controlling everything isn't a moat

thursday afternoon i taped an episode with a guy who has spent thirty years between business, technology and design. software engineer, architect, tech exec, filmmaker. now he helps founders build the product and the brand around it.

his booking form was about build versus buy. it's become easy to build almost anything. that, he says, is exactly the trap.

on air he put it like this.

a lot of companies believe that controlling everything is their moat... but what you don't realize is you're creating a ton of internal dependencies.

i think about this every day, because i'm guilty of it.

with ai i can build a scheduler, a crm, a dashboard, a booking page, in an afternoon. every time, it feels like a win. no vendor. no monthly fee. total control.

and the risk is what happens three months later, when the thing breaks late at night and the only one who understands it is me, or an agent that wrote it and doesn't remember why. i didn't remove a dependency. i moved it inside the house and made it my problem.

his answer is to define your core before you build anything. what's the one thing your company does that nobody else can? build that. own it. obsess over it. for everything else, buy, partner or borrow, even when building would be more fun.

every extra thing you build is something you now have to maintain, secure, explain to the next hire and fix when it breaks. it isn't free. it's a loan against your future attention.

he said something about clients that applies to founders deciding what to build. most people can't say what they actually need. his first move is to shut up and listen. then ask the questions that make them doubt their own framing. he told me he'd rather lose the deal than sell someone the wrong thing.

do that to yourself before you build. why do i want to own this? is it core, or is it fun? would a customer ever notice if i bought it instead?

he told a story from an earlier job that shows what listening buys you. companies that engrave stone monuments kept having their diamond tip machines break down, and those machines are expensive. people were standing next to them listening for trouble. he pitched sensors that capture the vibration, so you can see a failure coming and track it over time. a question about a noisy machine turned into a much bigger project.

that's what asking before building gets you. the problem gets bigger and the solution gets clearer.

one more piece i loved. he sees a lot of companies where the brand gets built by one team and the website by another, and the founder's vision gets lost in between. the right hand and the left hand tell different stories. then the founder walks into a pitch and the story doesn't match the site.

the way he sees it, your brand is your reputation. it walks in front of you, not behind you.

same principle. decide what you are first. then everything you build, or buy, tells that one story.

long term, i think this becomes the defining discipline for founders in the ai era. building gets cheaper every month. attention doesn't. the companies that win will build very little themselves, but they'll build that little thing better than anyone alive.

more on building with ai without fooling yourself is in the ai book.

so look at what you built last month... how much of it is your core, and how much is a loan you'll be paying back for years?