I was always technical. I just wasn't a developer.

I became a technology journalist because I was interested in how things worked.

Not only computers themselves, but what happened when new technology reached businesses and ordinary people. Why one product succeeded. Why another disappeared. How technical change altered work, behaviour and power.

I could understand the technology well enough to question it and explain it.

But I was always better at stories than maths.

That seemed to establish the boundary of my career quite early. I would write about technology. Other people would build it.

For most of my working life, that felt like a reasonable division.

More recently, I began to regret it.

Watching technical careers move ahead

Computer science, software engineering and data became some of the most valuable capabilities in the economy.

The work often appeared better paid, more strategically important and closer to the decisions that shaped products and companies.

Writing moved in the opposite direction.

Writers were frequently positioned at the end of the process. The product had been built. The strategy had been decided. The campaign had been planned. Somebody now needed to explain it.

The closer you were to building the system, the more valuable you seemed to be.

The closer you were to producing the words, the easier you were to treat as a supplier.

I occasionally wondered whether I had chosen the wrong side of the divide.

Perhaps I should have learned programming properly. Perhaps I should have moved towards data. Perhaps I had mistaken an interest in technology for a capability that could support a career.

There is some retrospective simplification in that. Not every developer is highly paid, writing and coding are not opposites, and software development does not require somebody to be a mathematical genius.

But the regret was real.

I could see technical careers creating forms of leverage that writing rarely offered.

Management never felt like the alternative

There was another route available to experienced writers.

Management.

You could stop being the person making the work and become the person allocating it, presenting it, defending it and talking about it in meetings.

That path never felt particularly natural to me.

I was not drawn to managing people, and I was rarely the most confident person in a room. Agency life made the difference particularly visible. Account directors seemed able to speak easily, manage clients and occupy space. I was usually more comfortable listening, finding the argument and making the work better afterwards.

Some of that may be temperament.

Some of it may come from growing up around a dominant parent. I learned to hold thoughts back, avoid confrontation and wait until I had worked something through before expressing it.

Those habits can help a writer observe carefully. They are less useful in environments where influence belongs to whoever speaks first and most confidently.

Tinnitus made the social and physical demands of that work harder too. Noisy rooms, long meetings and the pressure to remain constantly available could be draining in ways that were not always visible to other people.

I have become more confident as I have grown older.

But when I was trying to progress through agencies and larger organisations, these things definitely held me back. Advancement often depended on live performance: speaking quickly, managing upwards, projecting certainty and making yourself visible.

Loud frequently won.

Not necessarily because the loudest person had done the best thinking, but because their thinking was the easiest for the organisation to see.

AI has changed some of that for me.

It does not remove the need to communicate, persuade people or operate inside organisational politics. But it reduces the penalty for not doing your best thinking aloud and immediately.

I can take an incomplete thought away, develop it against existing evidence, challenge it, build something around it and return with work people can examine.

That might be a finished argument, a functioning website, a new framework, a prototype or an implemented change—not merely a suggestion that somebody else must interpret and deliver.

I can do this at something much closer to the speed a business expects.

The work gives me a stronger way to enter the conversation. I do not have to win the room before anyone can see the quality of the thinking.

The standard progression once appeared to offer two choices:

• remain a writer while the perceived value of production declined;
• or move into a managerial role built around meetings, stakeholder performance and responsibilities I had never particularly wanted.

AI is creating a third route.

It gives an experienced individual contributor more leverage without requiring them to become a conventional manager. I can coordinate tools and agents, maintain a system and move complex work forward while remaining close to the thinking and making.

I still need to present the work.

I just no longer need to be the loudest person in the room before the work exists.

My son thinks differently

My son finds maths relatively easy.

I can see a kind of fluency in the way he approaches it that I do not remember having myself. Numbers and rules appear to give him something stable to work with.

My own mind has always seemed to work differently.

I collect fragments.

A comment in a meeting connects to something a customer said six months earlier. An article changes how I understand an old research programme. A television programme suggests something about unreliable evidence. A disappointing LinkedIn result becomes a question about distribution, narrative or what audiences actually value.

