The Blast Door.

In Star Wars, the doors don’t close the way our doors close. On the Millennium Falcon, on the Death Star, on every ship in that universe, a blast door doesn’t slide down like a window, where you can watch the gap shrink and judge your time by it. A blast door closes from every direction at once. The opening is a circle, the circle gets smaller, and the thing about a shrinking circle is that the last part goes fast. One second there’s a hole you could step through. Half a second later there’s nothing, and it’s sealed like it was never open at all.

That’s the shape of the opportunity I want to tell you about. Not a window. A door, closing from all sides.

The chainsaw and the hacksaw

I jumped in early in 2023, back when all of this was chatbots and none of it could act on its own yet, and I made a bet that turned out to be half right. I saw immediately that this was an upskill everyone was going to need. So I built AI training and braced for the rush, because surely people would look at what was coming and say: I’d better learn this. I’d better be first. I’d better get really good at it.

The complete opposite happened. People rebelled. They’d heard AI makes mistakes. They’d heard AI was evil, that the robots were going to take over the world. I felt like I was handing people a chainsaw and watching them nod politely, walk back to the limb pile, and pick up their hacksaw. Why people don’t gravitate toward the power of a technology, regardless of the application, confused me then. It still does.

The argument, stated plainly

Maybe less than three years. That number is my belief, not a measurement, and I’ll hold it loosely. The direction is what I’m sure of.

Walk it backward from where this is going. Today, AI is a tool that solves problems a human brings to it. The human supplies the problem, the context, the taste, the decision to act. That’s the partnership, and it’s the whole basis of my consulting practice and every business I run. But the labs aren’t building toward better tools. They’re building toward systems that spot the problems too. When that arrives, whatever we end up calling it, the economics are brutal and simple: a problem gets identified, solved, and deployed for roughly the cost of the electricity it took to think about it.

Now put yourself on the wrong side of that. Say you had an idea. A garage project, a side hustle, a small business built on solving one real problem for real customers. The day AI can originate and solve that problem for the cost of energy, your business doesn’t get out-competed. It gets out-existed. Nobody beat you. The problem you were paid to solve simply stopped being worth money.

That could eventually happen to almost any business, and what happens after that day is a different essay, one I’m not writing today. This essay is about the time before it. Because between now and then, there is a stretch of time when the partnership still wins. When intelligence is nearly free but direction is still scarce. When a human who knows where it hurts, plus a machine that can build, can create things that neither could alone.

That stretch is the opening in the blast door, and it is closing from every direction at once.

Bets placed inside the opening

I’m not writing this from the sidelines. I run four businesses by myself from a home office in McKinney, Texas, and every one of them is a bet placed inside that opening. A home-repair-by-text service, because homes won’t stop breaking and people will always want the easiest trusted way to get them fixed. An AI hiring engine, because companies will always need to know who they’re actually hiring. A lead-response system for small businesses, because a missed call will always be a missed customer. And the consulting practice, which exists to pull other operators through the same door I went through.

Notice the pattern in those bets: each one sits on something about human behavior that does not change. That’s deliberate, and it’s the filter I’d hand any builder: build on what people will still need and still do, no matter what the machines learn. It’s also the honest hedge, because a business built on durable human behavior has the best odds of mattering on both sides of the door.

Here’s what I’m actually racing for. My family needs me as a provider, and they’re going to need providing for whether I’m here or not. If I build something that outlasts me and gives them stability, that’s the race. Money matters, but Ray Dalio draws a distinction I think about a lot: wealth is one thing, money is another. You cannot spend wealth. You have to sell it to get money, because you can only spend money. Recurring income is the engine that makes money keep arriving. That’s what I’m building, and I’ve taken on real personal risk to build it, because the math of the closing door changes the shape of the work. This doesn’t need my fevered pace for the next ten years. It needs it for the next two, maybe three. Then the benefits play out over the next ten, fifteen, twenty, and I plan to be around for thirty more to watch. Work the burst, then automate myself out of my own businesses. And the beauty of the model is that you don’t set it and forget it. You set it and monitor it. It’s a maintenance program. The hard work is up front.

Get your board in the water

So what should a person actually do with this? I keep coming back to something Eric Weinstein told Steven Bartlett on Diary of a CEO in the summer of 2025. Asked what he’d say to the average person about what’s coming, he said: “Get your board in the water and prepare to paddle. Like all get out. The tsunami of a lifetime is coming.” And on what to become, his advice kept landing on the same two words: “Get flexible, get good. Get good on a bunch of different stuff.”

That’s not a throwaway line. It’s the correct strategy for a closing door, and here’s why. You cannot know in advance exactly where the opening will be narrowest or longest. What you can control is whether you’re in the water when the set comes. Getting good, at AI, at your own domain, at the unglamorous mechanics of shipping something real, is the paddling. Waiting to see how it all shakes out is standing on the beach.

