What AI did to growth.

I watched something get made in about eleven seconds that I used to charge four thousand dollars for. It wasn't as good as mine. It was good enough, and good enough turned out to be the whole story. What people say is that AI makes everyone more productive, and more productive means more growth. I think that misses what actually changed. What it did was remove the cost of producing average work.

Which isn't the same as making everyone better. It's more specific and more brutal than that.

Think about what used to be worth money. A competent brochure. A serviceable ad. A reasonable read on a market. A decent first draft of a contract. A summary of a long document nobody wanted to read. A set of interview questions. A content calendar for the quarter. None of it exceptional. All of it fine, and fine had a price, because producing fine took a trained person several hours and there were only so many trained people.

Fine now costs nothing and arrives in eleven seconds.

So the floor came up. Everything that used to sit slightly above the floor and charge for the gap is now level with it, and the people who made a living in that gap are the ones feeling it, whether or not they can name why.

For an operator the practical question is which parts of your business were living in that gap.

Some obvious ones. Anything you charged for because it took effort rather than because it took you. Reports. Summaries. The first draft of anything. Work whose value came from the labour rather than the judgment inside it.

And some less obvious. Being informed used to be a position. Knowing the benchmark, knowing what the industry does, knowing the framework. All of that's now free to anyone who asks a decent question, which means it's no longer a reason to hire you.

The honest counter is that we have been here before and it turned out fine.

Spreadsheets did this to bookkeeping and there are more accountants now than there were. Photography did it to portrait painting and photography turned out to be a bigger industry than portraiture ever was. Every time the floor rises, work moves up rather than disappearing, and the people forecasting collapse have a poor record.

I find that genuinely reassuring about the long run and not very useful about the next three years. The move up isn't automatic and it isn't comfortable, and plenty of individuals don't make it even when the category does.

What I keep landing on is that the things which held their value are the ones with a person attached who can be held responsible.

Presence. Somebody physically there. Judgment where being wrong costs the person making the call something real. Accountability, which isn't a service but a relationship. A model will give you an answer and it will never lie awake about it, and that turns out to be most of what people were buying.

I hold this loosely, because it's exactly the sort of thing a person in my position would want to be true, and I've noticed that every trade whose value collapsed said something similar on the way down.