Built to be read by machines.

I write about pointing AI at the right things. It would be odd if this site were not built for that too.

So it is. Every page is clean, server-rendered HTML with structured data. There is an llms.txt index, and every white paper has a plain-Markdown twin you can hand straight to a model.

The machine-readable parts

/llms.txt A curated index of everything here, for agents.
/graph.json Jake as a typed graph: nodes, edges, and how they connect.
/sitemap-index.xml Every page on the site.
/papers/abundance-or-scarcity.md Abundance or Scarcity, as clean Markdown.
/papers/compute-then-narrate.md Compute, Then Narrate, as clean Markdown.
/papers/from-dashboards-to-decisions.md From Dashboards to Decisions, as clean Markdown.

Try it with your own AI

Copy a prompt, paste it into your assistant, and let it read.

Read https://jakeherridge.com/llms.txt and tell me who Jake Herridge is, what he has built, and what he writes about.
Read https://jakeherridge.com/papers/from-dashboards-to-decisions.md and explain its main argument in plain terms, then tell me who should read it.
Read the white papers indexed at https://jakeherridge.com/llms.txt and tell me the through-line across all of them.
Read https://jakeherridge.com/graph.json and describe how Jake's values connect to the things he has built. Follow the typed edges.