Reading notes, first half of 2026
Brief notes, mostly so I can remember what I thought later. No ratings — I never know what a rating is supposed to mean.
On maintenance
The useful idea here was that the interesting cost isn't building something, it's the years afterwards. Most of what I've had to fix on my own machines failed in ways that made sense in hindsight, because something quietly depended on a detail nobody wrote down.
On complexity
A recurring theme in everything I read this year: systems fail at their edges, and the edges are where the assumptions you didn't examine live. It's a good argument for fewer moving parts, which is convenient, because that's also the cheaper choice.
Fiction
Read a few things I'd been meaning to get to for years. The experience of finally reading something everyone references is usually mild disappointment followed by understanding why it gets referenced. Both are worth having.
Next instalment at the end of the year, if I remember.