From shovel to excavator: AI in software development

Last week, I noticed a wet patch in my otherwise dry garden. Since there hadn't been any rain in the past week, I decided to figure out where it was coming from. When I checked the main meter, it turned out that quite a bit of water was seeping away.

Full of good intentions (and a bit naïve) I started digging at the wettest spot. After about an hour of digging, I had discovered that the area contained a lot of roots and an old sewer pit. After another two hours of digging, I came to the realization that both the roots and the sewer pit were not making things any easier, and that I had actually made very little progress. There was still no sign of any leak, apart from the fact that the soil grew wetter the deeper I went.

Since manually removing that monstrosity of a sewer pit didn't seem particularly realistic, I called in the help of our gardener. He was kind enough to come to my aid with his... excavator. Ten minutes later, the sewer pit had been removed, we had dug another 80 cm deeper, and the leak had been found.

While watching the gardener at work, a thought struck me: the AI models we use today are our excavator. Until a few months ago, we — developers — were still mostly working with the shovel. Now we can build and solve problems much faster too.

It actually gave me an optimistic feeling. My gardener still has all the skills a good gardener needs (knowledge of plants, construction experience, dexterity, and so on), and he has had to learn how to work with his excavator. Thanks to that excavator, he can take on much larger projects, not to mention all the other equipment at his disposal.

That's why I'm (at this moment at least) convinced that (in the near future) the developer/architect of the future will still need the necessary skills to build quality software (architectural knowledge, creativity, the ability to understand code, analytical insight, and so on), and will be able to distinguish themselves through those skills. On top of that, they will need to learn how to work efficiently with all the available tools, which will allow them to deliver larger projects where the bar is set higher.

The fact that technology is advancing exponentially, that junior developers are having a harder time learning the ropes, and that the cost may become prohibitive: we'll set those aside for now. But that we need to swap our shovel for an excavator is beyond doubt.