“It works on my machine.”
Every developer knows this phrase. It’s become a running joke in the industry: you build something, deploy it somewhere else, and suddenly it breaks. The sheepish response? “Well, it works on my machine.”
With AI tools, we’re circling back to something that looks a lot like that old pattern. Except this time, it might not be a problem at all.
When One Machine Was Enough
When personal computers first became a thing, we were fine with software that only worked on one machine. You had a manual process, you automated it with a computer, and that was enough. The solution lived on your machine because that’s all you needed.
The problems started when businesses grew. You’d build something that worked perfectly for you, and then suddenly you needed another person to work on it. You’d add another computer, and everything would break down because the system wasn’t designed to run on two machines.
We spent decades solving that problem. We built systems that could scale, serve multiple people, and work on more than one person’s machine. But to build those systems, we needed professionals. We needed infrastructure. We needed complexity.
The Return of Personal Solutions
Now, with AI tools, I’m watching something interesting happen. Regular computer users can build professional-looking, fully featured solutions that solve their specific problems. Solutions that work for them, on their machine.
It’s “works on my machine” revisited, but with a twist. The tools we have now make it trivial to deploy that personal solution to the cloud and scale it as needed. We’ve already solved the scaling problem. The infrastructure is there.
What’s different is the starting point.
Start With Your Own Mask First
I’ve been thinking about this through the lens of the airplane safety instruction: put on your own oxygen mask first, then help others.
Maybe we shouldn’t give up on building a solution just because we’re already thinking too big. Maybe we shouldn’t start by trying to serve the entire world. Let’s start by making sure it works for us. It solves a problem we have, a need we have.
There’s another angle to this. Maybe you don’t have enough time to help others because you’re busy with certain tasks. If you have a need to help others, and the problem is that you don’t have enough time, you can use these tools to build solutions that do some of the things you need to do. Compress that time. Then use the freed-up time to help others.
That’s another way to put on your own mask first.
When you solve your own problem, something shifts. You feel fulfilled. It shows in your body language, your facial expression. And people notice.
They ask, “You look great, what’s up?”
And you say, “I had this problem, and I built a solution for it. Want to see?”
That’s when they might say, “Can I use it too?”
And then you figure out how to make it work for more than just one person.
My Own Examples
I have automations I’ve built that have recently helped with my journaling practice and my blogging workflow. They work on my machine. They work for me because of my specific process: getting thoughts out, processing them, and turning them into blog posts.
These two things might be separate for other people. For me, they walk hand in hand.
If I showed this to someone and they liked it, maybe they could adapt my workflow to theirs. Or maybe they’d try it, see what they learn, and refine it into something that works for them.
But I didn’t build this to make it available to others immediately. I’ve been documenting it. I’ve been telling people who ask. I’ve been blogging about it so I can look back and see how my system has evolved.
And I keep tweaking it, almost weekly. Not for the sake of tweaking, but to make it a little bit better. To achieve the outcome I want, either a little bit faster or a little bit better, or both.
I’ve done similar things with spreadsheets. I’d create a spreadsheet because it was the easiest thing to get work done. Then I’d feed it into an AI and get it to create an actual application: one that gathers the data, displays it in multiple ways, and lets me work with information rather than just raw data.
The Derek Sivers Story
One of my favorite stories about this pattern is the one about Derek Sivers and CD Baby.
Derek was a musician in the mid-to-late 90s. He wanted to sell his band’s CD online, which was difficult at the time. So he picked up a few books, learned how to build a web application, worked with banks to accept credit card payments, and started selling his band’s CD on his website.
One friend saw it and asked, “Can you sell my band’s CD there?”
“Yeah, sure.”
Then another friend asked. Then friends of friends.
The next thing he knew, he had a business. Not what he initially wanted. He only wanted to solve his own problem: selling his band’s CD. But he was able to help others, and it turned into a huge business. After 10 years, he sold it for millions of dollars.
The full story is in his book Anything You Want, and it’s worth reading.
What This Means Now
Pay attention to the problems you have. The needs you have. Solve those problems. Fulfill those needs.
Maybe other people will see it and want it. Then you have options. Do you want to offer this to others and monetize it? Or is it just something you want for yourself?
You can also look around from the start. Watch people’s body language and facial expressions as they navigate the world. When you see someone using a computer and letting out a loud sigh, that’s a person with a problem or an unfulfilled need.
Talk to that person. Maybe you can help them.
But start with your own mask first. Build something that works for you. Then see where it leads.
Leave a Reply