An open letter to anyone still building software the old way
Here's something every software builder knows but rarely says out loud: nobody gets it right on the first design. Or the second. Sometimes not even the third.
In the old world — which, by the way, was six months ago — this was a problem. A big, expensive, soul-crushing problem. Redesigns cost time and money. Rework burned goodwill. Most teams hit a point where they stopped iterating and started settling. You'd look at a flawed product and think, well, let's just make the best lemonade we can out of these lemons. Not because it was good enough, but because doing it over was too painful to justify.
I lived that reality for my entire software career. Human teams build the wrong thing. It happens constantly. And when they do, you're stuck — because it's too fucking expensive and too time-consuming to start over. So you ship something compromised and move on.
That era is over.
What actually changed
AI-assisted software development didn't just make building faster, though it did that too. It did something far more important: it brought the builder and the end user closer together than they've ever been.
Think about what that means. Layers of abstraction — project managers translating requirements, weeks-long sprint cycles, feedback loops measured in days — all of it compressed. With AI, I get feedback in minutes, not days. I iterate in hours, not months. And when something goes in the wrong direction? I don't feel an ounce of guilt asking AI to scrap it and rebuild from scratch. Try asking a human team to do that after two weeks of work. You know the answer.
We stopped saying no
Here's the part that really changed my thinking. In the software business, we are constantly fighting a two-front war: upselling new capabilities while battling scope creep. It's a motherfucker. We say no to feature requests all the time. Not right now. Maybe next quarter. That's out of scope. Every builder knows the dance — finding polite ways to wriggle out of awesome ideas from our users because the cost of saying yes was just too high.
With AI, we say yes. And it appears minutes later.
Read that again, because it's the part most people haven't internalized yet.
We. Say. Yes.
This isn't about speed. It's about magic.
Yes, software gets built faster. Yes, it's more performant. But framing this as a speed improvement misses the point entirely. This is about making software better. Making it magical. We can do things now that people couldn't even imagine requesting — because they'd learned not to ask.
Six months ago, I couldn't do what I can do today. Now I can. And so can you.
The quiet revolution
The world changed. No one made an announcement. There was no press conference, no ribbon-cutting. But we are living in a new world. The tools are here. The capabilities are real. And what we can build now — what any motivated person can build now — is nothing short of amazing.
So if you're still building software the old way, still saying no to good ideas, still settling for flawed products because rework is too expensive — I'm writing this to tell you:
You don't have to anymore.
The world has changed. Come build in it.