0: Summary.

The non-scalability of current major cities poses a challenge to economic growth; it is a burden that falls disproportionately on those least able to bear it.

Unfortunately, any would-be startup city must confront the network effects that pose a tremendous potential energy barrier to the attractiveness of a city that isn't yet mature. From day one, you need to compete with the hedonic attractiveness of San Francisco, in the eyes of the person who decides where to locate the company.

Movable houses mated to modular foundations--nice large customized houses, not small ugly trailers--can make cities more scalable and more hedonically attractive:

• By making it easier for groups to relocate and untangle themselves in a coordinated way, movable houses can increase the natural dispersion of the city as it grows larger.
• Since people could buy and retain customized houses manufactured with economies of scale, movable houses could have better technology and amenities not available today. (Contrast your current house to a modern car.)
• Movable houses could allow unprecedented opportunities to live next to your friends, or to groups of similar-minded people. You and your friends just need to find a set of available modular foundations close together.
• Movable housing might help on key points of governance and realpolitik, for example by making land value taxes more attractive, and decreasing exit costs.

(This was written in response to Y Combinator's request for ideas about the design of new cities.)

1: Introduction: The challenge of scalable cities.

Manhattan doesn't scale well. San Francisco doesn't scale well. Tokyo and London don't scale well. On the present system, each added resident comes with some tiny added degree in inconvenience for every other resident already living in a city, most notably in the form of traffic jams and housing prices. Yes, every added programmer also provides some tax revenue; but the city infrastructure, qua infrastructure, is not scaling well.

This is not all due to anti-housing housing policy, though we've certainly gone to every possible length to shoot ourselves in that foot. Limited road bandwidth is a real issue. The air in Manhattan is genuinely dirtier than in Berkeley--I acquire a sore throat every time I visit New York City.

Building cities that can scale better is a problem of overwhelming human importance. Some analyses have suggested that increased housing costs have eaten huge chunks out of middle-class income; that housing costs may be responsible all on their own for the stagnating middle class. The costs and inconveniences of our most active and growing cities, prevent people in less active economic areas from seeking better fortunes. Cities that don't scale are problems measured in trillions of dollars of lost economic growth and a crushing amount of real human despair.

1.1: All-robotic car fleets are an obvious first step.

The technology for an all-robotic car fleet is probably already available... if we take all the non-robotic cars off the road.

A robotic car fleet makes cities more scalable in many, many ways. The most obvious boost is widening space, compressing time, and removing the cost in stamina to travel. Robotic cars that don't need stoplights or speedlimits might let you get anywhere within 6 miles in 10 minutes. (Say, 2 minutes to get to a thoroughfare, 6 minutes to go 6 miles at 60mph, and 2 minutes to go the last mile.) A 6-mile radius is a lot of territory (113 square miles). You can be 10 minutes away from a lot of friends.

Robotic cars lend themselves to fleets that show up when needed, rather than individually owned cars. If every individual car is being used much more regularly, that changes the cost/benefit calculation for batteries versus gasoline. Which implies less pollution per added resident in concentrated populations. Fewer parking lots implies more space for people, and so on.

But robotic cars, by themselves, may not be enough. Designing a new city is the last place you want to stop and congratulate yourself on the improvements you already have planned.

1.2: The first business challenge is being hedonically attractive to corporate decisionmakers.

Cities run on network effects. Moving to a location with 1000 companies that might want to hire you is far less risky to moving to a location with 3 companies that might want to hire you.

Any new city faces a huge potential energy gradient during its youngest and most vulnerable stages. The gravity of Manhattan and Tokyo and London is absolutely enormous, drawing away your potential residents. You need to offer something extremely attractive about your new city if you want companies to move there instead of San Francisco. And your city's first critical business challenge is about attracting companies, not humans; humans go where there are jobs.

