metawritingsoftware-craft

Welcome to fidelitas.io — a manifesto, of sorts

Why I'm starting another technology blog in 2026, and what I'm hoping to do differently this time.

· 4 min read

OK so. I have, against my better judgment, started another technology blog.

There are already approximately ten million of them. Most are dead. The ones that aren’t dead are mostly either thinly-veiled marketing for whatever consultancy the author runs, or LinkedIn-grade “I’ve been thinking about leadership” takes from people whose actual job is to ship code. I read maybe four programming blogs a week and find one of them useful per quarter.

So why this one. Honestly, mostly for me. The job I do for a living involves writing a lot of internal documents that nobody outside the company will ever read — incident reviews, design docs, that one wiki page about the queue that keeps coming up in interviews — and at some point I noticed that the documents I’d put real effort into were the ones I’d want to read myself if I worked somewhere else. So I figured: why not just write them somewhere else.

That’s the whole pitch, really. There’s no business model. There is, as of this evening, no audience either.

What this is going to look like

Roughly four shapes of post, in descending order of how likely I am to finish them:

Field notes. Short. Something broke in production this week, here is the dumb reason, here is the slightly less dumb fix. The kind of post you wish you’d read on Monday and didn’t see until Friday’s incident review.

Tooling. Opinionated reviews of stuff on my $PATH. Sometimes these will be about a CLI that’s older than the internet. Sometimes they’ll be about whatever is currently being hyped on Hacker News. No affiliate links — I make zero money from this and would prefer to keep it that way.

Deep dives. Longer, slower, occasionally pedantic. When something is worth explaining properly, I’d rather explain it properly than turn it into a Twitter thread.

Essays. Once in a while. Nothing on the calendar. These tend to write themselves at 11pm after a hard week.

What this is not going to look like

A few promises, mostly to keep me honest later:

  • No tracking. No analytics, no fingerprinting, no Google Tag Manager hiding inside a “performance monitoring” library. The simplest way for me to find out if you’re reading is to ask, and I don’t intend to ask.
  • No newsletter. I don’t want your email and I really don’t want a Substack revenue line item. Subscribe via RSS or don’t.
  • No popups, no cookie banners, no GDPR consent modal. The site has nothing to consent to. (If you live in a jurisdiction where this paragraph is, technically, the cookie banner, then welcome.)
  • No AI-generated filler. I use tooling to spellcheck and occasionally argue with me about a paragraph. I do not ship words I didn’t write and stand behind.
  • No drive-by hot takes. If a post is going to be strong, it’d better earn it.

That last one is going to be the hardest. I have a lot of hot takes. Roughly 60% of them, when I sit down and try to write them out, dissolve into “actually it depends.” This is, on reflection, mostly a good thing.

What’s already in the queue

I have three pieces drafted, in varying states of readiness:

  1. Why your event-driven architecture quietly became a distributed monolith, and the small handful of things that bring it back from the dead.
  2. Performance engineering as a habit, not a project. (This one’s been in the drafts folder for over a year.)
  3. A love letter to curl. Long overdue.

And one I haven’t started but keep thinking about: Rust vs. Go, but calmer. The conversation about those two has gotten so dumb online that I’d like to attempt a sober version. We’ll see if I have the energy.

Housekeeping

If you want to follow along, the RSS feed is the only mechanism. There’s no comments section, which means there’s no place to argue with me directly. That’s deliberate. Disagreement is fine, even welcome, but it should live on your own blog where you have to take responsibility for it. (See also: the entire history of Twitter.)

Right. That’s enough preamble. Let’s write something.