The Parts Catalog.

Construction has been left in the dust when it comes to technological advancement and adaptation.

People keep trying to fix it, and they keep trying in the wrong direction. 3D printing companies want a better way to extrude concrete. There's a myriad of software companies trying to coordinate the mess. A dozen prefab startups have come and gone. Now AI is positioned as the final answer: AI to read plans, AI to catch conflicts, AI to coordinate trades.

None of it is working to have the effect it needs to. And none of it will.

The thinking behind every one of these attempts is fundamentally wrong, because every one of them is trying to optimize the existing process. They're adding layers on top of a system that was broken to begin with.

The Stagnation Nobody Wants to Talk About

Here is the real indictment. Manufacturing productivity has risen roughly 90% over the last twenty years. The global economy has grown by roughly 50% in the same period. Construction productivity has grown by about 10%.

Ten percent. Over twenty years. That's roughly half a percent per year.

Every other industry moved forward. Construction stood still. And while it stood still, we've been telling ourselves a story that the solution is always just around the corner. A better piece of software. A smarter algorithm. A clever new printer. None of it has moved the needle, because none of it is operating at the right level of the stack.

First Principles

How do we actually fix this? By doing what every other industry eventually had to do. Throw out the inherited process and rebuild the thinking from scratch.

David Senra devoted an episode of his Founders podcast to breaking down Atoms Are Cheap, Process Is Pricey by Max Olson. One line cuts directly to why construction is stuck:

"If materials are cheap and the real cost is all process and overhead, you need to control the process yourself."

A 2x4 cost is fixed. Concrete cost is fixed. Copper pipe cost is fixed. What makes a house expensive isn't the materials. It's the fragmented process of coordinating dozens of independent trades, each bringing its own raw materials, each billing its own margin, each solving its own subset of the problem without any unified system holding it together.

Here's what first-principles thinking produced at SpaceX: The Falcon 9 runs on nine identical engines. Not nine custom engines optimized for their position in the stack. Nine identical ones. The Falcon Heavy is three Falcon 9s strapped together. That is not a metaphor. It is literally three of the same rocket, bolted together, doing a heavier job.

SpaceX didn't invent new physics. They collapsed the number of variations. They asked a simple question: Does this problem actually require a custom solution, or are we just in the habit of building one? Nine times out of ten, the custom part was a habit, not a law.

The Tic-Tac-Toe Principle

There's a concept that captures this better than any engineering anecdote.

Tic-tac-toe looks like it should have an intractable number of possible games. Count every possible move sequence, every X and O placed in every possible order, and the board allows for 362,880 distinct games.

Musician and coder Marc Evanstein, in his video There Are Exactly 14 Different Games of Tic-Tac-Toe, argued something different. He didn't just strip out duplicate games created by rotation and reflection. He stripped out the games that only exist because a player made a move no rational person would ever make. Ignoring an open winning move. Failing to block an obvious threat. Playing into a fork anyone could see coming. Once you eliminate the duplicates and the irrational play, the number of genuinely distinct games collapses to 14.

Not 14 thousand. 14.

That's the part that matters for construction. Evanstein isn't saying the possibility space is smaller because it's mostly repetition. He's saying most of what makes the space look big is bad play nobody should actually choose. The space only looks massive because the rules of tic-tac-toe permit moves no educated player would ever make. Strip those out, and what's left is a solved game.

This is exactly what's happening in construction.

I'm not saying every building is a duplicate of every other building. I'm saying that if you eliminate everything that exists because of bad design in functionality, the possibility space collapses the same way. Most of what we call "every project is different" is not meaningful variation. It's the construction-industry equivalent of ignoring an open winning move.

A crew frames walls in driving rain because that's how it was scheduled. A plumber and an electrician arrive the same day and work around each other because nobody sequenced them. A change order cascades through six subcontractors because one detail wasn't caught in review. Custom millwork gets an 8-week lead time because every project is treated as a one-off. None of these are forced by physics. They happen because the industry's rules permit them, not because anyone would design things this way from scratch.

That's the construction industry playing 362,880 tic-tac-toe games when there were only ever 14 worth playing.

Two Parts

So what does first-principles construction actually look like?

Strip everything away and ask: what are the actual parts of a building?

The answer is shockingly small. At the most fundamental level, a building has two things: walls and floors. Everything else is a subsystem that runs through, on, or between them. Plumbing, electrical, HVAC, finishes, openings.

The future of construction is inventing the process that prints or molds these components at scale, as unifiable modular parts. Not assemblies of hundreds of smaller pieces coordinated by a dozen trades on a muddy site. A wall today is wood studs, plates, sheathing, insulation, drywall, wiring, plumbing, and finish. Hundreds of pieces touched by five or six different crews.

The wall of the future is one part.

A customer chooses from the catalog. Architects design to the catalog. Engineers move from site supervision to factory preconstruction. The parts are manufactured the way every other serious product is manufactured: indoors, at scale, by machines that improve with every iteration. The site becomes assembly, not fabrication.

The Hardest Part: Letting Go of Difference

Here's the part most people don't want to accept. In this future, buildings will look almost the same.

Homes will look like other homes. Offices will look like other offices. The endless visual variation we associate with a neighborhood will largely disappear, replaced by a small set of standardized forms with minor configuration differences.

This sounds dystopian. It's not. It's exactly what every other mature industry went through.

In 1920, every car was essentially bespoke. Today a Civic, a Corolla, and an Elantra are nearly indistinguishable from twenty feet away, and nobody considers this a tragedy. The variation moved from the physical form to the configuration layer: color, options, software. Cars, phones, computers, and televisions all followed the same path. The core product converged, and the world kept turning.

Buildings will follow the same path. They have to. The economics of a parts catalog require it.

What replaces visual variation at the form level is variation at the skin level. The facade becomes the canvas. Imagine the exterior of a building as a programmable screen panel, an e-ink surface, or a modular facade that can be reconfigured for branding, climate, or mood. The building underneath is the same building everyone else has. The face it presents to the street is yours.

The attachment to the idea that my building must be fundamentally different from yours is the biggest cultural obstacle to construction's first-principles moment. It's also the least defensible one. The 1950s row house on your block is not a unique creative work. It's one of ten thousand nearly identical houses built from the same pattern book, and the people who live in it love it anyway.

The future won't look like a science fiction movie. It will look like a normal street where the houses are, honestly, a little bit boring. Until you notice they cost a third of what they used to, went up in a fraction of the time, and the family living inside can afford one.

The Inevitable

I am not pitching an idea with all of the answers. If I was, I would be pitching this to a VC, not writing an article on it. I am, however, telling the story of what the future must look like.

Every other industry eventually got its moment. Cars got Ford. Computing got Apple. Rockets got SpaceX. Construction is the last trillion-dollar industry still waiting for the entrant who decides to stop optimizing the old process and start rebuilding from the ground up.

Someone is going to do it.

And when they do, everyone who spent the 2020s selling AI plan-readers and better project management dashboards will look like they were rearranging furniture in a house that was already being demolished.