Software is a cheeseburger

I'm not a chef, but I can make a cheeseburger at home.

It won't be the best cheeseburger you've ever had. But it'll be hot, it'll have cheese on it, and it'll be exactly how I like it. Nobody had to train me. I just had to want one.

I've started to think about software the same way.

Anyone can cook now

I did Ruby on Rails professionally for 15 years before I started working with AI. For most of that time, building software was like running a restaurant kitchen. You needed years of training to work in one, and most people never went in.

That's changed. Now someone who has never written a line of code can describe what they want and end up with something that works. Not a professional kitchen, but a home kitchen. And a home kitchen is enough to make a lot of things.

The kitchen gets better the more you put into it

The first cheeseburger you make at home is probably fine. The tenth is better. You buy a cast iron pan. You learn not to press the patty. You figure out which cheese actually melts.

Software is the same now. The more time you spend on your setup, the better your results: your instructions, your tools, the habits you build up about how to ask for things and how to check the answer. Two people with the same AI can get very different software out of it, the same way two people with the same stove can make very different burgers.

So why would anyone still buy one?

Even though I can make a cheeseburger at home, I still buy them all the time. For three reasons.

Sometimes I'm lazy. I don't want to cook. I want a cheeseburger, now, and I'm happy to pay someone so I don't have to think about it. Plenty of software will stay like this. I could build my own calendar app. I don't want to. I want one that already works.

Sometimes I want a really good one. There's a cheeseburger I can make at home, and there's a cheeseburger from a place that has made ten thousand of them, tested every part of the recipe, and fixed everything that went wrong along the way. I can't match that at home, and I'm not going to try. Software that's been used by thousands of people, broken in every possible way and fixed each time, is worth paying for. That kind of quality only comes from years of other people's mistakes.

And sometimes the one at home is good enough. It doesn't have to be the best. It just has to be mine and do the one thing I need. A lot of software will be like this: small tools built for one person, one family or one team, that would never make sense as a product.

Everyone still has limited time

This is the part I think matters most. AI made cooking easier. It didn't give anyone more hours.

Every cheeseburger I make at home is time I'm not spending on something else. So the real question isn't "can I build this?" Almost everyone can now. The question is "is this worth my time to build, or would I rather just buy it?"

That's where I think software is going. More people cooking at home than ever, and restaurants still full. The software people pay for won't be the stuff that's hard to make. It'll be the stuff that's worth not making yourself: because it saves you the time, or because it's better than anything you'd make in your own kitchen.

Anyone can make a cheeseburger. Some are still better than others.