"But we'll have lower housing prices!" you cry. Alas, it's a sad fact of life that the people who decide where to locate their companies face different incentives than the average employee in those countries. (As in the Dilbert cartoon where the CEO is moving the company to what happens to be his old hometown, because that way he can get free babysitting from his grandparents.) I once asked a Google employee why Google didn't start an offshoot campus somewhere with lower housing costs--surely, I ventured, Google had the scale to do that, and enough employees who dispreferred city living. My friend replied that no project manager at Google would want to be so far from the center of corporate politics.

The cofounder who decides to locate their hot startup in San Francisco may have plenty of reasonable and selfless motives to do so--better access to venture capital, better access to programmers. But it seems worth noticing regardless that this decision is being made by someone who can probably afford more Ubers and higher rents than other company employees. If the company did locate itself somewhere cheaper, it could perhaps pay lower salaries and so capture part of the gains from lower housing costs. The decisionmaker who decides where to locate the company may capture a part of the company's gains from those lower salaries. But quantitatively speaking, those fractions do not multiply out to 100%. The incentives are not aligned between the decisionmaker and the rank-and-file.

The overwhelming proportion of the real human good accomplished by a more scalable city comes in the form of schoolteachers who can afford the rent. But if you want that city to boot up successfully, you need to think about how to make it personally, hedonically attractive to the decisionmakers who choose where to put their companies.

Robotic cars help here too! Even compared to taking a private car through traffic, taking a robotic car through an absence of traffic might be more attractive to a CEO.

Slap some flying drones riding the cars, to take off and make deliveries as the destination approaches, and you can have amenities that non-robotized cities just don't have! If delivering hot meals across a 3-mile radius became truly cheap and ubiqitous, so that even non-CEOs could afford it, then we would see much greater specialization on selling meals. Meal-creation businesses would compete on cost and healthiness, instead of restaurants competing on location and dining experiences. It could become typical to order a not-so-expensive, non-high-calorie, home-cooked-style hot meal that would have taken you an hour to make from scratch, and have it delivered in 10 minutes. Maybe the CEO of a big company has a home chef who can compete with that--but it's an attractive amenity for the CEO of any middle-sized company or hot startup.

Again, the vast majority of real good comes from giving the thousands of other residents access to healthy food without a huge cost in personal labor. But you have to appeal to the decisionmakers before anyone else gets that chance.

Having remarked on this cool and futuristic amenity, we are still in no position to relax and declare ourselves done. We need every possible attractive feature for our shiny new city to exert a draw that is remotely competitive with Manhattan.

1.3: Inside or outside the US?

I will digress at this point--though it will turn out to not really be a digression--to consider the possibility of locating the New City outside the United States (and outside the UK and the EU and Australia).

By far the hugest attraction of this possibility for cofounders would be EASIER IMMIGRATION. If your city is otherwise competitive in terms of first world amenities and safety, has lower housing prices, and cofounders can bring people there without Eternal Visa Hell, startups will flock to you that couldn't form anywhere else.

Avoiding regulatory molasses will be another huge attraction for some companies and potential companies. Cough cough PATENT MONSTERS cough cough.

If the new city were located in, say, a special district negotiated with the government of Uruguay, this might permit some basic utopian legal features. For example, not living in a country where marijuana is theoretically illegal if you're white and rich, and potentially life-destroyingly illegal if you're black or poor. For many cofounders or company-owners, this is a point of principle rather than practice; but to some people it is a matter of practice. Or some people, with decision-making power even, may dislike at a deep emotional level the prospect of living in a country where a SWAT team can kick down their door at any time.

The United States is unique among nations in insisting that it can tax the income of its citizens even if they live and earn outside the US. But the tax policy surrounding your city still matters to anyone who isn't a US citizen, or to any US citizen earning not much more than $92,000 per year.

Act 20/22 in Puerto Rico enables people moving to Puerto Rico for the first time to negotiate a two-decade contractual exemption from US federal income taxes. Puerto Rico is just about the only place on Earth, inside or outside the US, that a US citizen can go to not pay US federal taxes. For this reason I'm fond of saying that the three obvious places to start a new city are 2 hours from the Bay Area, Puerto Rico if you're going to be located in the US but not near any particular existing metropolitan area, and Uruguay (hurricane and earthquake free, extremely sensible and stable government). Not paying federal taxes in Puerto Rico would be a hugely favorable attractive gradient--for the decisionmakers that locate companies, in particular, but heck, even for programmers making a dinky $150K per year.

