Library

Summary

In spring of 2026, I sent off my entire genomics and developmental biology library to be digitized — both because I wanted an easier way to review my reference stack myself, and because I wanted to make the entire corpus available to an agentic assistant via RAG.

Textbooks can age quickly, so papers soon made their way into the library as well, and an Obsidian-like concept graph followed in short order.

Somewhere in there came a reminder of the replication crisis. Biology isn’t quite so rife with it — I think — as the social sciences, but it’s hardly immune.

The answer was trivially easy, and borrowed from the web-of-trust functionality I’d already built into the tutor/assistant. Each paper’s citations were already tracked; adding a reputation cascade to that citation history — one that updates paper, researcher, and lab trustworthiness based on past behaviour — was a comparatively small lift.

The application itself still has some rough edges before it’s ready for public release, but expect a skeleton package to hit GitHub late 2026.

Working notes

• 2026 · 08

  Reputation cascade in place.

  Citation chains already tracked; the trust cascade rides on top of them, updating paper, researcher, and lab standing from past behaviour.