I have been building software for more than 30 years. One thing I learned during all these years is that software that looks good is not necessarily good software.
If you have enough experience, you can usually find the difference. You look at the architecture, the database, how errors are handled, and what happens in situations that were not part of the original plan. Most importantly, you can often understand if the people who built it really understood the problem they were trying to solve.
AI is changing this.
Today, you can describe an application to an AI and, very quickly, get something that looks surprisingly professional. The screens are there, the workflows work, the database is connected, and there is authentication, validation and even automated testing. For someone who is not a software engineer, it can be very difficult to see anything wrong.
And this is what I find interesting. AI is not only making software easier to build. It is making it harder to judge.
The difficult part of software development was never only writing code. The difficult part was understanding what should happen when things do not go according to plan.
What happens when two people change the same information? What happens when a payment goes through but the confirmation fails? What happens when the data is wrong? What happens when the customer changes the process six months later? And what happens two years later, when the original developers are no longer there and nobody remembers why something was designed in a particular way?
These things are not visible in a screenshot. They are also not very impressive in an AI demo. But they are usually what makes the difference between software that works and software that you can trust.
This makes me think that something important may happen in the next few years. As AI gets better at building software, it may become harder to know if the software is actually good.
For a long time, technical competence was relatively easy to recognise. You could look at a product and, with enough experience, understand a lot about the people who built it. This may not be true anymore. The product can look professional even when the thinking behind it is not.
So perhaps we will need to look at different things. Who built it? What questions did they ask before building it? How did they make the important decisions? Do they understand the business problem, or did they simply describe it to an AI?
I am very positive about AI and what it is doing to software development. It allows small teams to do things that required much larger teams only a few years ago. I use these tools myself, and I see the difference every day.
But I think this creates a new responsibility for software companies.
Building software is becoming easier. Knowing what good software is may become more important.
And that is something AI cannot solve for us.
