The code I write today is terrible.
My 25-year-old self, who wrote code for monolithic government systems, would be appalled. He would probably fire me.
But he was optimizing for something different. Which brings up the only question that matters.
What are you optimizing for?
In a large organization, you have customers, revenue, and a reputation.
You have something to lose. Losing 5% of your customers is a disaster. Gaining 5% is a rounding error. The optimization is for stability. For preservation. Don’t break the thing that works.
In a startup, you have nothing.
Your biggest risk isn’t a bug, it’s obscurity. It’s not shipping. It’s not getting that first check before the money runs out. You have no customers to disappoint.
If the site goes down, you can call the three people using it and say you’re sorry.
Your customers won’t leave you because of a bug. They’ll leave you because you didn’t solve their problem fast enough.
So today I write messy code. I deliver value and kick the technical debt can down the road. I'm betting that our growth can outrun the debt.
I’m hoping we survive long enough for this code to become a problem. Because having to fix the messy foundation of a successful company is a great problem to have. It means you’re still around.