International development has its own negatives. Most obviously, the extra barrier posed by being a lot further away from existing centers of gravity; compared to starting up 2 hours from the Bay Area, or in coastal Oregon, or Nevada. It's one thing to live a 2-hour drive away from venture capitalists, and quite another to be separated from them by a 6-hour plane flight.

There may also be language issues--especially if your company needs to interact with government agencies whose employees don't speak your language. Google Translate will only continue to improve in the future, lowering language barriers further; but the technology isn't there right now to render international communication frictionless.

These barriers will be particularly fatal during the early stages, when the potential energy barriers surrounding your city are the highest.

What we really want--one might think--is to initially start up our city 2 hours from the Bay Area. But to have some mysterious feature that made it unusually easy for many of the city's inhabitants and whole companies to relocate to coastal Oregon, to Puerto Rico, or to Uruguay, once the system had been proven.

Especially if this mysterious new feature also improved the hedonic attractiveness, the cost profile, and the fundamental scalability of our startup city...

2: Movable housing.

Let's assume at the beginning that we're talking about an early town small enough to be composed mainly of houses instead of apartment buildings.

Imagine that in this town, houses are portable objects with standard connections that match up to standard modular house-foundations. We can disconnect a house from its water and electricity and Internet cables, have a standard vehicle pick the house up from its foundations, and gently drive the house over to somewhere else inside the city.

We could also imagine apartments that slot into towers on rails. We can imagine office buildings built up from modules that could be taken apart at need. But as we'll soon see, this new city might be able to scale a lot further before it needed apartment buildings and office towers.

Movable housing would require some amount of new technology, and more importantly, a green-field city--the roads need to take the weight (unless we can replace or supplement carriers with skyhooks or other lifters); the robotic-vehicle idiom would also help, because it means we can clear any road of cars as a house-carrier is passing. I'm pretty sure all of this is within reach of human technology; the question is whether it's too expensive.

And whether anything is 'too expensive', of course, depends on what benefits you're buying.

Benefit #1: You can live closer to where you most want to live.

In the world today, when you want to move to a new city or district, you hunt around until you find a house that is shaped sorta like you want a house to be shaped, which isn't located too far from where you wish you actually lived. And all of your friends and coworkers--network effects will shortly become very important here--are implementing that same process.

Movable housing decouples the problem of finding a good location from the problem of finding a good house. It disaggregates the two businesses, as my father would say.

To find a nice place to live, you just need to find any available plot whose modular foundations match your current house. (There'd better not be more than three varieties of foundation, though, or we're back to the same old matching hell.)

Benefit #2: You can own a better house.

A modern car contains vastly more interesting technology than a modern house. That's because there are big-company car manufacturers with centralized factories and R&D departments competing to offer the car that you'll like the best.

Imagine if instead, a boutique car-constructor company sauntered over to a parking space and built a car there over a year or so. And then you looked around for an available parking space that wasn't too far from your house and had an associated car that was acceptable. If no car was good enough, you could buy a parking space and its car, trash that car, and have a different car handcrafted in that parking space over the next year by the small boutique car company of your choice.

Cars would be a lot less advanced, to put it mildly.

You wouldn't realize what you were missing. You'd never have seen cars with electronic stability control and automatic parallel parking and collision alarms and radio keyfobs. Maybe some rich people would have hand-constructed cars with primitive versions of those features, to go with the gold-plated ashtrays.

