TL;DR
- 01Inheritance is what makes a code library reusable. It is also what makes an institution survive the people who staff it.
- 02Institutional knowledge work fails at handoff. Generic AI stores text. The knowledge layer stores typed reasoning.
- 03The kernel is fixed across cases; per-case extensions are logged, named, and reviewable. Provenance is mandatory.
Institutional reasoning fails at handoff. Generic AI stores text; the knowledge layer stores typed reasoning that survives the analyst. The kernel is fixed, the extensions are logged, the provenance is mandatory.
The handoff failure mode
The analyst rotates; the desk officer moves on; the mediator retires. What gets inherited today is a folder of PDFs and a few decks. What needed to be inherited is the reasoning — which actor owns which interest, which commitment is still active, which event triggered which policy shift, which narrative stopped being credible last March.
Why generic AI does not solve this
- ▸LLMs store text and generate text; they do not maintain typed structure across sessions.
- ▸Self-generated context accumulates and accuracy decays in ways that look fluent and read confident (Context-Bench, 2025).
- ▸A chat transcript is not a persistence layer; pretending otherwise pays the inheritance bill twice.
What the TACITUS layer does
The kernel is fixed across cases. Per-case extensions are logged, named, and reviewable. Every claim has a source span. Every commitment is bi-temporal. The graph survives the analyst because the analyst was never the storage layer in the first place — the typed graph was.
SOURCES
