I started with AI badly. That’s the whole point.

Squares on paper, 1980.

My dad taught computing for a living, and he brought it into my life early. I learned to program by colouring in squares on paper and feeding punched cards into a machine. My first computer was a Sinclair ZX80 with 1K of memory. If the tape stretched, your program was gone.

Since then I’ve made a living riding most of the technology waves that came along: theatre lighting boards, broadcast editing, 3D visualisation, digital film, drones and offshore inspection robotics. I trained more than two thousand broadcast professionals as a master instructor. Today I’m a Senior Robotics Operations Engineer, and I use AI every working day.

AI doesn’t tell you when it’s wrong. You have to know enough about the subject to catch it yourself.

When AI arrived, I still got it wrong. My early prompts were weak, and where I didn’t know a subject well, the answers sounded right and weren’t. That lesson is the foundation of everything here: the person using AI is the quality control, so your years of knowing what good looks like are the part the machine can’t supply.