I put things together until they begin to reveal an argument.

That is why journalism made sense to me. An interview rarely arrives as a finished story. You gather observations, quotes, contradictions and context, then decide what they mean together.

It is also why I moved naturally into B2B technology writing. The job was rarely just explaining a product. It was understanding the market around it, the problem it claimed to solve and the story the organisation wanted people to believe.

I did not think of this as technical ability.

I thought of myself as a writer.

AI has made that distinction much less useful.

A different kind of technical thinking

Working with AI has shown me that I think technically through relationships rather than calculation.

I am interested in:

• how pieces of information connect;
• how one decision affects another part of a system;
• how evidence becomes an argument;
• how different tools can perform different roles;
• how knowledge can be stored, retrieved and improved;
• and how separate capabilities can be orchestrated into an outcome.

This does not make me a software engineer.

But it is a form of systems thinking.

Before generative AI, expressing that kind of thinking usually ended in a document. I could propose a framework, write a strategy or describe how a better system should work.

Turning the idea into something operational often required developers, designers, data specialists and a budget.

AI has weakened that boundary.

I can now work with a model to develop an idea, ask an agent to inspect a repository, change a website, record a decision or connect new evidence to an existing framework.

I still need technical systems. I still depend on code. But I no longer need to be the person who writes every line of it before I can build something useful.

AI did not teach me to code.

It made the boundary between understanding technology, explaining it and building with it much less important.

Why agentic work feels natural

I recently tried to explain how I work agentically with ChatGPT, GitHub and Claude Code.

The workflow sounds technical when I list the tools.

But the underlying behaviour is familiar.

I notice something. I connect it to previous thinking. I decide what it might mean. Different tools help develop, record, challenge and implement it. I review the result and feed what I learn back into the system.

That is not very different from how I have always worked editorially.

The difference is that the connections can now produce more than an article.

They can change the system holding the work.

An observation can become a research question. A research question can alter a framework. A framework can change a website page. Audience response can be recorded as evidence for the next decision.

The workflow moves through capture, connection, creation, activation and learning.

I find this way of working unusually intuitive because it resembles the way my mind already works: not in a straight line, but by assembling pieces and looking for the relationship between them.

The AI provides execution and reach.

I provide direction and judgement about which connections matter.

Editorial Intelligence uses more of me

Editorial Intelligence may be the first thing I have built that uses the whole of my career rather than the job title I happened to have at the time.

Technology journalism taught me to understand unfamiliar subjects and question claims.

Agency work taught me how organisations construct messages around products, markets and commercial goals.

In-house content work showed me how much evidence companies generate and how quickly they lose it.

Research programmes taught me that the published report is only one possible use of what has been learned.

Writing taught me to recognise an argument and make it understandable.

The difficulties mattered too.

Being quieter made me a better observer, even when it made conventional progression harder. Tinnitus made me more aware that a working model built around constant meetings and live interaction was never going to bring out my best. Remaining an individual contributor kept me close to the substance of the work.

AI has given me a way to connect all of those experiences into a working system.

Editorial Intelligence is partly a methodology for turning organisational evidence into ideas, narratives and reusable knowledge.

But it is also a response to a more personal question:

What can somebody who understands technology, evidence, systems and stories build when the cost of technical execution suddenly falls?

I am still working out the answer.

The regret has changed

I still occasionally wish I had learned more technical skills earlier.

There are things I would understand better. There are systems I could build more independently. There are opportunities I might have recognised sooner.

AI does not erase those differences. Access to an agent is not the same as years of engineering experience.

But it has changed what I think I missed.

Perhaps I did not choose between being technical and being creative.

Perhaps the available careers forced those capabilities into separate categories, and I followed the category that best matched the tools I had at the time.

Those categories are now becoming less stable.

Experienced people can combine subject knowledge, judgement and technical execution in ways that previously required a larger team.

The scarce capability may not always be writing the code or producing the words. It may be understanding what should be built, what evidence it should use and how the pieces should work together.

That is much closer to the work I have always been trying to do.

Perhaps this is not the path I was always destined to take.

But it may be the first path that makes sense of everything I took before it.