The Building Company

A vignette

The post below is a story about the future, but it is also a story about the policy choices we face today, and their tradeoffs.

A shipping container with treads is sitting in a field. Out of it come a drone and a rover. Together they survey the field, starting along the perimeter and working their way to the middle. They collect topographical data, soil samples—the kinds of things you must give the regulators. But they also collect a 3D representation of the field and its surroundings. The data these robots collect is being streamed up to an agent in the cloud. This agent is making a plan—a plan to turn this field into a jobsite.

First the agent makes a basic sketch of the work to be done: the customer ordered a 75,000 square foot warehouse, with a human office module—one of The Building Company’s earliest SKUs. Then the agent feeds the 3D data into a world model, simulating each step of construction with real-world physics. This is computationally expensive, but worth it: within a few hours, the agent has simulated every step of construction 10,000 times. It’s found the pitfalls and accounted for them. It’s created a detailed schedule and a bill of materials optimized precisely for this site. It’s prepared a legal and regulatory analysis worthy of a team of white shoe lawyers.

Before a single inch of foundation has been laid, the construction crew has spent more time thinking about this job than the average human laborer works in their entire life.

The agent files the paperwork with the city planning department. The city planning department’s agent gives the nod—for this site, a fait accompli. The city optimizes for model cost; The Building Company optimizes for performance.

Larger trucks, carrying traditional excavators running autonomy software, show up. They begin digging, preparing the building’s foundation.

A few days later, several more autonomous trucks roll up to the job site, carrying more shipping containers with treads. The containers roll off the trucks and get themselves into position on the field.

Rovers hook the containers up to the field power source—in this case, a 1-megawatt nuclear microreactor. Once they are in position, they deploy an extendable roof and a charging bay. This is where the construction crew will rest.

This particular site has soil with high clay content, so some of the rovers roll over to the piles of dirt dug up by the autonomous excavators. They will mix this dirt with other cheap materials to make adobe—a building material nearly as old as settled human habitats.

There are no plumbers in this crew. No electricians, either. This crew is organized in a way that would be utterly alien to us now: carriers, shapers, joiners, pourers, sealers. The Building Company, being vertically integrated, was free to re-conceive of construction altogether.

When you bring end-to-end autonomy to bear on an industry, you examine it as an alien might: coldly, rationally, with no sentimentality. You don’t see noble occupations or skilled trades with rich histories and dignified practitioners. Instead you see abstract tasks—in the case of construction, things like carrying, shaping, joining, pouring, sealing. Then you design robots to solve those tasks. This crew has things that crawl, things that burrow, things that fly, things that step, and things that drive.

But The Building Company’s crew has no plumbers or electricians or carpenters; it simply has builders.

The builders get to work on the field, executing the plan created by their agent. They proceed in a manner that no construction company today could ever contemplate, taking advantages of optimizations that no human crew could ever execute. They work day and night (though in many municipalities, work after hours is still restricted). The agent is constantly monitoring every robot and making tiny optimizations along the way, tweaking, for example, the robots’ battery recharging schedule to minimize downtime.

In some jurisdictions, a human safety inspector is required by ordinance, rule, regulation, or statute to be on site while work is ongoing. And when the robots malfunction, The Building Company keeps human workers on call to handle repairs. But for the most part, the agent supervises construction.

Every action on this jobsite is recorded by the cameras the robots use to see the world. Precise telemetry is available for every pipe cut, every tile laid, every inch of concrete poured. The city planning department has a stream of data. They—or really, their agents—can continuously monitor the jobsite for violations of any kind. In a sense it is surveillance, but it is the surveillance of computers by other computers. Some privacy advocates complain, but regulators, insurers, lawyers, and banks love that there is a clear record of every step taken during construction.

So, too, does The Building Company. With all the data they collect, they can make a digital twin of every building they manufacture for customers. Among other things, this enables optimally timed predictive maintenance—and recurring services revenue for TBC.