So if you can buy a house from a company that has a real R&D department and is actually trying to please the customer and has real competition… well, I'm not an entire R&D department myself. But my personal starting list of feature requests for an Apple iHouse might include:

• Completely blackout shutters over my windows, which open gradually and automatically at the time of morning when I actually wake up.
• Noiseproofing. Extremely serious noiseproofing, with thicker walls, better-sealed doors and windows, active sound cancellation, and maybe white noise generation if that still wasn't enough. When I close my door I want silence.
• Centralized, extremely powerful LED lighting modules with the light carried by optical fibers to whichever room I was actually using at the time. Dimming to red as my bedtime approached, of course.
• Centralized humidification with per-room control plus per-room temperature control, so I'm not running a loud hot humidifier in my bedroom at night. (I need a lot of humidity, personally.)
• A Jacuzzi on the roof, open to the stars. (You think this is some kind of huge luxury and that I'm spoiled for even wanting it, but that's because you live in a civilization where everything house-related costs too much.)
• Anti-insect laser bug-zappers on the roof which fry any mosquitos, wasps, hornets, other stingy things, or even houseflies that get too close. (Yes, I know we're all supposed to be too tough to be bothered by little things like insects; just like there was a time when British people made fun of people who carried umbrellas on drizzly days. I observe that you do, in fact, work inside an office instead of outdoors. I bet those little annoyances you're supposed to be too tough to care about have an awful lot to do with that cultural decision in practice.)

Finally, bigger rooms! More storage space! Relatively empty square meters in a house just should not cost that much on the margin!

Extra marginal people impose costs on cities. Electricity costs money. The central LED lighting modules and Powerwall batteries will cost money. An extra square meter of floor or ceiling, even an extra story on a house, should not cost that much more on the margins.

It does feel to me like the small, crowded apartments are an arcane penance that we are imposing on ourselves. Even people who don't like expanding cities, ought to see more people as being the issue, not more square meters of living space; having more square meters per person seems like something our civilization should be able to manage. Our ancestors lived in bigger houses. We really ought to be able to get that back.

(This does mean tolerating higher roofs, or larger apartment-tower frames for modular apartments on rails, or less grass on the hill. Or living further away with lower local population density; see Benefit #4.)

Some of us have chosen to live in large house-complexes with friends (though it's not obvious how much of this is just due to modern housing costs). Imagine a central 'house' that has the kitchen and the game room and the Jacuzzi, and six modules radiating out from it, each module with a bedroom, bathroom, storage space, and small reading room. Want to try out a different group house for a month? Pick up the bedroom module, connect it to a different house, put it back down again.

Your own wishlist would look different, I expect. And in another 5 years the Apple iHouse 4e would offer us features that neither of us imagined. But one point is sure: under the present system of the world, we'll never get a chance to find out.

Benefit #3: Friends living near each other.

All sufficiently tight clusters of friends can find a set of empty foundations, rent them, and live literally next door to each other.

We could have neighbors again.

We could have tribes again.

Yes yes, I know, you're thinking of that one person who's friends with a bunch of your friends but who you don't want on the same block as you. But really, human beings have been living in tribes for a while. You can probably work it out. Many people could benefit a lot from living in close proximity to fifty people whose company you can tolerate--not necessarily your coworkers, even; you might be choosing to live next to your fellow D&D 3.5e roleplayers who also like dubstep.

It could give us back something human that we lost, and start to make headway against lives that are sad, never mind the poverty.

This notion will admittedly create some coordination problems: "Does everybody in the block need to be in our group for us to have the kind of... parties... that we want?" and "How do we try to make sure that a lot of foundations go empty at the same time so a new group can move in?" and "Uh, what if 60% of the group prefers somebody not move in to their cluster, is that going to be enforceable at all and doesn't that imply a whole new level of local government?"

I can think of some obvious policies to try, but I have a strong suspicion that policy v.1 won't work all that well and it will take until policy v.3 before work tolerably well. Sometimes you just gotta try things.

Benefit #4: Small towns can scale better.

And not just to larger sizes, but to higher levels of economic activity.

This is the prediction that I'm least sure about. But one of the reasons why I'm talking about movable houses, instead of modules on rails in a huge apartment-tower frame… is that maybe, a lot of us won't even need the huge apartment-frames?

Maybe you, dear reader, positively prefer to live in a huge apartment building. Maybe you don't care where a robotic car can take you in 5 minutes. Maybe what you really want, more than any other style of life, is to be able to walk out onto your sidewalk and see lots and lots of human faces (even unfamiliar ones), then turn right and walk into that lovely coffeeshop that is downstairs and half a block to the right of your eighth-floor apartment. Maybe you like the hustle and the bustle. There is no arguing with terminal components of the utility function, as David Hume observed.

