Self-driving motorhomes are the solarpunk future.
Acknowledgements: Co-authored with Opus 4.7 and 4.6 as an experiment in AI-assisted writing. Update: the problem with AI-assisted writing is that the AIs don’t think good sometimes. This essay has a fundamental flaw that invalidates the whole argument. Let me know if you spot it.
This essay was inspired by John Loeber’s excellent piece on how self-driving cars will change commuting patterns. He makes the point that once commute time goes from drudgery to quiet personal time, People will be significantly more willing to tolerate long commutes and the whole landscape of cities and suburbia would change. Some days after reading it, I was trying to fall asleep and my brain decided to take a giant bong hit and went “brooooooo….what if you had a house on wheels… that drove itself??? broooooo… that would be really cool”.
Anyway.
San Francisco rents are what they are because of the people in it. The bay is beautiful, the climate excellent, and the architecture occasionally striking. But this is true of quite a few places. What you’re paying for, when you pay San Francisco rent, is proximity to a particular collection of humans and the opportunities that emerge when they gather: the engineers, the founders, the capital, the conversations, the parties where you meet someone who later hires you or backs you or marries you. The land beneath your apartment is the cost of access.
Economists call this an agglomeration effect. People in dense networks of other people are more productive, more creative, more likely to start things, more likely to encounter opportunities. It’s why a city twice as big as a town has far more than twice the economic output.
But in our current system, to regularly bump into the cool people, you have to own or rent a fixed patch of ground near them. Even though the agglomeration is human, access to it is a matter of spatial geography – as a real estate agent might put it: location, location, location. We’ve welded civic agglomeration and land together so completely that we hardly notice they’re different.
So what we call the housing crisis in dense cities across the world, is really an access crisis. The only way to be near the people who matter for your life and work is to buy a piece of the planet they happen to be standing near, and there’s only so much planet to go around. Zoning and NIMBY politics make the supply tighter, but those are aggravating factors on top of a deeper problem: we’ve coupled human proximity to land ownership.
The coupling cuts the other way too. La España vaciada — empty Spain — is an example of what happens when an agglomeration moves and the houses don’t. Over sixty years the inland villages of the Spanish countryside lost their people to Madrid and Barcelona. The villages themselves are still there, beautiful ruins and slowly crumbling, but the people are not. The same thing happened to inland Portugal, to rural Japan, to the Welsh valleys. In each case the center of economic gravity shifted away, and the people who had been living there faced a brutal choice: leave the homes they knew and loved, or stay in a place that was hollowing out around them.
What would it look like to decouple human agglomeration and land?
Several candidates already exist in partial form. Remote work is one — you cluster on Slack and Zoom instead of in offices, and the agglomeration becomes informational rather than spatial. Coliving and monthly-stay networks are another version of sorts: you rent proximity by the month rather than the decade. Network states, as Balaji and co have argued, are a more speculative version: start with the network, find the land later.
The version I find most interesting is physical, near-term, and delightfully steampunk: self-driving motorhomes.
Full self-driving technology, when it actually arrives, will be transformative for motorhomes, RVs, camper vans and whatever you want to call them. Right now the recreational vehicle lifestyle requires you to drive a giant van around all day, which is exhausting and limits where you can go. But if the motorhome can drive itself while you sleep in the back, the cost-benefit landscape changes dramatically.
Picture this. It’s Sunday evening. You and the kids have spent the day at a nature reserve, hiking until you’re all tired and pleasantly out of conversation. You walk out of the reserve and straight into your motorhome, parked at the gates. You make dinner, put the kids to bed, and turn in yourself. Sometime in the night, the motorhome quietly starts to move. When you wake up, you’re outside the city where you work. You fix breakfast as the streets slip past. At nine the motorhome stops outside your office. You kiss the kids goodbye and step out. The motorhome carries on, drops the kids at school, and finds somewhere quiet to wait until you all summon it back.
This is what life could look like when your home is no longer tied to a single location. You sleep in your own comfy bed and wake up where you need to be.
Of course, there are lots of details to work out, like sewage, water, power, and parking. I suspect the first three will be solved once there is demand for solutions. Parking is trickier because there is a public commons that needs to be managed, but there’s a window of opportunity to get it right.
