Did running your business actually change?

Strip away the keynote slides and the funding-round theatre, and ask the plain version. Twelve months of AI everywhere. Is your company a different company now, or is it the same company with a few processes that used to take three people and now take one?

For most firms it's the second one. That's fine. That's normal. It's what every technology wave has ever done.

I've watched a few of these now. Twenty-three years, give or take, across SAP, Oracle, and pricing software. I sat through the ERP rollouts that were going to transform the enterprise. I sold software that was going to make pricing decisions so people wouldn't have to. Then came RPA, the bots that were going to empty the back office. Every wave arrived saying the same thing. This changes everything. And every time, it automated a few tasks and moved some work around. The business came out looking a lot like the business.

So yes, I think a good part of the current AI valuation is a bubble, and parts of it will burst. Not because the technology is fake. The technology is real and some of it is genuinely good. The bubble is the gap between what gets promised on stage and what shows up on the P&L. That gap always closes, and it closes downward.

The doom headlines get the next part wrong, though.

The fear underneath all of it is jobs. The machines are coming for the work. We've run this experiment before, for two centuries, and the result is boring and consistent. Technology kills specific jobs and creates more than it kills. In 1900, around half of Americans worked in agriculture. Today it's under two percent. Those people didn't stay unemployed. They became something else. More than half the jobs people do today didn't exist eighty years ago. When McKinsey looked at fifteen years of the internet in France, it found half a million jobs destroyed and 1.2 million created. The net is positive. It almost always is.

What changes is the kind of work. The jobs automated first are the ones nobody dreams about as a career. Who grew up wanting to spend forty years closing support tickets? Reading the same script to the same angry caller? Keying invoices from one screen into another? That work goes first. Good. It frees people for the work that's harder to automate and a lot more human. Judgment. Relationships. The messy stuff a model can't sit in a room and do.

I won't pretend the shift is painless or even, because it isn't. The same research that shows net job creation also shows the new jobs lean toward higher skills, and the people displaced are rarely the people hired into the new roles. A factory packer does not become a data analyst on Tuesday. That gap is real. It's a training and policy problem, and waving it off with "innovation always creates jobs" is its own kind of dishonesty. Both things are true at the same time. The economy nets out positive over a decade, and real people pay the cost of the transition in real time. You don't get to pick one and call it honesty.

Ignore the noise about transformation. Ask the boring question instead. What number in this business would you actually move, and would the thing you build still be running in two years, once the hype has moved on? The AI that matters reaches production, you own it, it moves a real number, and it survives contact with how your people actually work. The demo was never the point.

That part has nothing to do with the bubble. It'll still be standing after the bubble is gone.

The call center job won't be. Ask anyone who had it whether they'll miss it.