Systems Playbook
Durable systems judgment for the AI era
Browse the playbook ↗As AI makes implementation faster, the scarce work shifts to judgment: choosing boundaries, abstractions, constraints, tradeoffs, and goals. Systems Playbook compiles exactly that, the durable engineering wisdom from influential books, so you have one place to check a vetted idea when you’re making a real decision instead of re-reading a shelf or trusting a half-remembered takeaway.
The unit of knowledge is the concept, not the book. Books are sources; when a new one repeats an idea the graph already has, it enriches the existing entry instead of adding a summary. The graph currently holds 137 concepts distilled from 19 books, with each source kept as a reference summary for provenance.
Ideas are wired together, not filed away: full-text search, category pages, and curated reading paths that walk a cluster in order, from a start-here path through deeper routes on software design, failure, and decision making.
Everything is written to be read by coding agents as well as people. The site serves an llms.txt and the entire graph as a single JSON file, so the same judgment that guides you can steer the tools building alongside you.
Concept
A single idea you can define on its own, like Deep Modules. The atomic unit of the graph; every book gets distilled into these.
Pattern
Two or more concepts combined for a recurring situation, like Modular Boundaries. Promoted only once the combination shows up across sources.
Mental model
A stance about how things work, like Complexity Should Be Hidden, reinforced across many books rather than argued by one.