Goodbye, hello

Goodbye Substack, hello Tuhat

When I started writing on Substack last November, I told myself it was a pragmatic choice. Self hosted blogs had come and gone over two decades, victims of my own neglect and the slow entropy of running things alone. Substack had an audience baked in, a clean editor, and the comforting illusion that publishing somewhere with other writers would make me one. I would post my photos, share some thoughts, and the habit would stick. That was the theory.

Six months later I am leaving. Not in a huff, not because anything in particular went wrong, but because the longer I stayed the harder it became to ignore what the place actually was, and what it was slowly turning my own writing into.

I have written about this before, in a piece called Subsocial, so I won't rehearse the full argument. The short version: Substack is a social media platform that has convinced its writers it isn't one. The notes feed, the recommendation engine, the relentless nudges towards engagement, the small dopamine puzzles dressed up as community. All of it serves the same metric every other platform serves, time on site. The fact that the revenue comes from subscriptions rather than third party ads doesn't change the underlying physics. They still need you on the page. They still optimise for the things that keep you there.

What I didn't fully appreciate, until I'd been there a while, was how much that gravity pulls on the writing itself. You start composing headlines for the algorithm without meaning to. You feel the small tug of, this would do well as a note. You notice that the pieces which get attention are the ones with a divisive premise or a clean takeaway, and the ones which get ignored are the ones where you actually sat with something for a while and refused to land on a tidy conclusion. Over time, without anyone forcing you, you write less of the second kind and more of the first. That is the algorithm working as intended.

The bigger problem

There is a deeper thing here too, beyond Substack specifically. The shape of the internet I use every day has been decided, almost entirely, by a handful of American companies, mostly headquartered within a few hours' drive of each other, mostly funded by the same small pool of venture capital, mostly answering to shareholders who want their growth curve to keep bending upward forever. The platform I read the news on, the platform I publish my writing on, the platform my photos sit on, the platform my email lives on, the platform that runs my phone, the platform that hosts most of the websites I visit. Almost all of them, the same five or six firms, the same handful of billionaires, the same monoculture of business model.

I don't think the people who built these things are uniquely malicious. I think they are caught in a financial structure that demands a particular outcome, and that outcome is a smaller and smaller number of larger and larger companies, each of them trying to be the layer through which everything else passes. Enshittification, to use the inelegant but accurate word, is not a moral failing on the part of any individual founder. It is what happens when you point that much capital at human attention and tell it to compound.

The result, from where I sit in Cape Town, is that an extraordinary amount of what shapes daily life (what I read, who I talk to, what my children will grow up assuming is normal) is being decided by people I will never meet, in a country I don't live in, optimising for outcomes I didn't ask for. I am not anti-American. I am anti-monoculture. The two are not the same.

Planting a small tree

The honest response to all this, for someone like me, isn't to write a manifesto. It is to build something small, and then to use it, and then to invite a few other people to use it, and to see what happens. Not a revolution. A tree.

That is what Tuhat is. Tuhat is Finnish for one thousand, and the rule is exactly that: every post must be at least a thousand words. No notes. No threads. No hot takes. No algorithm sorting writers into winners and losers based on how often they post or how spicy their headlines are. You get a page at tuhat.net/u/you, and your readers find you the old fashioned way, through a URL, an RSS feed, or an email subscription you actually own and can export as a CSV.

The constraint is the point. A thousand words is enough room to make an argument, tell a story properly, or sit with something difficult without rushing to a punchline. It is also enough friction that nobody publishes here for the dopamine of it. If you don't have something you genuinely want to say, you won't bother. That is by design.

The pages themselves are plain HTML and CSS. No JavaScript, no tracking pixels, no advertising, no machine learning quietly building a model of who you are based on which paragraphs you lingered on. The whole thing runs on solar powered servers in a small studio called 8by3, which I happen to...