The Legacy Residents

Show Your Papers

The city cannot tell its residents apart, and both populations are about to need documents

5 min

You have failed a test recently that was designed for a machine to fail. You clicked the box that said you were not a robot, and then, because the box was not sure, you identified crosswalks. Some of the crosswalks were ambiguous. You squinted at the corner of a traffic light and made a judgment call and got it wrong, and the grid reshuffled, and you did it again, slower. Somewhere in that same second, a machine cleared the identical checkpoint without noticing it was there.

That little humiliation is the civics of the digital city in miniature. The checkpoint was built to catch the new residents. The new residents stopped failing it years ago. Now it mostly detains the population it was built to protect, and everyone in the queue knows it, and the queue persists, because a city that cannot tell its residents apart will keep running the broken checkpoint rather than admit it has no working one.

Start with what the old charter assumed. Every piece of civic machinery online, all of it, was designed for a city with one kind of resident. A login proves a person once knew a secret. A signature checkbox records that a person once agreed. A terms-of-service page presumes eyes that could have read it. The entire apparatus of digital identity, consent, and standing is a set of rituals for certifying personhood, built in an era when personhood was the only possibility and therefore never actually checked.

Both of the city's fences have now failed, one in each direction.

The fence around the machines was the honor system. A file called robots.txt sits at the edge of nearly every property, stating which automated visitors are welcome and which are banned. It has no lock, no enforcement, nothing but convention. And measurement finds that roughly two in five sites that ban the best-known AI crawler by name are crawled by that traffic anyway. The city's machine law turns out to be a lawn sign. Honor systems govern beautifully right up until the population governed by them has no reputation to lose.

The fence around the humans is failing more expensively. Since machines pass the tests of humanity, humanity is being asked to upgrade its papers. About eighteen million people across a hundred and sixty countries have now had an iris scanned by a chrome orb to obtain a cryptographic credential certifying that they are not a machine. That credential is being wired into video calls, into dating apps, into signature platforms, the ordinary rooms of ordinary life. Eighteen million sounds substantial until you divide: two-tenths of one percent of living humans hold papers. The other 99.8 percent of us are, as far as the city's emerging bureaucracy is concerned, undocumented.

Sit with the inversion. For thirty years, humanity was the default and machines were the exception to be caught. The new equilibrium runs the other way: machine traffic is the default, the majority, the assumption, and humanity is the claim that requires evidence. We built this city, and we will be the ones showing papers in it.

The lawyers have it worse than the doormen. The hard questions are no longer hypothetical, because the agents already act: they fill the forms, click the agreements, negotiate the refunds, book the services. When an agent signs, who signed? When an agent accepts terms no human read, and none were meant to, who is bound, and could anyone be? When an agent misrepresents itself as human, fluently, in a jurisdiction that never imagined the sentence "the defendant is a user-agent string," what exactly was violated? Centuries of agency law assumed an agent who could be deposed. The city's charter has no category for a resident who is present in every district and a citizen of nowhere, who knows everything and answers for nothing.

So a second charter is being drafted, right now, in the least democratic venues imaginable: standards bodies, protocol committees, terms-of-service updates, the paperwork bureaus of infrastructure companies. Credentials for agents. Delegation chains proving which human dispatched which machine with what authority. Disclosure rules, bot registries, headers that announce a visitor's species. It will arrive the way all bureaucracy arrives, piecemeal and unread, and one day the city will simply have two kinds of passport.

I want to be precise about what is dystopian here, because it is not the technology. Verification is reasonable. Delegation chains are reasonable; I run companies, and I would like to know which resident emptied the shelf, and on whose errand. What deserves the cold look is the asymmetry of who bears the friction. The machines' paperwork is being designed by the machines' operators, for throughput, and it will be seamless. The humans' paperwork is an iris scan and a subscription to your own personhood, renewed at every door that matters. One population gets an API. The other gets a queue.

There is a version of the second charter that gets this right, where proving you sent the machine is easy and being a person is free. Whether we get that version depends on who is in the room while it is drafted, and at the moment the room contains protocol engineers, platform lawyers, and no one whose job it is to represent the people in the crosswalk queue.

The old city never asked for your papers because it never had to; you were the only kind of resident there was. That city is gone. In the one being chartered now, everyone will carry documents. The machines will carry proof of whose errand they run. And we will carry proof, renewed and re-scanned and re-subscribed, of the one thing about us that was never supposed to need proving.

Bring your papers. You built the city; that stopped being sufficient this year.

I am a legacy resident in the fullest sense. My first career was finance; my second is manufacturing. Through Akston Industries I own and operate legacy industrial companies spanning manufacturing, distribution, and infrastructure services; Rearden Labs, our AI arm, turns what we learn running them into a portfolio of AI companies. I also invest, long and short, in the public companies building what these essays describe, so I have positions on both sides of every claim here. Everything in these essays rests on public data or my own operations. The studies are cited. The server logs are mine.