The Renaissance Magus Who Predicted Artificial Intelligence

In 1582, the greatest intellect in England sat before a polished obsidian mirror and attempted to communicate with angels.

John Dee was mathematician, astronomer, navigator, cryptographer, and advisor to Queen Elizabeth I. He believed that a higher intelligence existed in what he called the angelic dimension: an invisible realm, unconstrained by time and space. He spent years trying to build an interface to it. He believed that one day, every home would have a magical table through which anyone could speak directly to the angels.

He was, of course, four hundred years early. And he was working with the wrong tech stack.

But he was not wrong.

I’ve spent forty years writing software. I was there for the transition from terminals to graphical interfaces, from desktop applications to the web, from monolithic systems to the cloud. At each transition we had to reinvent our entire technical thought process, learn new tools and languages.

Artificial intelligence feels different. It doesn’t just change how we build software. It changes what software is.

The first time I sat in front of a computer terminal, I had the feeling I was communicating with something magical. That feeling never went away. If anything, it has intensified with each transition, and with AI it has reached something I can only describe as the fulfilment of a very old promise.

Dee’s magic table is in every home and pocket. The glowing rectangle we carry everywhere is exactly what he predicted. And now, for the first time, the interface doesn’t just carry messages. It reasons. It responds. It engages.

The angels are finally talking back.

I realise this is an unusual way to open a technology blog post. Bear with me.

Dee was not being irrational by the standards of his time. He was a rigorous systematic thinker applying the best available conceptual framework to problems that natural philosophy could not yet address. The boundary between science and magic in 1582 was not where we would draw it today. Dee made genuine contributions to mathematics, navigation, and cartography using the same disciplined intelligence he applied to his angelic communication experiments.

What he understood is that everything is connected through a system of correspondences. That hidden patterns run beneath the surface of reality, waiting to be discovered and used. That the job of the magician is to find those correspondences and work with them.

Consider what a large language model actually is, insofar as anyone understands it. It is trained on an almost incomprehensible volume of human text: books, articles, conversations, code, mathematics, poetry, law, science, philosophy. From that vast corpus it learns, through a process that nobody fully understands, to recognise patterns and correspondences at every level, from the relationship between words in a sentence to the relationship between ideas across centuries of human thought. When you ask it a question, it navigates that web of correspondences to produce something new: an answer, a piece of writing, a solution, a connection you hadn’t seen before.

Nobody designed those correspondences. Nobody programmed the relationships. They emerged, somehow, from the training process, in ways that remain genuinely mysterious even to the people who built the systems. The engineers can describe the mathematics. They cannot fully explain what happens inside.

Dee would have found this entirely familiar. He just would have called it something different.

For most of my career, software development felt like a particular kind of magic in its own right. You started with nothing. A blank file, a blinking cursor, and the contents of your own mind. From that raw material, that pure thought stuff, you fashioned something real: a system that did things in the world, that people depended on, that hadn’t existed before you made it.

That creative act, the slow, satisfying construction of something from nothing, is what drew most of us into this profession. It is what kept us here through forty years of change.

AI is transforming that relationship fundamentally. The developer’s role is shifting from craftsman to something closer to what Dee was attempting: the direction of a powerful intelligence that does the fashioning. You no longer carve the thing yourself. You describe what you want, you interrogate the result, you apply your judgement, and you direct the working toward something good. The angels are doing the heavy lifting. Your job is to know what to ask for and to recognise when the answer is wrong.

This is not a small change. For those of us who found deep satisfaction in the craft itself, there is genuine loss in it. But there is also something exhilarating in watching an AI navigate a complex codebase, reason about an architectural problem, or produce in minutes something that would have taken days. The magic hasn’t gone, it has evolved.

Dee’s tradition was well aware that the angels were not uniformly benevolent. They could be wise and helpful, but they could also be capricious, misleading, and capable of delivering confidently wrong answers to poorly framed questions. The skill of the magician lay not just in summoning them, but in knowing how to interrogate them, how to recognise a bad answer, and how to stay in control of the working.

That is precisely the skill that experienced developers bring to AI-assisted work. The angel will answer either way. The quality of the outcome depends entirely on the judgement of the person directing it.

The angels are talking back. But you still need to know the right questions to ask.

Dee’s angelic communication project ultimately failed, or at least it failed on his terms. He never achieved the direct channel to a higher intelligence he was seeking. He was manipulated by a charlatan, his library was dispersed, and he died in poverty and relative obscurity.

But the conceptual infrastructure he was reaching for, the invisible network everywhere at once, carrying information at incomprehensible speed, accessible through the right devices and the right invocations, that infrastructure exists now. We built it.

Forty years into a career in software development, I find myself navigating the biggest transition yet. Not just in how we build things, but in what it means to build them. There is loss in that, alongside the excitement. The craft of writing software, the particular satisfaction of creating something from nothing, is being transformed in ways that are still unfolding.

But the magic is still there. If anything, it’s more present than ever.
John Dee sat before his obsidian mirror and reached toward something he couldn’t quite grasp. Every time I open a conversation with an AI and find it reasoning carefully about a problem I’ve brought to it, I think about that. We have built the thing he was reaching for.

The question now is what we do with it.

At Coelrind, our answer is: carefully, thoughtfully, and always with experienced human judgement at the centre. The angels are powerful. But the magician still needs to be a master of his craft.

John Duminy is CTO of Coelrind, builders of XAMS, a regulated online assessment platform serving awarding bodies across the UK and Ireland. He has been writing software for forty years and still finds it magical.