Eighteen months ago, I sold a car so I could afford to have the first version of Coachstack built.
Last week, I built a new feature for it in an afternoon.
Those two facts have been sitting next to each other in my head for a while now, and they keep circling back to a question I've started getting a lot: if you were starting Coachstack over today, from scratch, would you build it the same way?
The honest answer is no. Not even close. But probably not for the reason you'd expect.
What I thought the hard part was
For most of this journey, I believed the hard part was the building.
It's why I sold the car. It's why I spent months collecting quotes from agencies, why a number like $50,000 and six months felt like the simple price of admission, why so much of my energy went into one question: how do I get this built, and who's going to build it.
I have an MBA. I could have given you a confident lecture on competitive moats over a beer. And I still spent two years quietly believing that my moat, the thing standing between me and a real product, was my ability to write code. I can't build it myself, I told myself, so I need money, or a team, or both.
What actually got easy
That constraint is mostly gone now.
If I started today, the build would look completely different. I'd sketch the wireframes for the whole app visually in Claude Design in an afternoon. I'd take those screens back to the coaches I interviewed, get their honest reactions, and make changes on the spot. I'd turn the parts we agreed on into clear product requirements, and hand those to Claude Code to actually build. I'd run several pieces of it at once, talking to the agents out loud instead of typing, something I watched Peter Steinberger talk about months ago and thought was completely unhinged at the time. I get it now.
The point is this: the thing I once sold a car to afford, I can now do in about a week. Building stopped being the bottleneck.
What didn't get easy
Here's the part that surprised me.
When building gets that easy, you'd assume the whole thing gets easier. It doesn't. The difficulty just moves.
Because if I can spin up a custom coaching platform in a week, so can anyone. The barrier that used to protect a good product, the sheer difficulty of building it, is mostly gone. What's left is the thing that was always actually hard, and is now the only thing that matters: knowing exactly who you're building for, and what they truly need.
A generic, AI-built coaching tool is a commodity now. Anyone can generate one this afternoon. The only reason to build Coachstack instead of one of the thousand lookalikes that will exist by next year is that I spent more than 200 hours and 30+ conversations understanding a specific kind of coach, and the specific problem keeping them up at night.
That understanding can't be vibe-coded. There's no prompt for it. You get it the slow, unglamorous way, by talking to real people over and over until you understand their world better than they can put into words themselves.
So if I started over today, I'd spend more time on customer development, not less. The cheaper building gets, the more the entire game comes down to judgment, taste, and knowing your customer cold. Those are the last things left that are genuinely hard to copy.
The order I'd build in
The other thing I'd do differently is the order.
When I built the Bubble version, I tried to build the whole house before anyone moved in. Portal, billing, scheduling, contracts, all of it, on the theory that I needed a complete product before I could let anyone in the door.
If I started today, I'd build in the order the pain actually shows up. After all those interviews, the loudest problem was never "I need a better client portal." It was "I can't get enough clients." So I'd build the growth engine first, the website, the newsletter, the lead magnet, and leave the back office for later, for when coaches tell me they genuinely need it.
It took me a year and a half and a hard goodbye to learn that the right first feature isn't the one you find most interesting to build. It's the one sitting directly on top of your customer's biggest pain.
The takeaway, if you want one
We're walking into a strange new world where almost anyone can build almost anything. And the instinct, the one I had for two years, is to celebrate that the building finally got easy.
But here's what I'd tell myself if I were starting over. The tools got radically better at building things. They did not get better at deciding what to build, or who to build it for, or in what order. That part is still entirely on you. If anything, it is the whole job now.
So if you're starting something today and the building feels intimidating, I have good news and a warning. The good news is that the building is the part technology is quietly solving for you, faster than you'd believe. The warning is that everyone else gets that exact same gift, which means the easy part can't be your edge anymore.
Spend the time you just got back on the part that's still hard: understanding, better than anyone else, exactly who you're for.
That's the piece no model can do for you. At least not yet.
See you next Saturday.
Peter
