ARCHIVE / ABOUT
ABOUT
THE ARCHIVE.
This is a public working memory for software, artificial intelligence and the engineering of systems that have to behave under uncertainty.
D/
Dean
Software / AI / Systems
I use this site to keep ideas from evaporating after a project ends. It is deliberately closer to a laboratory notebook than a personal brand page.
Working philosophy
Knowledge should compound.
I study how intelligent systems are built, how they fail, and how they become useful when connected to real software. The archive exists to turn reading, implementation experience and failed experiments into reusable mental models.
Some entries are polished. Many are not. What matters is whether a future version of me can reconstruct the reasoning, find the evidence and see what changed.
01
Write for retrieval.A note should be findable by problem, not only by date or title.
02
Preserve disagreement.Old conclusions stay visible so revisions have context.
03
Link concepts to experiments.Ideas become durable when they are attached to observations and failure modes.
04
Prefer systems over prompts.Reliable behavior usually comes from architecture, state and feedback — not one clever string.
Archive timeline
Recent checkpoints
SELECTED HISTORY
2026
AUG 25
NOTE
Context Engineering Is Becoming Infrastructure.
AUG 19
CONCEPT
Agent Memory map revised around working / episodic / semantic state.
AUG 11
LAB
Tool retry experiment reaches 121 failure scenarios.
2025
NOV 02
SYSTEM
Archive migration to linked content collections and structured metadata.
MAR 14
MILESTONE