TBC started out as a space company. You wouldn’t know it looking at them today, though their ubiquitous rovers, vaguely Curiosity-like, are a reminder of their spacefaring roots. Their original ambition was to build a base on the moon. The prospect of human travel to the moon is much less harrowing, after all, if robots have erected fully habitable buildings before a single human boot touches regolith.

It was a neat technical idea, but a small market with uncertain prospects. So after a couple years of engineering and R&D, they pivoted to Earth.

They first discovered product-market fit in the construction of urban and suburban warehouses. Autonomous deliveries, not to mention the robot-staffed kitchens that were upending American cuisine, created a need for a new kind of industrial facility: a small warehouse that can blend into urban environments. This required creative architecture, clever use of materials, and surgically precise workplans—all things that would become TBC hallmarks.

TBC became known for these micro-warehouses. They were far more aesthetically pleasing than any urban industrial facility that had come before. To some, the presence of these warehouses in a neighborhood was a marker of status, an indication that a community was forward thinking. But others complained about how the uniformity of these warehouses had a homogenizing effect on the architecture and aesthetics of cities around the world.

These concerns became louder when TBC made their first move into the consumer market. The floorplan of a home is much more complex than that of a warehouse, so TBC had to redesign all their robots to be able to navigate and work in this new, more constrained environment. It was a major risk for the company, but it paid off: TBC could offer unparalleled build quality, rapid construction timelines, and far lower costs than any non-autonomous competitor.

Because TBC eliminated labor costs (well, it converted the marginal hour of a laborer’s time into a marginal hour of energy cost), they could afford to incorporate advanced technology into their homes: indoor air quality equipment, sensors for health and security, top-of-the-line appliances. Even the concrete foundation has sensors embedded in it. With its digital twin and access to the sensors in homes, TBC’s agent in the cloud could, for a monthly fee, plan maintenance tasks and manage home security. It could even orchestrate smart home automations, noticing residents’ patterns of behavior and modifying the home’s operations accordingly.

But the feature that really drove consumers wild was the kitchen. In an earlier stage of its business (the micro-warehouse stage), a large portion of TBC’s income came from constructing buildings for robotic ghost kitchens. The projects became so common that TBC’s management saw it fit to acquire one of the larger autonomous cooking startups. That acquisition ended up bearing fruit in TBC’s consumer play: every TBC home came with an autonomous kitchen, capable of preparing chef-grade dishes from a diverse array of cuisines. All the customer needed to do was keep the ingredient bay stocked.

Sales of TBC’s home product immediately took off. When TBC began offering autonomous demolition and debris removal, sales really took off.

While critics worried about the homogeneity that would come from TBC’s conquest of the American residence, others had precisely the opposite concern. TBC homes were far more customizable than homes constructed by most traditional builders. TBC’s first home package was orderable through a snazzy online configurator. Customers could pick from a wide range of architectural styles, including modernist, Greco-Roman, Georgian, French provincial, Ranch, and others. Floorplans, too, could be adjusted arbitrarily. Within the four walls of a TBC SKU, just about anything could be customized. The only limitation was floor count: TBC buildings everywhere were limited to one story. The TBC robots, for all their strengths, still could not, for the most part, go up and down stairs.

The critics who focused on homogeneity would turn out to be right in the long term: though TBC buildings did indeed vary in style, they always were recognizable as TBC buildings. TBC gradually “ate” the built environment of all American communities that allowed them to operate—though of course, many municipalities have blocked them altogether. An even greater number of cities have created procedural hindrances for TBC.

A few years ago, agents themselves, working on behalf of corporate entities they created (legal in Eastern Europe and a smattering of US states), became customers of TBC. Their buildings alarmed some, but others pointed out that 83% of the time, there were human investors backing the agent-led companies. The fact that an agent was the CEO was a mere convenience, they argued, and those who objected were already behind the times; we have existing securities law, after all, and it can be applied evenly without dampening innovation.

Some American cities proudly stand opposed to autonomous construction, and there are even some people who view humanmade buildings as a status symbol.

But other, less sentimental places have been transformed. The built environment of the far-flung future co-exists with the buildings of our legacy, diffusing gradually, as most things do. The future must be built step by step, and few understand this more intimately than The Building Company.