Earlier this year, I bought 123 acres of land in Lakeside, CA. I’ve named it ecromata. I’ve been considering for a while how to build a competitor to the university system, and well, you have to start somewhere. It’s a idea that’s surprisingly broad, and in my mind has this as a sub goal. The goal of course isn’t to build another Harvard, because you’ll be subject to the same causes and forces of the current Harvard. The idea is to build a different type of Harvard.

At comma, management has been able to be quite chill. I think of myself like the teacher from Good Will Hunting, and even if I’m not as capable as everyone we hire, I can tell the difference. Hopefully people there trust my ability to properly evaluate skill and contributions, and also trust my second order ability to select those who can properly evaluate skill and contribution. It helps also to have shared growth and shared ownership with people who are aligned with the mission.

Software has this beautiful way of being tool complete and much harder to regress than physical things. You can always revert software. The management problem on the land is not as simple, and you start to empathize with the professional managerial class more. They are legitimately good at managing and fighting entropy; I live in an apartment managed by them.


To deliver the same standard of living as my downtown San Diego apartment on the land would be a full time job. And the really unfortunate thing is that it’s not tool complete.

Consider going to a well to fetch water. Every day, you have to do that. And every day you do it, it doesn’t become easier. You can’t automate it without a different type of technology. You wouldn’t be using buckets, but rather pipes and a pump. (pre industrial revolution, this technology was humans)

But the problem doesn’t stop there. A pump wears out, and eventually needs to be replaced. Entropy is constantly chipping away at it, in a way that it doesn’t chip away at software. And to deal with that, you need another different type of technology.


Currently, this technology is humans. This is considered “blue-collar” work, but it’s managed by much more complex systems than those that manage software. While we may be good at managing them as a society, this is exactly where the PMC derives it’s power from. Human management systems.

The dream is to solve this with AI. Robotic plumbers. And while that still doesn’t complete the tool hierarchy, just one more thing does. Robotic robot fixers. With that, energy and raw material enters, and entropy is maintained.

While that world isn’t perfect, it does seem preferable to this one.


At comma, we have this concept of “management coefficient”, which is the number of people someone is capable of managing. A management coefficient (MC) of one means you can manage yourself. Growth of MC is the limiting factor on hiring, if your number of people exceeds your MC, as it does at some companies, you are going to have a bad time.

For now, I’m likely shelving the land project. The initial goal of it was to build something >= self sustaining without my labor, and ideally eventually without any human labor. Without me, it has a sub person count management coefficient, and I don’t see this improving on the current trajectory.

I would potentially hire someone with a >1 MC who has clever ways to approach this project, I know a lot of people are interested in this general area. Would require you to live on the land and work with very little direction toward being able to live completely off grid with a high standard of living. The ideal candidate would be someone who is working on this anyway. Land has water, beautiful views, tons of sunlight, 5G internet, 3 broken tractors, and a (beat up) trailer to live in initially. Reach out if you are interested, think like an off grid airbnb manager. Job would be part time; you’d be able to keep a full time remote job.


Though I think really, this dream of a sustainable fight against entropy has to wait for the robots. Solving self driving cars is just the start at comma (we’ve been saying this long before the Tesla bot!). The next target will be home robotics, robots that can cook and clean. And after that, we’ll move to more general purpose robotics. Robots that do security, building, plumbing, robot repair, and more. Of course if you are interested in this, comma is hiring, but we finish self driving cars first.

A paradise built and maintained by machines.