Playing with AI—have you actually made money? Three paths, and the precondition most people skip
An uncomfortable topic: after all this time playing with AI, have you actually made any money?
I break it into three paths—cut cost, copy, take it apart and rebuild. Each one is harder than the last, and each one has a higher ceiling. But what decides the outcome isn't which path you pick. It's the precondition under all three: you have to be making money already. AI is an amplifier, not an engine.
The short version
There are really only three ways to make money with AI: cut cost, copy, and take the business apart and rebuild it.
Each path is harder than the one before, and each has a higher ceiling. But more important than which one you pick is the precondition sitting under all three—
You have to be making money already.
AI is an amplifier, not an engine. It can make something that already earns faster, bigger, and more automatic, but it does not conjure a business out of nothing.
Path one: cut cost
The easiest one to see. Work that used to take ten days now takes one.
For an individual, the time you free up can be reinvested: learn something, deliver more value, turn it into a promotion and a raise. Even the spare hours you save can go into something on the side.
For a company, the drop in headcount cost is real. Across the clients I work with, whenever AI replaces a business process, it takes out roughly 60% of the headcount cost.
But there’s a trap here. The moment that cost drops, a company’s first instinct is usually “can I run with fewer people.”
Going down the layoff road is a vicious cycle. The savings never turn into new capacity—you just made the operation smaller. Next round of cost-cutting, the only lever left is another cut. The end of that road is shrinking yourself out of existence.
Path two: copy
One level up from cutting cost: take the thing, the process, that already monetizes repeatably and turn it into an automated product—or serve and follow up with more clients on the same headcount.
This path is hard in two places, and both are real:
Most companies right now are short on orders, not efficiency. With no demand, extra efficiency has nowhere to go.
Everything looks the same. If you’re a one-person business, you’ll find that once AI is behind everyone, 99% of what people deliver isn’t fundamentally different. It turns into a bidding war, and very few people genuinely command a premium for the level of what they deliver.
So for a company, the more valuable use on this path isn’t on the production side at all. It’s upstream:
- expanding the lead pool
- lifting the conversion rate
- following up, at a decent standard, on all the leads nobody was ever going to touch
That last one gets underrated. Those leads went untouched because they couldn’t be afforded—not enough hands, and the expected return per lead didn’t cover the labor. AI changes that premise.
Path three: take it apart and rebuild
The hardest path, and the one with the highest ceiling: use AI to pull the whole business apart and put it back together.
It tends to show up in three shapes:
- The conversion path gets shorter — steps in the middle get eaten.
- What you deliver changes — what used to be a course becomes hands-on coaching with AI behind it. The vehicle changed, and so did what the client walks away with.
- You meet demand that was previously impossible — serving newly created needs in a way that couldn’t have worked before.
The third shape is the one worth sitting with. I’ve talked in an earlier video about using AI for personalized, gene-level nutrition management—something there were previously neither the resources nor the means to do, and now it’s easy.
The point isn’t just “one more product.” Inside my existing pool of users, when I hand them a product I can add value on top of it and raise what a single customer is worth.
Same pool, higher value per customer. It’s genuinely harder to pull off, but it’s the only one of the three that changes the shape of the business.
The precondition under all three
All three come with one precondition: I’m making money right now.
Everything I’ve done and everything I’ve learned over this past stretch was to make the way I already made money faster, more scalable, more automated.
Not this: see a new technology, go learn it. See a skill, go install it. That’s not it.
The difference looks small and it decides where you’re standing a year from now. The first is multiplication on something already proven. The second is addition—and you’re adding in someone else’s direction.
I do the same thing when I work with a client: find the point of highest value in the business they already have, and put AI behind that one point. Not roll out a whole new toolkit.
So
You can genuinely stop watching the videos about how the sky has fallen again.
Spend that time settling down and taking a hard look at how you can actually make money right now—and move from there.
This episode is me on camera, 2 minutes 47 seconds, with the bottom half of the frame showing a live screen recording of me driving CC from my phone. The full transcript is under the “Transcript” tab.
Source: EP0074 original audio track · ASR model gemini-2.5-pro (parallel segments) · full text of the original video
[00:00] A pretty uncomfortable topic. Everybody playing with AI—have you actually made any money? I’ve been looking rough lately, haven’t shaved in days, look at me. I’ve been iterating my web-based Claude Code the whole time, ten days, 240-odd small updates. Out of 50 people who signed up, 16 closed, and those 16 aren’t the ceiling on sales, they’re the ceiling on what I can serve. Because the whole system, besides letting you use CC in the cloud, has a lot of other pieces to it, including answering questions and consulting. And it happens that tomorrow I’ve been invited to give a talk, on the underlying logic of making money with AI, or call it the starting point.
[00:25] Even though I’ve been jumping back and forth between one big company and another, if you actually add it up, the time I’ve spent running my own one-person business already comes to 16 years. So the way I see it, the basic starting point for this whole thing is that I already make money. Even if I’m still working for somebody else—work that used to take ten days, I can now finish in one. The time that’s left, I can spend learning, delivering more value, to the point where I can get promoted and get a raise. Or even take the spare time I saved and do a little something on the side. For an individual or for a company, either way,
[00:50] the cost of headcount comes down. If a person can run a side project, then a company’s first instinct is naturally, can I cut a few people. And from the clients I’m working with right now, whenever AI replaces a business process, it cuts roughly 60% of the headcount cost. But that approach is actually a vicious cycle. The better way is to copy—take the thing, the process, that I can monetize over and over, and turn it into an automated product. Or, with the same headcount, be able to follow up with and serve more clients. That situation is actually
[01:15] harder, because most companies right now are short on orders. There’s no demand. If you’re a one-person business, what you’ll find is that with AI behind you, everything we deliver—99% of what people deliver isn’t fundamentally different. And then it turns into a bidding war. The people who genuinely get a premium for the level of what they deliver are very few. But for a company, this is the more valuable thing to get done with AI, including expanding the lead pool, including lifting the conversion rate, including all those leads nobody was going to follow up on, which I can now
[01:40] follow up on at a pretty high standard. So that’s the second case. And the third case I mentioned, what’s that? It’s using what AI can do to take the whole business apart and recombine it. Maybe the conversion path gets shorter. Maybe what you deliver changes. Maybe what used to be delivering a course becomes hands-on coaching with AI behind it. Or you might even meet some newly created demand in a way that couldn’t be done before. For instance, the thing I brought up in an earlier video, using AI for personalized, gene-level nutrition management. That used to be something there were neither the resources nor
[02:05] the means to do, and now it’s easy. In my existing pool of users, when I give them a product, I can add some extra value on top, raise the return I get from a single customer. Pulling that off is genuinely harder. Now, all three of those points come with one precondition: I’m making money right now. Everything I’ve done and everything I’ve learned over this past stretch was to make the way I already made money faster, more scalable, more automated. Not that I see a new technology and go learn it, see a skill and go install it. That’s not it. Same with how I serve clients—I go find, for
[02:31] the business they already have, the point with the most value, and put AI behind that. So really, stop paying attention to those videos about how the sky has fallen again. Spend the time settling down properly and take a look at how you can actually make money right now, then move on from there. That’s it for today’s video. See you next time.