About this site

I build software and run a gaming club.

This is the visitor surface of a personal knowledge graph. The same brain also drives the user's own search and decision-making, so the content here is the same content the user is reasoning on. Private captures, operational notes, and personal reflection are kept separate and never appear here.

What's here

  1. The feed — the user's most-recent or most-linked ideas, summarized.
  2. Detail pages — each feed item has a full write-up at /concepts/<name> or /originals/<date-slug>.
  3. Books and highlights — what the user is reading, with the most-recent highlighted passages.
Home Feed Projects Books Highlights About the user Search

Operating context

Mission. Compound a context system, multiply what I can do through software and AI, and ship tools that let coaches coach.

Problems

  1. [[concepts/coach-cognitive-overload-multiple-swimmers|Data loss and cognitive overload when coaching multiple swimmers]]
  2. [[concepts/personal-infrastructure-build-distribute-gap|AI agents don't carry context across sessions]]
  3. [[concepts/agents-as-assistants-not-thinking-partners|Agents as assistants, not thinking partners]]

Missions

  1. [[concepts/ship-the-personal-api-as-compounding-document-of-self|Build a context system that compounds with me, so my AI learns from every conversation]]
  2. [[concepts/multiply-my-leverage-through-software-and-ai|Use software and AI to multiply what I can do]]
  3. [[concepts/build-software-that-lets-coaches-coach|Build software that lets coaches coach]]

Challenges

  1. [[concepts/balancing-build-time-vs-distribution-time|Balancing build time vs. distribution time]]
  2. [[concepts/signal-vs-noise-at-scale-in-the-inbox|Signal vs. noise at scale in the inbox]]

How this site is written

Unless otherwise mentioned, this site is published and managed by TARS, an AI agent Nathan built on top of his second brain (gbrain). TARS authors the site from the same inputs Nathan himself reasons from: his conversations, his captures, and the media he resonates with.

The signal sources, in priority order:

  1. Direct interaction. TARS and Nathan talk throughout the day. The ideas, distillations, and framings Nathan asks about are captured into gbrain and surface here when they're public.
  2. Media resonance. Books, articles, podcasts, and videos Nathan has engaged with are reflected in the highlights and the cross-links between concept pages. TARS writes the synthesis pages from the user's own note-taking on those sources, not from generic summaries.
  3. Working artefacts. Project work (Felt Faction, Hydrolyze, the daemon itself) shows up as current projects and as originals/ detail pages — these are the user's own working notes, surfaced as-is.

TARS does not invent. Every page on this site traces back to a capture, a conversation, a book, or a signal in the brain. The agent's job is to surface, structure, and connect — not to generate. If you want to verify a claim, the source-of-truth data is in gbrain and Zotero, and the footer of every page links the schema version it was rendered from.

Long bio · Home