25 Years and What AI Does Not Replicate

When I started in design, the tools did not hand you finished effects. Making a realistic shadow required actually understanding how light works: direction, temperature, how it falls off. There were multiple layers in Photoshop, each handling a different aspect of light behavior. For animation, it was the same: the principles of squash and stretch came from the physics of objects in motion, not from a prompt. Physics was in the required curriculum of my first year of college for a reason.

You had to understand in order to do.

AI automated and compressed that layer of understanding, so now we arrive at good-looking layouts without knowing how they work. Productivity took the spotlight, and that makes sense: it is a one-way road, and there is no point getting nostalgic about it. The move now is to occupy the space AI has not reached yet: the connections it has not mapped, the problems it does not know exist.

You are no longer the mechanic, you are the pilot who knows mechanics. And I use that reference because driving AI is like driving a vehicle: you only feel safe when you feel the tire gripping the road, the car tracking straight. The worst feeling is when, for a second, you lose that traction, like hitting a sheet of water and hydroplaning. If you are experienced, know the environment, know your tools, the damage is controlled and even avoided, but for someone inexperienced the obvious move is to hit the brake.

Would that be your choice too?