Some of us--and I'm not saying we have better preferences, but we have different preferences--some of us would rather live in Rivendell.

Or failing that, we'd like to live in a quiet little house on a green hill, where the technologically advanced soundproofing doesn't need to work that hard.

If we had robotic cars and movable houses, it would be a lot easier for all of us who work at my nonprofit to not need to live downtown--not even in downtown Berkeley. Part of the reason we located in downtown Berkeley is that it has sufficient density of housing that we originally could, and new employees still can, find available non-horrible apartments within walking distance of the office.

But in a city with robotic cars and movable houses, we could perhaps all live in the same section of green hills, 10 miles or 20 miles from the skyscrapers, and move our office modules there too. Robotic cars could teleport us to the large, centralized supermarket that served a big area and therefore had just as much selection as a supermarket in a big city. Or some of us might live closer to the skyscrapers, or with the Burning Man tribe; and for those who made that choice, a robotic car would take them to the green hilly workplace in 24 minutes while they browsed their cellphones.

24 minutes isn't far from the time it takes to cross San Francisco right now! Anyone who considers that good enough could live anywhere within a 20-mile circle, served by robotic cars with no traffic lights cruising at 60mph down the central throughways.

Even so, some offices and some people would need to be closer to the center of gravity, or would just yield in preference to the siren call of network effects. Sometimes there are just too many graph links that all need to be located within 5 minutes of one another. Those companies might still need to live in office modules slotted into towering office-complex-frames at the New City's center.

But if you can coordinate locations more cheaply, move more cheaply, and travel more cheaply because of the robotic cars, then it becomes a whole lot more feasible to have more software companies located in the quiet hills. It's not a panacea and it won't work for every organization, but more people will be able to live in real houses like our parents owned.

Even more of us will be able to live in Rivendell when virtual reality tech matures, which will loosen (though not cut) the bonds of spatial distance that much further. Or VR might not really make any pragmatic difference, but it's worth trying to think 5 years ahead at least; the headsets will improve. It will be one more marginal force exerting a bit more quantitative push towards the attractiveness of living in a bigger house in a quieter place, if your new city can offer that with fewer than usual disadvantages.

Benefit #5: Land value taxes.

Economists since Adam Smith have observed that land is the ideal thing to tax. Literally, tax the square meters of planetary surface. We can't make more land, so it's not like a land tax discourages the production of land, the way that income taxes discourage work. Somebody is going to collect the implicit rent on land; so long as the relevant collector doesn't tax more than the price point at which the supply of land balances the demand for land, the tax doesn't change that price. It's a frictionless tax on a HUGE flow of land rent, and the alternative to taxing that huge flow of land rent, is frictionful taxes. Taxing almost anything other than land (except carbon dioxide emissions), before taxing land, is one of those insanely stupid aspects of modern civilization that make economists want to stab themselves to death with a butter knife.

With movable housing, the houses are moving around and the foundations are staying in one place. This makes it even more obvious that you might as well have the rent on the foundations supporting government services.

I pay $2500/month on my little 2-bedroom apartment in Berkeley, the supermajority of which is certainly land rent. That land rent is probably more than I pay in state and federal income taxes. Which makes me want to gouge out my eyes with a spoon. Because--no offense to my innocent landlord who paid a corresponding price for that land and is not earning an excess profit in on it after mortgage costs--I am paying two huge taxes where an even slightly saner system could be charging me one tax.

Yes, you can see how it might go wrong if one government tries to own all the modular foundations and therefore all the urban land. (The more so if there are no competing cities: see benefit #6.) But the economic factors here are huge enough that it seems worth trying to do things the sane way just once.

Since most taxes are federal, making real headway on eliminating the 'double tax' (income tax plus land-rent) might require locating in Puerto Rico or Uruguay. But even without that, to whatever extent the modular foundations have a natural equilibrium rent, that rent can provide for fire trucks, maybe even health care if the land rents equilibrate high enough--without additional taxes. Where, again, the current model is to have the residents pay one stream of highly frictional tax-rent to the government, and an entirely separate stream of land-rent.

