I could build that

Some of you, watching the tech do its work through these pieces, will have had a very particular thought: “I could build that myself.”

It’s a fair thought, and an honest one - I’d probably have it too. You’re a capable person. Wiring up an AI to read a profile and draft a message is not, in 2026, some impossible feat. So let me take the idea seriously, because the answer is more interesting than a simple yes or no.

Start with what you’d actually be copying. The software you can see is the easy part - genuinely. The hard part is invisible, and it’s underneath: years of working out who is even worth reaching out to, what makes a first message land instead of getting deleted, which flicker of attitude in a profile tells you this person will build and that one won’t, how a discovery call should feel, what a healthy node looks like and what a dead one looks like. What looks like a clever bit of software is really scar tissue with an interface on top - a decade of getting it wrong, distilled into something that now gets it right. You can copy the interface in a weekend. You cannot copy the decade. It was never the idea that was hard; it’s the execution, and in untrained hands the very same idea mostly fails.

But say you could. Say you had the years and the judgment to build a real one. Should you?

Almost certainly not - and here’s the part worth sitting with. If you built it, you’d be pouring your scarce, valuable hours into specialising in the wrong thing. Tech and AI aren’t the differentiator anymore; they’re the given, the background hum, the thing everyone will simply have (that was the last piece). Your edge - the thing that’s actually rare - is your own domain. The work you do, the value you bring, the years you’ve banked. That’s where your hours should go: getting better at the thing only you can do, and letting the tech run behind it. Let the people who’ve spent their decade on the tooling handle the tooling, the way you’d want people to trust you with the thing you’ve spent yours on. Everyone specialising where they’re deep, AI quietly behind all of it - that’s not a compromise, it’s the whole shape of how good work gets done now.

And there’s one last tell, the one that really settles it. When a capable person builds a thing for themselves, they build a tool. A tool works in exactly one pair of hands - their own. But the thing that actually changes your working life isn’t a tool, it’s a system - and the difference is that a system duplicates. It carries the judgment inside it, so it works in hands that don’t have the years: you can hand it to an assistant and the assistant is suddenly operating at your level, because the hard-won part lives in the system, not in the person. Build-it-yourself, at its very best, gets you a clever tool. What you actually want is the thing that duplicates.

So don’t go and specialise in the plumbing. Specialise in you - and in the people you choose to build alongside.

Which is the real question hiding under all of this: of everyone you could build with, who is actually worth it? That’s next.