Two pillars on purpose

The next two posts on this site are about a 15 kilometre run and a piece of broken software. In the first, I am two kilometres into a long run and everything in me wants to stop, with thirteen still to go. In the second, I make an AI project work by deleting most of the documentation I had carefully written to help it. One is not a warm-up for the other. They are here together on purpose, and that is worth explaining before either of them lands.

Every guide to starting a blog gives the same first instruction: pick one thing. Choose a lane, own it, become the person people think of when they think of that subject. Write only about Postgres and you become a Postgres site. It is the most common advice in the genre, and this site ignores it.

The advice is not wrong. It is optimising for things I have decided not to optimise for. A site that is obviously about one subject builds search authority faster, because a search engine can tell what it is an expert in. It grows an audience that knows exactly what it signed up for. It is far easier to make money from. If the point of this site were reach or income, one subject would win, and I am not going to pretend otherwise. The point is something else: an accurate picture of the person doing the work, which is a worse strategy for traffic and a better one for the thing I actually want.

There are people who allow the mix, and they tend to justify it two ways. The first is to find an umbrella. Call yourself a lifestyle creator, brand the overlap, turn two subjects into one product so that travel and cooking become the food you eat while travelling. That is a marketing move, tidying two things into a single sellable shape. The second is the version developers reach for: show the whole person behind the code, be authentic. It sounds right and it says nothing. It would equally justify posting a photo of your lunch beside a code snippet and calling the pair a brand. I do not want either. Not an umbrella, and not authenticity as a slogan.

The real reason the run and the broken software belong on one site is that I do the same thing in both. I read the signal and respond to what it is actually telling me, rather than what I assumed it would say.

On the run, the wall at two kilometres was not the famous one. The famous wall comes late in a long effort, when the fuel genuinely runs out. Mine arrived ten minutes in, because I had gone out too hard, and it was pacing and mood wearing the costume of something serious. The whole skill is telling those two apart: is this a real warning, or is it noise. That is discrimination, not willpower, and it is the thing that carried the long project of getting from 137 kilograms to 90, where the movement kept changing and the reading underneath never did.

At the keyboard it looks completely different and it is the identical move. The AI project got worse the more documentation I gave it, so I took the documentation away and it got better. A 404 I chased for two days turned out to be a dev server I had started earlier in the week and forgotten was running. An admin panel rendered blank in production, twice, because one committed file did not match what the build expected. None of that is expertise on display. It is reading what is in front of me instead of what I expected to see, on a road and in a terminal, and it is the same person doing the reading.

The fair objection to all of this is that it is a story I tell to avoid choosing. That refusing to niche is not a philosophy, it is indecision with better production values, and the through-line is a permission slip I wrote for myself. Maybe. The test is whether the writing earns it. If a post about running here teaches you nothing you could not get from a hundred other running posts, the objection wins that day, and no amount of talk about a shared approach saves it. The connection has to show up in the specifics or it is not real, which is a standard each post has to keep meeting rather than a claim I get to make once and bank.

And then there is the part no marketing framework has a slot for. Some of what I write here will not serve anything. Not an audience, not the work, not the clients the work might eventually bring. I will write about running because I run, and about building things for no reason beyond wanting them to exist, and sometimes about whatever I am actually thinking about that week. A personal site is one of the few places left that does not have to justify its contents to an algorithm.

So there are two posts coming, a run and a broken build. Read one, or read both. Same person either way.