Why AI Feels Familiar to Anyone Who Lived Through Multimedia

If you were around computers in the late nineties, you may remember when they stopped feeling like office equipment and started feeling like a shed.

A real one. A bench. A radio playing something faintly irritating. Cables everywhere. Software that crashed if you stared at it too hard.

Multimedia made that possible.

If you could picture an experience, you could usually assemble a rough version of it. Video, sound, animation, clickable curiosities that sometimes worked and often didn’t. Making carried its own momentum.

Most of what came out of that period has faded. Animated logos. Corporate CD-ROMs. Interfaces that dated quickly. What people carried forward was the pace. An idea in the evening could be visible on a screen the next day. Curiosity moved faster than explanation, and effort mattered more than finish.

In the early 2000s, I once spent a full week writing a small 3D engine for a project. It was inefficient and inelegant, but absorbing. I knew what I wanted to see, and the shortest route was to start typing and let the problems present themselves. Progress arrived through adjustment and repetition.

What I was really doing was creative tinkering

Today, the same kind of work unfolds differently. I write a PRD in the morning. By the end of the day, I am reviewing code that broadly reflects what I described. I am still deciding. Still shaping. Still present. The effort has shifted away from operating the machinery and toward directing it. The loop remains human.

Over time, the shed drifted out of view. Platforms took over. Process filled the gaps. Creative work learned how to behave. Experimentation became something to plan rather than something to fall into.

AI has brought back space for that tinkering.

Tools like Claude from Anthropic make it possible to describe something you want to exist and see a version appear quickly enough to work with. Sometimes it runs. Sometimes it stumbles. Either way, the idea gains substance. It responds. It becomes something you can examine rather than talk around.

Much of the current debate misses this shift.** **The change is in how people behave:

• When producing a first version is easy, explanations give way to tests.
• Rough artefacts appear while thinking is still flexible.
• Weak spots reveal themselves before they becoming consequential. This is creative tinkering in its modern form. A person, a set of tools, and enough space to explore without performance being a factor.

**AI accelerates that rehearsal. **

Half-formed thoughts turn into diagrams, interfaces, or working code quickly. Enough to show where the idea needs adjustment.

Ideas leave the head and meet reality sooner. For me, that has meant spending more time working on the problem and less time working inside the process that surrounds it.

The open question is how long this phase is allowed to remain loose. The pull toward formalisation and optimisation is fast arriving. The most valuable work often happens before that pull takes hold, while experimentation remains ordinary and change is still easy.

The tools continue to change. The habit of creative tinkering is what keeps the opportunities coming.