It's not the age of the one-person company. It's the age of the build mentality.

AI has added a new question to the way I think.

For most of my career, when I spotted an interesting problem my instinct was something like: is there a story here? Or an article. A useful argument. Something worth explaining.

Increasingly I’m asking something else.

Is there something I could build here?

The minimum viable organisation is collapsing

I’ve been thinking about this after reading all the discussion around the rise of the “one-person company”. I’m not particularly interested in becoming one.

The more interesting change is that the minimum amount of organisation needed to turn an idea into something useful seems to be collapsing.

A few months ago I started developing Editorial Intelligence as an argument about how editorial work could change with AI. It didn’t stay an argument. It became a website, a knowledge system, research infrastructure, publishing workflows, assessments and experiments.

Then I had a completely different problem. My wife needs to prepare for the Life in the UK test, and most of the material isn’t designed for someone who understands the concepts better in Thai but needs to recognise and answer the questions in English. Previously I would probably have researched it, made some notes and found the best resources. Instead I’m building her a Thai-first learning app.

Even planning our bathroom renovation somehow became a small tracker rather than a spreadsheet or a pile of links.

None of these things makes me a software engineer. That’s actually what interests me.

AI hasn’t suddenly given me 20 years of development experience. It’s lowered the cost of crossing the gaps between the things I already know and the things I don’t. Research. Editorial. Product thinking. Design. Code. Data. Distribution.

I can carry an idea much further before reaching the point where I genuinely need somebody with specialist expertise. That’s a different argument from the one I made about the unit of capability changing, but it’s the same shift viewed from inside a single week rather than across a content organisation.

Why the speed is temporary, and what it revealed

This also explains why I’ve been working at what sometimes feels like a ridiculous pace.

I’m currently on a long holiday in Thailand. I’m spending time with family, travelling, eating far too much fruit and doing normal holiday things. But I also have something I rarely have at home: stretches of time with no full-time job competing for my attention.

So when an idea occurs to me, I’ve been following it. Sometimes quite a long way.

An observation becomes a note. The note becomes an argument. The argument becomes a prototype. A few hours later there’s something working that didn’t exist that morning.

That is also the honest explanation for why so many of these pieces carry dates within a few weeks of each other. It isn’t a content push, and it isn’t twenty years of thinking arriving at once, although the work behind most of it is genuinely that old. It’s simpler than either. Almost everything here would previously have gone into a notepad as a field note and stayed there.

What changed is the distance between noticing something and having a version of it worth showing anyone. These are my ideas being made visible, at the speed I actually have them.

I’m not pretending that will be my permanent pace once normal life and work resume. But having this temporary space at exactly the moment the tools became capable enough has shown me how much further I can take an idea than I previously assumed.

The uncomfortable part

That has created a slightly uncomfortable thought. I sometimes feel as though I’ve wasted a lot of time by not building things earlier.

I know that isn’t entirely fair. The years spent interviewing people, writing, editing, researching and making decisions about what matters are probably why I can use these tools effectively now. And before generative AI, many of these ideas would simply have hit a boundary. I would have needed a developer. Or a designer. Or more money. Or more time. Or another team.

But there’s still a strange sense of urgency.

How many ideas did I previously stop at “that’s interesting” because actually making the thing felt outside my territory?

It’s a weird thing to realise in your forties. Not that AI has made me want to produce more. That it’s made me wonder how much more I could have made.

So while I’ve got this unusual amount of space, I’m moving quickly. Not because I want to spend the rest of my life working at lightning speed. Because I want to find out where the new boundaries actually are.

The models are not the interesting part

I’ve written separately about why I’m not Team Claude or Team ChatGPT, so I won’t repeat the argument here. The short version is that access to the models isn’t much of an advantage if everyone has access to them.

The harder questions are still:

What deserves to exist?
Who is it useful for?
What’s the smallest version worth making?
What’s rubbish even though the AI produced it beautifully?
And when should you stop building?

AI makes it dramatically cheaper to build useful things. It also makes it dramatically cheaper to build useless things.

A build mentality, not a business model

So perhaps this isn’t really the age of the one-person company. Maybe it’s the age when far more people can develop a build mentality.

See a problem. Understand it. Build the smallest useful response. Put it in front of somebody. See what happens. Improve it or abandon it.

I don’t know whether Claude, ChatGPT or something else will be the tool I use a year from now. That’s almost beside the point.

What I don’t want to do is spend another decade stopping at the idea when I can now try to build the thing.