A weekend bicycle ride to Vũng Tàu, the road that made it worth doing, and eight photographs of walls. Notes on riding for the pleasure of it, and on why a town you arrive in slowly gives you different pictures.

Bloom filters answer one narrow question - definitely not present, or maybe present - in a fixed amount of memory, and that narrow answer is enough to save databases from millions of disk reads for keys that were never there. Here is how they work, the math behind the false positive rate, and where RocksDB, Cassandra, and PostgreSQL actually use them.

A practical walkthrough of signing and verifying container images with cosign's keyless mode: how Fulcio and Rekor replace long-lived keys, a working GitHub Actions job, and the cosign v3 migration details that trip people up.

Extended thinking is not a smartness switch. It is extra sampled tokens, billed as output, trained by reinforcement learning on verifiable answers, and it has a measurable point past which accuracy goes down.

PostgreSQL and RocksDB both answer point queries and range scans, but their on-disk layouts could not be more different. Here is how B-trees and LSM-trees actually store your data, why one trades read speed for write speed, and how to pick between them.

Free-form JSON prompting frequently breaks in production due to syntax errors and schema drift. Here is how to combine native provider schemas, client validation, and repair loops for rock-solid LLM outputs.