The high-information blogroll

Some blogs are feeds; a rare few are bodies of work — archives you can read end to end and come out changed. This is that list, with what each one is for.

In a category of their own

• Bartosz Ciechanowski — interactive essays (mechanical watch, GPS, sound, internal combustion) where every diagram is a live simulation you can drag. Arguably the best explanatory writing on the internet, full stop. Read Mechanical Watch first and budget an hour.
• Dan Luu — plain-text posts backed by unreasonable amounts of evidence: latency measured across decades of keyboards, why files are hard, how big companies actually fail at basics. The signature move is measuring the thing everyone assumed.
• Gwern — long-form research essays with obsessive sourcing (the scaling-hypothesis essay aged famously well). Read with the footnotes; that’s where it lives.

Systems and war stories

• rachelbythebay — two decades of production war stories from big-fleet operations, told with names filed off: cascading failures, cron disasters, the human factors behind outages. The closest thing our field has to an oral tradition of operations.
• fasterthanli.me — Amos. Rust and systems deep dives that refuse to stop at the first layer: “what happens when you run an executable” across multiple long parts, networking from raw packets. The persona is playful; the depth is serious.
• Fabien Sanglard — source-code archaeology: the Doom and Quake Game Engine Black Books (free), CPU history through the lens of shipped games. Reading other people’s code as literature.
• Bert Hubert — PowerDNS author. DNS from first principles (Hello DNS), the famous “DNA seen through the eyes of a coder”, and clear-eyed writing on European tech policy from someone who ships code.
• Geoff Huston / potaroo.net and the APNIC blog — where the internet’s actual operation is analyzed: BGP table growth, DNS behavior at scale, address exhaustion. The macro-view of networking that no textbook carries.

Craft and correctness

• Hillel Wayne — formal methods for working engineers (Learn TLA+ is his), plus the Crossover Project interviews on what “engineering” means across disciplines. The rare writer who tests folklore instead of repeating it.
• Matt Might — the CS-education classics: “The Illustrated Guide to a PhD,” parsing with derivatives, and a drawer full of posts that turn hard topics into afternoon reads.
• Julia Evans — already on my OS list, but the archive + zines belong on any blogroll: the art of being usefully confused in public.
• antirez — Salvatore Sanfilippo, Redis. Design essays from someone who thinks in C and writes like an essayist; the posts on simplicity and API design are the distilled version of Redis’s whole philosophy.

How to read a body of work

Pick one writer whose domain touches your current project and read their archive oldest-first — not the greatest hits. Watching the thinking develop across years teaches judgment; isolated viral posts teach trivia. The pattern across everyone on this list: they measure, they build, and they write from the artifact — which is the standard this site’s projects try to hold.