This isn't really a technology problem. But having houses moving around, so that the rent collected on the stolid immobile foundations is entirely separate from any handcrafted structure nailed to them, makes the solution that much more obvious and maybe more politically feasible.

Benefit #6: Exit threats and political relocations.

When Patri Friedman proposed 'seasteading', a lot of the hoped-for good systemic properties came from the fact that sea-based platforms would be easier to move around. Movable housing can be seen as trying to get several similar systemic properties, without taking on the engineering or political challenges of building at sea.

This analogy extends over to one of the primary political ideas motivating seasteading, the notion of "voice and exit".

When your BATNA to staying in a place is to pick up your houses and/or your office modules and move them somewhere else... then that creates a different negotiating position than when you need to destroy your painstakingly handcrafted structures, create new handcrafted structures somewhere else, rearrange all of your personal possessions, etcetera.

Yes, there's still friction--you have friends who may not follow you, maybe your kids end up going to a different school. But there's less friction with movable houses, the coordination problems are that much easier to solve in groups; it could make a quantitative difference.

This improved BATNA could apply at the level of a whole city that negotiated a special economic zone with some state or country. Maybe it's wacky to think that "Kay thanks bye" would ever be a plausible reply on that scale if the host country tries to "alter the bargain, pray I do not alter it any further". There might just be too many non-movable objects creating too much inertia. But even having a large fraction of the potential victims, having the option of putting their movable modules on a cargo ship and heading elsewhere, could make a difference. It matters quantitatively to a victimizer whether making conditions worse for their victims means that 20% of their victims leave or 2% of their victims leave.

I'm trying not to go on too much of a rant here. But one of the enormous overlooked questions of the modern age is how poverty still manages to exist, when agricultural and economic productivity have risen by a factor of literally 100 since the time when 98% of the population was farming. We have fewer poor people, to be sure, the life of the lowest income quartile is a lot less horrible than it was in the 13th century. But there's still some sense in which it seems a little embarrassing to imagine going back to a world where people managed to survive despite being 100 times less productive, telling them that we are now 100 times wealthier, and then having to explain why there are any horribly poor and desperate people in our country at all.

When a condition is that sticky, we should suspect it to be an equilibrium with strong restoring forces. There must be some powerful factor that makes some people be poor, no matter how much wealth is flowing around--a factor that gets stronger as more wealth flows, even by a factor of 100.

One of the obvious forces that could be stabilizing a Poverty Equilibrium is if the standard state of affairs, for human civilizations in general, is for there to always be a few groups here or there that can extract a little more value. The Ferguson Police Department, issuing 3 warrants per household per year, is one obvious example of this idiom. But you should also be thinking of taxi medallions, licensed haircutters, NIMBY house-owners, and health insurance companies without much statewide competition. I don't mean to single out one group as a target for the Two Minutes Hate. There can just be these endless small sets of local factors with the power to drain one more dollar; and these factors will collectively go on draining one more dollar until they can't drain any more dollars without some victims dying. Actually, the equilibrium for multiple extractive forces is a commons problem--Alice knows that if she doesn't steal a dollar from your pocket, Bob will steal it instead, so Alice might as well steal that dollar even if the result is disastrous. Which means that in many cases the little extractions do continue past the point where people riot.

This is one reason I'm skeptical of the ability of a Guaranteed Basic Income to solve poverty in general, leaving aside various other technical problems. We increased economic productivity by a factor of 100 and there are still poor people. Is a GBI really going to be the last marginal improvement that solves it all? A GBI might still help--just like increasing economic productivity by a factor of 100 helped the people who are still living lives of awful suffering and desperation. But after you introduce a GBI, I'm guessing, there will be a number of factors that start to extract one more dollar here, one more dollar there. The Ferguson police department issues another arrest warrant per household, the state increases its court costs, hey, people can afford it now, they've got a GBI right? And what do you know, almost everyone will still have to get awful jobs just to survive.

