Origins

First came the urge to take things apart

Some of my earliest technical memories are physical ones: a children’s electronics set, wiring switches, resistors and lights, and the fascination of building an FM radio on a breadboard. I remember trying to work out how the VCR worked and taking apart more than a few broken electrical things in the hope that understanding the inside might somehow lead to repairing them. Success was optional.

The important part was the instinct: technology was not a black box that had to stay closed.

Then the computer became something I could change

I copied BASIC programs out of Commodore 64 magazines, character by character, and then waited for the next edition to print the addendum containing the bug fixes. It was an extremely slow software-update channel, but a good way to learn that programs were understandable, fallible and fixable.

One of the first original programs I remember writing, probably on the Amiga, was for my dad, who bred budgies. Given two parent birds, it tried to predict the colour of their offspring. The implementation was awful: a vast pile of IF statements covering every combination I could think of. But it worked. Real-world knowledge had become executable rules.

Before university, software was already doing real work

As a teenager I built two Microsoft Access applications for my dad. He was Head of Security at a prison, and one of the applications registered and tracked vehicle number plates. I have forgotten what the second one did, but he later told me the applications were used there for years before he retired.

That mattered. Software was no longer only something I was learning or playing with. People could depend on something I had made.

I studied Computing even when I was not studying Computing

For GCSEs I was not allowed to choose Computing because I already had too many technical subjects: dual science, Graphic Communication and Design & Technology. Physics fascinated me, although I was frustrated by being taught the answer without enough of the why. Graphic Communication introduced me to technical drawing and AutoCAD. Design & Technology gave me an early, rudimentary robot arm.

Being unable to take Computing did not stop me asking the teacher for the reading material and exercises and completing them for fun.

At A-Level I finally took Computing formally. I was top of the class, studied subjects including PGP encryption, sometimes found myself explaining things back to the teacher, and finished with an A.

Reading gave the obsessions names

I studied Applied Computer Science at the University of Reading from 1997 to 2001. My favourite subjects came from the Cybernetics department. Professor Kevin Warwick was a major influence, along with one of his graduate students.

The lecture that stayed with me was called, or at least centred on, a deceptively simple question: what is a chair? What properties make something a chair? Shape? Purpose? Context? How would a machine know? That question captured something I still find interesting: familiar things become much less obvious when you have to describe them precisely enough for a machine.

My original final-year project was an attempt to write an ELIZA-style conversational program from scratch in Pascal. I read into natural-language processing, tokenisation and the machinery required to turn language into something a program could reason over. It was too ambitious for the time available, so I dropped it rather than pretend I could finish it properly.

The replacement project was not exactly conservative. I built a home automation system using sensors and lights, and a robot arm controlled by it.

BT taught me to think about what happens when systems fail

During my year in industry I worked at BT in Martlesham Heath on integration testing for Directory Services. I also wrote the rollback guides for Directory Services Y2K incidents.

That was an early production lesson in something that would keep returning throughout my career: independently built systems have dependencies, assumptions fail, and recovery needs to be designed before the failure happens.

Around the same period I spent a lot of time with SuSE Linux. Long before my career became heavily associated with Microsoft technology, I was already interested in what sat underneath the desktop and what happened when the operating system exposed more of itself.

The cultural influences were less respectable, but just as real

The Lawnmower Man and the early 1990s VR arcade machines started a fascination with virtual and mixed reality long before headsets became consumer products. Hackers, and especially Sneakers, fed an interest in security, systems and the gap between how something is supposed to behave and what somebody can actually make it do. I read the Jolly Roger’s Cookbook repeatedly as a teenager. It was an artefact of early internet hacker culture that mixed technical curiosity with plenty that should not be mistaken for good guidance.

Those interests kept resurfacing: PGP and Linux, enterprise security, VR experiments, mixed reality, robotics, cloud architecture and eventually autonomous software.

Then I became interested in the other complicated system: people

After university I moved to London without a huge amount of social confidence. In the early 2000s I discovered the pickup and self-development literature around The Game and disappeared down another rabbit hole.

The lasting fascination was not the manipulation or the more dubious parts of pickup culture. It was psychology: body language, communication, persuasion, confidence, social feedback and the idea that human interaction could be observed and practised rather than treated as magic. I read widely, including material around Neuro-Linguistic Programming. Some of its stronger claims have not aged well scientifically, but the interest in how people communicate absolutely did.

Years of consulting, community work and public speaking came later. This was one of the places where I first started deliberately studying the human side of communication.

Looking backwards, four threads keep crossing

The maker: electronics, BASIC, Access, robot arms, automation, products.

The investigator: taking things apart, physics, encryption, Linux, cybernetics, integration testing, resilience and security.

The world builder: elaborate childhood scenes with Star Wars figures and Action Man, then VR, simulations and eventually environments populated by autonomous agents.

The observer of people: psychology, body language, communication, consulting, speaking and the question of how meaning survives between one mind and another.

I would not pretend there was a master plan connecting any of this. There was not. This page is hindsight, not destiny. But it does explain why the current work feels less like a pivot into AI and more like several very old interests finally meeting in the same place.