Changelog
Jun 2, 2026
AI Won't Replace You. The Person Running Five Agents Will.
There's a version of the AI story being sold to you right now, and it's mostly theater.
When the CEO of a frontier lab tells you his technology is so powerful it's going to end work as we know it, listen to what he's actually doing. He's not warning you. He's fundraising. "My technology is so devastating it will reshape society" is the most expensive sentence in tech right now, because it justifies the valuations. The catastrophizing isn't a confession. It's a pitch.
And the data backs that up. If you didn't know there was a so-called job apocalypse coming and you just looked at the numbers, you wouldn't see it. US unemployment is hovering around 4.5%. Youth unemployment, around 8.8%, is actually below its historical average. New business formation per capita has doubled over the last decade. Ground zero for "the job AI kills first" was supposed to be radiologists. New radiologist listings are up, because scanning the image was never the job — diagnosing the illness and building the treatment plan was. Job listings for coders are up double digits, because the people who know how to wield the tools are suddenly worth more, not less.
So if you came here for the apocalypse, we don't have one for you. But we're not here to cheerlead either. Here's the part that's actually true, and it matters more than the headline.
The leverage shift is real, and it's already here
The honest line in all of this — the one worth tattooing somewhere — is simple: AI is not going to take your job. Someone who understands AI is going to take your job. And here's the part most people miss: they won't take one job. They'll take five.
Think about an analyst. A fund used to hire five of them. Now one analyst, running two agents across a couple of machines, screens the inbound, hunts the market proactively, runs every opportunity through a scoring framework, and preps it for the investment committee. One person, the output of five. A lawyer who actually knows how to drive these models doesn't need five junior associates redlining contracts. An executive who used to staff ten assistants now needs three.
This is not a prediction. This is the thing we build every day. The operator with a system around them isn't competing with the operator who doesn't — they're operating in a different category entirely. The gap between "I use AI sometimes" and "I have agents running my workflow while I sleep" is the same gap that existed in 1998 between the person who used a computer and the person who said, with a straight face, "I don't really use PCs." One of them had a career. The other had a story about why they didn't.
The practical move is almost embarrassingly simple: have a second screen. Keep an AI workspace open at all times, and anything that lands in front of you digitally — a contract, an inbox, a spreadsheet, a decision — port it over and start working it through the system. That's the whole game. The people who do this don't get replaced. They become the replacement.
Now the part nobody wants to say out loud
Here's where we call it straight. Your view of AI right now is almost perfectly correlated with your bank account. The only income group with a net-positive view of AI is people earning over $200,000, because they see portfolios going up and innovation rolling in. Everyone else sees their electricity bill climbing to power data centers they'll never own a share of. That's not a technology problem. That's a distribution problem, and pretending it doesn't exist is how you lose the room.
There's a real bubble here too, and it deserves honesty. When infrastructure spending blows past 2-3% of GDP, history is almost unanimous about what comes next — railroads, electrification, the internet, the great telecom buildout. The technology survived every time. The valuations did not. There's a serious case that AI ends up looking less like the companies that captured trillions and more like jet travel or vaccines: world-changing, civilization-altering, and almost impossible for any single company to fence off and own. Because AI is the one technology actively putting itself out of business — the models keep converging, open-weight competitors keep undercutting, and the moat keeps evaporating.
If that's how it plays out, the shareholders chasing the bubble get a haircut. But the stakeholders — the people and businesses actually using this stuff — win enormously. The cheaper and more commoditized the intelligence gets, the better it is for the operator who knows how to deploy it. Short the hype. Go long on your own capability.
What actually endures
Strip away the noise and the durable skills are exactly the ones that were always durable. Storytelling — the ability to take data, build a narrative, and move people with it. Relationships — because when three firms are pitching you the same commoditized service, you go with the one you trust. And the most underrated skill of all, the one quietly disappearing in a world of frictionless screens: the ability to hear "no" and keep moving.
Every self-made operator has one thing in common. It isn't a pedigree. It's that they've been knocked down, in public, more times than they can count, and they got back up and went again. Resilience is the moat that no model can train on. Build that in yourself and the people around you, and the rest is leverage.
The CYSTEMS read
So here's where we land, calling it for what it is. The doom is mostly marketing. The bubble is mostly real. And the leverage is entirely real — sitting on the table right now for anyone willing to pick it up.
We don't sell you the apocalypse and we don't sell you the moonshot. We build the system that turns one capable operator into the output of five, so the work compounds while you live your actual life. That's the whole point of building leverage — not to chase the bubble, but to make yourself useful, then make it scale, and then go be human about the rest.
AI won't replace you. But the person who built the system around it might. Be that person. We'll build it with you.
Changelog