Fancy RVs currently cost on the order of $100,000, a lot, but still several times cheaper than a typical house. Self-driving improves the value proposition in a way that leads to a much larger market, which (hopefully, but plausibly) leads to prices falling down a learning curve. Personal finance looks very different (and way less stressful) if your biggest life expense falls from “house” to “fancy car”. There’s a whole set of logistical and macroeconomic details to get into here, but let’s move back to agglomeration for now.
Notice what having ubiquitous convenient mobile homes does to the clustering. The same people still meet at the same coworking spaces, coffeeshops and parties, but the houses they sleep in don’t have to be there. The cluster is sustained by participants’ choices: not where they could buy or rent. Agglomeration becomes even more of a pure coordination equilibrium instead of being locked into place by geographical contingency.
More importantly, the cost of creating new agglomerations becomes far lower. This, to me, is the far more interesting part. The frontier of civic self-assembly opens up. What brave new forms of self-organization might we see?
Consider one existing approach: intentional communities. They become a lot more tractable. Most attempts to start new ones fail because of an irreversibility problem: you can’t bootstrap critical mass without first movers, but nobody wants to be the sucker who uproots their whole routine and gives up their creature comforts for what turns out a year later to be an empty field. Mobile housing lowers the cost of being a first mover. Nothing has to be irreversible. You try a place for six months and leave if it doesn’t work. What used to require founding a town now requires the temporary coordination of a few hundred people, and that coordination can dissolve without ruining anyone’s life if the experiment fails.
What effects will this technology-driven increase in mobility have on our urban economic geography?
One, we get to expand our conception of housing beyond the story of supply and zoning. The YIMBY-NIMBY fight is a tussle over who captures the rents from a particular instance of welded-together agglomeration. The way to expand the Pareto frontier of win-win solutions is to decouple the welding, which makes the rents disappear into competition.
It’s unclear to me whether this leads to more distribution or more concentration of people into existing urban agglomerations. Even if it leads to superstar cities winning even more than they currently do, the fact that rent-seeking is harder with easier exit rights leads to huge welfare gains.
Cities will have to compete on what they actually offer. Right now, expensive cities can be badly governed and still extract enormous rents, because the cluster of valuable humans is partly held together by sunk costs — the houses, the leases, the schools, the social inertia. Decouple the houses from the place and the cluster has to justify itself on an ongoing basis. Cities that host existing agglomerations can still win the competition — most of what makes San Francisco valuable is genuinely valuable — but they have to win it, rather than collecting tribute from people who can’t leave.
Of course, not all places will appreciate an increased transient population, for some of the same reasons people dislike new entrants to their places today. But the point of dramatically increased mobility is that people can go where they’re wanted, and avoid the places where they’re not.
Many of us have learned not to obsess over what we’ve given up to be where we are. A job that would have been a better fit, but in the wrong city. A relationship that didn’t survive a move. The parent we see once a year instead of twice a month. We make these choices, and then we make them again, and after a while we stop calling them choices. They’re the shape the mold of life compresses us into.
But that shape was imposed by a coupling of human proximity and land proximity that was historically ubiquitous but is in fact technologically contingent. Once you see that, or read, say, Ada Palmer’s Terra Ignota, which describes a world of globe-skipping flying cars that dissolve nation-state boundaries and allow everyone to have breakfast in Brazil, lunch in London, and dinner in Doha, if they so choose, you can start to imagine what we’re missing.
Not just abstract economic gains, but specific concrete lives we could live. The friendship that thinned because the friend moved. The version of ourselves that might have flourished in a place we couldn’t afford or couldn’t justify or couldn’t reach. This is the small steady tax that location imposes on every life lived in a developed economy. We pay it without noticing. It seems as inevitable as death and the bigger taxes.
But it isn’t inevitable. It really isn’t. It’s an artefact of a particular coupling between human proximity and physical land, and when that coupling breaks — and the technology to break it is being built right now — the tax goes away. Your grandma moves with the rest of the family. You lose your job but keep your home. The relationships that survive are the ones that should survive, not the ones geography happens to permit.
Self-driving motorhomes won’t fix every dimension of the housing problem, and for them to achieve their true potential we need a lot of details, like congestion charges, to go right. But they break the load-bearing assumption underneath it: that you have to own or rent a piece of the planet to be near the people who matter to you. Once that assumption lifts, a lot of things that currently look like intractable zero-sum fights start looking like temporary artefacts of how we used to live.