Encyclopedia of things that shouldn't work but somehow do — mass point geometry, zero-knowledge proofs, Bloom filters, Rowhammer, GPS's relativistic clock correction, the Y combinator. Each entry leads with the holy-shit hook, then deep-dives the mechanism. Wonder preserved, not flattened. Plus the ambient magic we forgot was magic.

A book about building software for an audience of one. How to spot the tool you need that nobody else has, scope it tight, ship it with AI, and not be precious about the result.

A fun way to learn a bad thing.

Code points, code units, bytes. UTF-8 vs UTF-16 vs UTF-32. Normalization, collation, grapheme clusters, and why "é".length lies. A patient tour of the text encoding system every programmer has been bitten by and few actually understand. Practical, honest, no hand-waving.

Blockchain isn't the only way to reach consensus. A tour of the BFT algorithms that power databases, distributed systems, and permissioned networks — Paxos, Raft, PBFT, HotStuff, Viewstamped Replication. Real tradeoffs, real math, no hype.

Proof of Work vs Proof of Stake — real tradeoffs, real math, no hype. A patient engineer's tour of how blockchains actually work.

Emacs for Goodness' Sake

How to use AI as a cognitive lever for genuinely novel thinking — breaking out of mental ruts, exploring alien perspectives, and reaching creative territory your brain can't access alone. All the mind-expansion, none of the acid.

A practical guide to leveraging AI for information triage without falling into filter bubbles. Covers attention economics, AI-assisted relevance filtering, the dangers of non-adversarial algorithmic curation, and the underrated art of strategic selective ignorance.

Paxos, Raft, PBFT, Viewstamped Replication, and more — compared and contrasted with brutal honesty. Covers failure models, FLP impossibility, classic and modern consensus protocols, Byzantine fault tolerance, and when to use what. Stop copying what Kafka does and actually understand the tradeoffs.