So it's not at all a side issue, or a mere bugaboo of the independent-minded, to think about the political power of a cheaper exit. To consider whether mature VR, and to a lesser extent, movable housing, might make it a little bit harder to extract value from victims anchored too solidly to run away. The mobility of labor might affect how fast the poverty equilibrium restores itself.

I'm not saying that corporate taxes are the correct level of organization on which to have any tax at all... but it does happen to be the case that taxing corporate profits located in your country is very hard to do, at least to large corporations, because they just locate their profits somewhere else. Making individual human beings and small companies more mobile would grant them some of the same power of resistance.

No, let me be more blunt. If your shiny new city would otherwise be generating a huge amount of excess value for the people inside the new city, and the people inside the city have no credible threat of exit, the people inside your city will not be allowed to keep that value. There are things in the ecology that like to eat free energy, and your city will not be allowed to keep that energy indefinitely if it is so temptingly available for a little more taking every year. It could be eaten by any level of regional government, or any organization empowered by any level of regional government. If you're dumb enough to let somebody patent the connection scheme of the modular foundations, they can let you build out the city, watch to see how much excess wealth is being generated, and then jack up fees to try to capture nearly all of that value. It could be an invasion of patent monsters under a national jurisdiction that permits them. It only takes one factor that can threaten to shut down your whole process, to extract nearly all of the free energy from your city.

So if your well-meaning goal is to make your new city generate lots of excess wealth that the people inside the city get to keep, you'd better make it as cheap as possible for coordinated groups to leave.

Benefit #7: Lower-frictional flow of people through economies.

I'll finish by remarking that, in a very generic and boring sense, friction is usually bad for economies and reduced friction is usually good. It is a terrifying sign of stagnation that people in the United States, especially the less wealthy states, are moving house less often. Housing prices are probably a bigger part of that problem than any one-time cost of shuffling possessions. But every little bit helps, and if you reduce the physical cost and emotional wear of moving from Point A to Point B, it will matter some. Alleged benefits #3 to #6 mostly reduce to, "Some nice things might happen if we reduce the cost and friction of being somewhere else". So benefit #7 is just the generic observation that movable houses would reduce economic friction in a very generic way and therefore other nice things might happen as well.

We could also see movable housing as a kind of modularity that has the same kind of benefits as modularity in code; you can think about fewer things at a time. Time to extend the city? No need to plan Housing Projects and Development. Your job is just to dig a bunch of new foundations, and hook them up to water and electricity and Internet and roads. There, now your city is bigger. Everything else will sort itself out.

3: Conclusion.

I would put high odds against any of this happening in real life before further events are derailed by a near-lightspeed expanding front of von Neumann machines eating the galaxy (as happens in both good and bad AI scenarios). But if advanced AI does happen to take that long, or if Elon Musk gets bitten by the movable housing bug and makes it happen in two years, I'd enjoy living in Rivendell for some of the remaining time.

Imagine living in a higher-tech bigger nicer house, with robotic cars to teleport you wherever, and drones to deliver hot meals. Imagine that the rents not being so damned high--or even that 18th-century utopia where you just pay rent instead of rent plus tax--has produced a decrease of economic friction and a corresponding economic boom, so that it's less hard for people to make a living and get by. Is that a little more like that legendary future our parents were promised, of which it is said that instead we got 140 characters? Better steel factories didn't just produce shinier steel, back in the day; people made more tools that required steel, and used those steel tools to make other things. Scalable cities are something in that class. Movable houses, I think, might be a significant incremental piece of something in that class.

To me, it seems nearly certain that none of that will actually happen. This essay is not a prediction of future glories ahead of us, more of a wistful sigh. If we lived in a civilization where we could have nice things at that complexity level, we'd already have nicer versions of much simpler things.

But so long as Y Combinator is asking for essays on new cities anyway, this seemed worth writing up. I hope you enjoyed it!

PS: Please pave the sidewalks with the bouncy kind of pavement so that people can run on sidewalks without destroying their knees.