Documentation
The wiki covers any technical subject, at the grain of one mechanism to a page. Some pages are working notes, some are opinionated guides, and some are attempts to explain a thing properly from first principles. They assume a technically astute reader with no particular exposure to the topic at hand.
Nine sections so far, ordered here as they appear in the sidebar:
- Economics — how value gets created, priced, and moved. Almost all of it is finance so far, and most of that is DeFi: the largest section of the wiki, running from what a blockchain is up through the options machinery built at bitsy.services. Payments holds the bank rails DeFi is measured against, regulation the constraint both run into — the Bank Secrecy Act and the anti-money-laundering regime that decides what any of it is allowed to do — and fraud covers what happens where the rules do not reach, one page per mechanism from rug pulls to pig butchering.
- AI — two halves. Getting useful work out of coding agents (context engineering, agentic workflows, prompt caching), and taking the models apart (neural networks, large language models).
- Security — the operational side: handling keys, secrets, and credentials on a developer machine.
- Social — building and moderating communities on social platforms.
- Computer Science — the durable ideas underneath the rest: entity addressing, DAGs, content addressing and what it costs to keep the bytes alive, zero-knowledge proofs, and a recurring interest in canonicalization bugs.
- Git — an opinionated workflow rather than a command reference: small complete commits, linear history.
- Networking — how packets get between hosts. So far that is NAT traversal: why two machines behind home routers cannot simply connect, and the protocols that get them talking anyway, from STUN and hole punching to the TURN relay of last resort, all orchestrated by ICE inside every WebRTC call. Trystero sets those calls up with no server of the application’s own, by passing the setup messages through public Nostr relays.
- Web — publishing and serving documents: how a site gets built, and how it reaches a browser. So far that is Hugo, the static site generator this wiki runs on — where URLs come from, which of the files in
layouts/renders a given page, the asset pipeline, and the internal-link checking Hugo does not do. - Microsoft — programming against Microsoft’s platforms, starting with Outlook.
Pages link to each other heavily on purpose. Following the links tends to be a better way through the material than reading a section top to bottom. A subject that fits none of the nine gets a tenth.