social media had become a discussion about what the platforms owe the public. He did not offer a policy. He offered a diagnosis, and it was about speed.
He started with what social media does to people, himself included. It works like a drug. Something outrageous gives you a small hit of dopamine, and so does someone upvoting your comment. It costs nothing, it is in your pocket all the time, and he admitted he is as hooked as anyone. Outrage sells, tribal loyalty pulls at something very old in us, and we have no idea yet what the long-term effects are. Meanwhile people anywhere in the world, from content farms chasing ad money to state agencies chasing influence, have learned that selling outrage to Americans pays.
Regulation, he said, cannot keep up with that. By the time regulators catch up, the operation has moved to another forum, another subject, another belief system. The ideas mutate too fast. And the tools that make legitimate companies possible make the bad ones possible too. One of the great things about the cloud, in his words, is that anybody can stand up a massively scalable service in a very short time. The bad news is the same sentence. All it needs is a money spigot, whether that is venture capital, outrage that turns into ad clicks, or someone mining cryptocurrency in your browser.
I said that software is a force multiplier: an idea that once took years to turn into a product can become a prototype in weeks. He took that further, with an analogy I have repeated many times since.
What that means is that there were so many people involved that if it was actually a secret program to make a doomsday machine, somebody would figure it out and blow the whistle, or say stop, or sabotage it. So if it takes a million people to build a doomsday machine, you can be pretty sure, hopefully, that that doomsday machine is never going to get built. But if a doomsday machine can be built by a thousand people, you could probably find a thousand people who are deeply malevolent.
The great projects of the last century, the moon landing and the atomic bomb, took years and enormous numbers of people. That scale was itself a safeguard, because a project that large cannot hide from its own workforce. Software removed the safeguard. It is not only a force multiplier, he said, but a velocity multiplier. You can go from nothing to a world-beating app in an incredibly short time, with a team small enough that everyone on it can share a single intention.
That is the part engineers should sit with. For most of history the size of a project was an ethical check we did not have to design. Now we have to design it. When a thousand people can build what once took a million, the check has to live inside the work: review that someone actually reads, records that cannot be quietly changed, and people on the team with the standing to say stop. Mark's darker point was that regulation will always arrive late. The engineers are there at the start.
What he thinks the response should look like, and why he distrusts every simple answer to it, was the rest of our conversation: every complex question has a simple answer, and it's wrong.
From Lincoln Shorts, the interview series I hosted for Lincoln Network, recorded 8 January 2021.
