I need to confess something. I’m coming clean about what my company is and why it’s not good for you.
Valni is one person, me, running what used to take companies of thousands. Valni’s Switchboard routes AI requests, meters usage, handles billing, manages keys, enforces spend controls, and reports analytics. Any one of those verticals is a whole company somewhere else. At least one is a billion-dollar company that does less than Switchboard. I built it alone, in months, and I’ll likely need very little investment. I will never need more than a handful of people. Operating and development costs are so low that profit arrives early, removing the need for continual investment. On the current trajectory, every dollar Valni makes flows to one place, forever.
Mine.
I’m not telling you this to brag. I’m telling you because I am the thing coming for you. I got here faster than anyone expected, and there are ten thousand people who will do exactly what I did inside the next two years. Most of them won’t stop to write this essay first.
This is the story of what’s coming, why it’s different from every concentration before it, and the one structural move that still works before the window closes. That move has a name, usership. Like ownership, but earned by use instead of capital. By the time you finish reading, you’ll understand why it’s the only word missing from the language we use to talk about power.
I’ll tell you what I considered first.
I debated whether to fake a headcount. I thought about raising money to impress my business school class of 2016. Both would have worked, but they would have been unnecessary, and potentially risked my ultimate success.
After I built Switchboard and understood what was coming, I felt obligated. I tried to help my employer see it, educate my coworkers, warn the people around me. Every effort landed the same way. Skepticism, engineering cliques, the comfortable certainty that AI can’t do what I’m good at. I couldn’t make them see it. You can’t explain a realization. People have to arrive at it themselves.
After realizing I was in a strange position where failure was less likely than success, I wanted to find a co-founder, someone to share in the success. Maybe previous co-workers I’d liked, or important people as advisors. So I reached out to friends with a motive I’d never had before. Not for help. Not for money. As a favor to them, so they could come along. I wanted to be able to say I gave them a chance, kept them in the loop, so when it works they don’t blame me for not giving them a chance. But everyone still felt like it was for the opposite reason. I couldn’t show them what I see. No words would make them understand.
Then I did the math. Adding people costs more energy than doing it myself. I was already faster alone.
The gradual theft is over. The instant one begins.
You already know the system is tilted. You feel it in the rent, in the math that doesn’t close, in the raise that didn’t come. You know the last fifty years moved money steadily from doing to owning. You don’t need me to prove that.
Here’s what you don’t know yet.
That was the gradual version. What’s coming is the instant version, and it arrives around 2027, not with a crash, not with a headline, but quietly, the way all the real ones do. One morning it just stops being deniable that one person plus AI does what used to take a thousand. No drama. The drama shows up later, in the charts. And by then the window is closed.
Your ability to participate in success is no longer guaranteed
There have always been two ways ordinary people got a stake in the economy. You got hired, and wages bought you in slowly. Or the company went public, and you bought a piece.
Both doors are closing at the same time.
The hiring door. Valni will never need more than a handful of people. The companies being built at agent scale hire in single digits, not thousands, not because they’re cruel, but because they don’t need more people. A handful of seats is not a door. The work still happens. The money just stops moving through people on its way somewhere.
The investment door. Valni will likely not raise past a Series A. Companies built at agent scale are so cheap to build and run that profitability comes early, and early profit removes the need for continual investment. Capital becomes a choice, never a lifeline. In 1956 a factory worker could buy Ford stock the week Ford went public. That route (effort converts to wages, wages convert to savings, savings buy a stake) is the entire history of how regular people got into the money. It’s closing. And when it closes it doesn’t reopen.
This concentration, unlike every previous one, is permanent
AI compounds. The returns to owning it grow every year. Human effort decays toward zero as a wage. When effort no longer converts to ownership, and we are close, there is no mechanism left to buy your way back in. Every previous concentration could be unwound by the next generation’s effort. The robber barons got broken up. The monopolies got regulated. The next generation could always work or invest their way back in. Not this time. When AI removes the hiring door and the investment door simultaneously, there is no ladder left.
Call it The Severance. Not a metaphor, the literal event. Every human gets laid off from the economy and handed a payout. They’ll call it generosity. It’s severance. The acknowledgment that you’re no longer needed, dressed up as a gift. And like every severance, it happens once, it closes the door behind you, and the company moves on without you.
Whoever holds the compounding assets at the moment of The Severance holds them permanently. Not by law. By arithmetic.
That’s what’s different. Not that the system is unfair; it’s always been unfair. That this time the unfairness doesn’t reverse.
Your attention is the only thing they still need from you
The specific, economic power of leaving one product for a better one. That power survives even when everything else is worthless, because no AI can want the product for you. Usage is the one human contribution that cannot be automated away; it is the entire point of the machine. You can’t automate demand. The user’s seat is the only seat in the AI economy that permanently requires a human.
You are owed more than a severance. Usership is coming.
Every previous economy paid people through work. The next one has no payroll. But it will always need users, because users are the point. The entire machine exists to serve them. That’s not a sentiment. It’s a structural fact, and it’s the one lever that survives the transition.
Usership is the word for what comes next, and the concept is simple. Companies win users by handing value back to them. It already comes in many forms. Venture-backed companies undercut incumbents with subsidized prices, investor money flowing to users. Passthrough pricing hands users the wholesale rate. At the far end, companies let users take the profits themselves. The forms will keep multiplying, but the direction is the same in every one. The value users create flows back to the users who create it, earned by use instead of capital.
Here is the prediction. AI is collapsing what software costs to build and run, so prices race toward free, and price stops winning users. When everything costs nothing, the only move left is to hand users more than free, a piece of the upside. Competition will drag every company along that spectrum whether its founders like it or not. The ones that move early will take the users.
Part of it is already live, in how Switchboard charges. No subscription. No fixed fee. You build, you get users, and only then do I get paid, a sliver of the usage that flows, months behind your success and only in proportion to it. I carry the provider risk, the spend controls, the accounting. That is not a payout program; it is the thing that has to come first, a company that cannot make money unless the people building on it do. Alignment is the foundation, and usership is what gets built on top.
I predict usership is the new method of economic participation, one that has the potential to rebalance some of the structural issues of the economy. The hiring door is closing. Companies don’t need your labor. The investment door is closing, for software. Companies don’t need your capital. But they will always need you to use it. Every AI product, every platform, every service built on these rails exists because people choose to use it. That choice has always been the only real power users held. Usership is what happens when that power finally counts.
One company changes nothing. The standard changes everything.
Valni is one company. One company changes nothing at scale. A user preference changes everything. When users see a company offering usership, they prefer it. Not out of ideology; out of self-interest. Why use a product that keeps everything when you can use one that gives something back?
That preference is the pressure. Companies that don’t offer usership lose users to ones that do. So they adopt it, not because they were asked nicely, but because the alternative is watching their users walk.
No government has a plan for the world after The Severance. No regulator is ahead of it. The gap between what AI will do to economic participation and what policy is doing about it is the largest unaddressed risk of our time, and the failure mode isn’t just inequality. It’s what inequality becomes when enough people have nothing left to lose. Civil unrest. Government overreach.
Usership is the reverse Uno. The move nobody saw coming, because everyone was looking at voting, lobbyists, legislation, and capital to solve it. They missed the one thing that never stopped mattering. Us. Users. The only input AI can’t replace, the only role that survives The Severance, the only seat at the table that belongs to everyone.
Usership is the answer. A gradual transfer of wealth back that doesn’t require companies or governments to act against their own self-interest. It doesn’t require war. Just users choosing. And capitalism, a system that won’t survive The Severance, can help us transition through it.