And if you already have a job, I’ve written about your version of this. The consequence clock says you work at the speed you are checked, not the speed you are capable of. If you’ve crossed the competency line with AI and your company still measures you monthly, you are a second-hand person in an hour-hand building, and your finished work is sitting scheduled for delivery while you do what, exactly? That freed time is the most valuable asset you will ever hold, and the blast door is the deadline on it. Use it to build something you own. Not a hobby. An income source with your name on the deed, however small it starts.

The part that will cost me some friends in the feed

There’s a story going around, and I’ve heard it told on television by people I otherwise respect, that the money side of this is easy now. That the AI labs and the funds around them are so hungry for ideas that a person can more or less walk in the door with a plan and walk out with a check. I’ve heard it put almost exactly that way on the podcast circuit, what I’ve started calling the tour of talkers: the same guests rotating through the same shows, agreeing with each other that opportunity has never been more available to the little guy.

I call bullshit, and I’m allowed to, because I tested it.

I have real businesses with real receipts. Working products, live customers, revenue that arrived this week, business plans built to be read by exactly these people. I’ve started submitting to the doors that actually have addresses, the angel networks and the funds, and I’ll keep submitting. You know what earning the investment community’s attention has gotten me so far? Silence. Not rejection. Rejection would at least mean somebody read it.

I’m not above asking. That’s why I keep walking into those rooms, and my own operating system is what finds the doors worth knocking on. But let me be precise about what a check would even mean, because this is where the circuit gets it backwards. If a fund wrote one tomorrow for my home-repair business, I wouldn’t call that success. I’d call it a rounding error against what that business is actually building toward. Here’s my real benchmark, and I’ll put it in print: if you live in one of our markets, within two years it will be inconceivable to handle a home repair any other way. A local contractor, on your schedule, arranged entirely by text, paid from your phone. The friction is gone. It is simply too easy. That is the valuation. A check is just fuel.

I’m not bitter about the silence. I’m reporting it, because somebody has to say the quiet part to the people at home: if your plan requires a stranger’s check to start, your plan is behind the door already. The money finds you after you no longer need it. That was true in every era, and no amount of AI hype has repealed it.

Two tells, from a guy who has seen this movie

First tell: the suffix. In 1998 a company was worth whatever its profits said it was worth. In 1999 the same company added dot-com to its name and was suddenly worth fifty times that, with no new profit, no new margin, no new anything. We know how that ended. The same suffix magic is happening right now with dot-AI, and yes, I’m poking myself, because my own shingle hangs at kerzie dot AI. The difference I’d offer is that I’m not asking you to value the name. I’m showing receipts. That’s the line that separates a business from a costume: revenue you can point at versus a suffix you rented.

Second tell: monthly income. When someone tells you they make $500,000 a month, ask yourself why they didn’t say $6 million a year. A year is more impressive. A year is also harder to fake, because a year means you survived twelve months of reality. I learned this pattern in the early 1990s in Amway, and I’m not ragging on Amway, I learned real things there about distribution and hustle. But the income was always quoted monthly, because a good month can be manufactured and a good year has to be earned. Thirty years later the manufactured month is back, wearing an AI t-shirt, screenshotted in your feed between the four-prompts-to-riches posts. Same play. Same tell.

None of that circus changes the real argument. It just means you should run your race without waiting for their permission or their check.

Two hours and forty minutes

Let me tell you about the Titanic, because I can’t stop thinking about the timeline.

From the moment it struck the iceberg to the moment it went under was roughly two hours and forty minutes. For most of the first two hours, the ship was calm. The band played. Drinks were poured. The first lifeboats were lowered half empty, seats rowing away unfilled, because standing on a warm, lit deck, the urgency wasn’t believable yet. The mathematics of the situation had already been decided, the ship was already sinking, but the experience of the situation said otherwise. Then came the last thirty minutes, when the deck tilted enough that everyone believed at once, and suddenly there were not enough boats, not enough seats, not enough time. The people who perished didn’t lack information. The ship told them what was happening two hours before the water did. What they lacked was the willingness to act while acting was still cheap.

The door doesn’t announce the last half second. It just closes.

While the deck still feels level

That’s the whole essay, really. Not that the world ends. It doesn’t. Something comes after, and thinking clearly about that is work I’ll do in another piece. This one exists to say the simpler, harder thing: the era when an ordinary person plus a nearly-free intelligence can build their own lifeboat is running right now, it is the best innovation opportunity any of us will ever see, and it is closing from every direction at once, at a speed that will look obvious only in the rearview.

Get across the competency line. Get good, at more than one thing. Get your board in the water. Build something you own, however small, on top of something people will always need. Do it while your work is scheduled for Friday and your evenings are quiet and the deck still feels level.

And here’s where I plan to be standing when the door closes. The measure of success I actually want is not a number. It’s a community. People who read these essays, took me up on the setup or the session or the full build, and got through the door while it was open. People who someday say thank you. Thank you for pointing it out. Thank you for making it simple. Thank you for taking the scary out of it. Thank you that the old projects got dusted off, the business grew, and the family is better for it.

I want a community of winners. That’s what I want. The door is still open. Get your board in the water.