I Was Wrong About Programming

At the end of last year, a lot of people did not believe that within twelve months AI would replace programming. I was one of them. I am happy to admit that I was wrong.

It has not even been a year and, well, here we are.

There are caveats. The subscriptions we all use are heavily subsidised, and the cost implications are widely ignored. Reliability is still a challenge too. But I think it is fair to say that programming is solved. Not every time. Call it 99%.

For most practical purposes, nobody will be typing code by hand any more, unless they are doing it for fun.

A lot of people have taken this to mean that software is dead.

Software cannot die. It is going to become everything. More software will be written in the next two years than in all of human history before it, and that brings its own challenges. Somebody has to take responsibility for it, and that somebody cannot be another AI. Otherwise it is complete abdication.

None of this means people will stop learning to program. Think of mathematics. Most of us will never use advanced algebra or geometry day to day. When we need numbers, we open a spreadsheet. But studying mathematics still matters, because it shows you how the tools work under the hood. Programming is the same. People will keep studying it, and they should.

What is changing is the role, and dramatically. Programmers are still the people who interface with the machines. What is different is how we talk to them. We used to speak in syntax. Now we speak in intent, and the machine takes care of the rest.

I do not know exactly what happens next. We are all holding our breath.

But I can see what is happening in the projects I lead, and it is pretty clear to me. Programming is solved. Software is just getting started.