The first generation of AI-made products could end up like the first photos taken on mobile phones.

The more experienced among us will probably remember the moment when mobile phones first incorporated a camera.

Suddenly we were photographing everything. We abandoned the old digital cameras; analog ones were already gone, and everything started to be captured: friends, parties, dinners, trips, absurd moments inside the car, everyday scenes that before we would not even have thought about saving.

A new world opened up.

We took thousands of photos. Like never before.

But over time something became evident that at first almost nobody noticed.

The quality of many of those photos was mediocre, sometimes very poor: images taken quickly, without attention, without selection, and very often without any real criteria. And a part of all that was simply lost along the way, forgotten in old phones, broken hard drives, and folders that nobody ever opened again.

That explosion was real. But so was the noise.

And with AI something similar may be happening.

In the middle of the hype, it seems that everyone programs. That programmers are no longer needed. That anyone can build products, tools or software at any moment and from anywhere.

And yes, surely some of this will be valuable.

But it is also possible that, with time, we will discover the other side of this explosion: products that become obsolete too quickly, low-quality solutions, projects built on services that stopped being paid for, broken tools, lost prompts, logic that nobody really understood, and code that nobody wanted to maintain.

Much of what today seems good enough may not age well.

And much of what today dazzles us may end up lost, abandoned or irrelevant.

Then we may remember something important.

That it was never only about producing more.

It was about understanding what was being built.

About being able to sustain it.

About knowing why it was built that way, what dependencies it had, and what would happen when the initial enthusiasm disappeared.

Then we may also understand that the real problem was never producing fast.

The problem was confusing speed with solidity.

Because generating more does not guarantee building better.

And sooner or later, the same things as always will matter again: judgment, structure, vision, and technical responsibility.