I think it’s great that we’re in the age of AI, but I also feel we’re leaving something behind. These tools let someone with technical skill produce more in many scenarios (they clear out repetitive and even boring tasks), yet maybe that piece we are removing is fundamental to the puzzle.

We’ve never had a time so practical and capable for building things, with free or low-cost tools, easy access, and robust documentation. Still, maybe the true integrity of value lived in the difficulty of access and in the formal learning curve of a subject. I’m not saying new teaching modes aren’t welcome (they are). I’m saying that using AI every day for everything erases the “secret” inside difficult learning, and maybe that very difficulty carries lessons we cannot replace. Today those lessons get drowned out by a bunch of simple instructions (copy, paste, run).

Take AI-generated media as an example (videos and even dubbing). Working with audio engineering taught me something crucial: the artist or voice actor owns the intellectual and sonic property of their track. Now picture a current scenario where a voice actor is easily swapped by a machine for a teenager’s school project. Often the property of that voice ends up used not only in educational projects but also in commercial ones (which directly affects the actor’s career).

I’m not here to mourn the end of careers (new kinds of work will come). My concern, as a person, as an engineer, and as a developer who has been curious since childhood, is that heavy use of AI might dim the curiosity and the joy of building and designing projects in an organic way.

Inside big companies we already hear the line that, in the future, we won’t need so many professionals. Areas that used to have a talent deficit can now be reduced by hundreds or thousands within the same company. But are we actually taking the right path as individuals and as humans? Times of transformation and powerful tools have always existed. I keep asking whether tilling the land with a hoe or with a tractor leads to the same outcome (not only in terms of the final artifact, but in terms of the joy of building, the motivation to continue, and the hope of solving problems).

I don’t see much of that joy today among most developers I meet. Even the self-styled builders are often making replicas of replicas instead of crafting robust, high-quality software. It feels like the art of caring about code problems turned into a systems-design prototype exercise (rather than a truly exploratory environment).

How many fintechs launched in the last few years since LLMs started to spread? Now think about how many real software houses you saw emerge in that same period.

Yes, this is about supply and demand. But maybe we’re too focused on “saving” in areas where intellectual capital should develop alongside creative capital. New technologies and LLMs do make access easier (that is good), yet they still feel premature in many ways. Seeing a tool as a worker might be riskier than seeing it as a long-term productivity boost (and that difference matters).