A friend of mine studied sound engineering. For thirty years he placed microphones and measured rooms down to the finest frequencies. He could walk into a room and hear it before the first note played, instinctively, by feel.
He does not do that anymore. Software measures the room now, and the result tests as perfect.
I think about him whenever the debate about AI-made content comes up, because that debate usually skips past what his career actually contained. AI today makes films without a set, a camera crew or actors. It generates images, voices and music at a quality that used to require a full studio. The barriers are falling, and AI companies call this democratisation. That much is true. Many more people can make things than before, and much of what they make works well.
The barriers were also professions, quiet skills and whole ecosystems of people working together, so ordinary that nobody thought to name them.
A film set is a place where people with very different kinds of perception meet in the same room. The camera operator sees what the director misses. Standing at the mirror when the character becomes visible for the first time, the make-up artist reads the character differently. While everyone else is still building the room, the sound engineer already hears it. Each person brings a perception the others do not have. The result becomes something nobody could have made alone, because getting there ran through other people’s heads and hands.
Replace that whole process with one person at a laptop and something still gets made. But it comes from a single perspective, amplified by a machine that is very good at meeting expectations. The output sounds right and looks professional. What is missing is the friction: the moment someone from another corner of the studio says this does not work, let’s do it again.
Craft is a particular kind of knowledge. It lives in the hands, the ear and the eye, and it takes years to get there. A carpenter feels in the wood whether it will move before he even cuts the board. After two notes, a musician already hears whether the room carries the sound or swallows it. For a photographer, a light that will be gone in two minutes leaves one choice: shoot now or not at all.
This kind of knowledge comes from repetition, mistakes and a thousand hours in which nothing remarkable happens. It runs against efficiency, and it produces character, which optimised output does not.
The debate around AI settles quickly on scale: more content, faster, cheaper. That is an economic argument, and it is correct as far as it goes. It leaves out what disappears along with the scale: jobs go, and so does a form of knowledge tied to those jobs, the kind that cannot be exported into a file.
When nobody listens to a room anymore, that listening disappears as a capacity of a whole culture. After a while, there is nobody left who could even notice the difference.
The strange part of this loss is that it stays invisible. You do not notice what is missing if you never knew it was there. An AI-generated soundtrack sounds good. The image on screen looks like studio work. Without a baseline, nothing marks the gap: the people who could have supplied one are doing something else now.
My friend does consulting now. He advises companies on acoustics, it is going well and he probably earns more than he used to. Still, something is missing for him: the work with his hands, walking into a room, listening and knowing exactly where the microphone had to stand. Nobody, honestly, needs that knowledge anymore.
Everything can now get made at the push of a button, and more gets made than ever. What is missing is the years of trained hands and ears that used to make it.