Notebook Entry 021
Pre-Marination Loops — Start Research Months Before You Need the Output
May 2, 2026
Can an agent start being useful before the moment it's actually needed?
What happened
While discussing a future business consultant agent, Amanda proposed running Ralph Wiggum Loops (P-16) via scheduled tasks NOW — before the agent is built — to accumulate synthesized business intelligence. The insight expanded: this isn't just about one agent. It's a general pattern for any capability you'll need in the future. Start the research/synthesis loop early, let it marinate, and by activation time the knowledge base is already expert-level.
Root cause
Most AI implementations cold-start their agents — build the agent, then start feeding it data. The time-to-useful is measured in "how fast can we load context?" But context loaded all at once is shallow. Context accumulated over time with iterative refinement is deep.
What changed
"Pre-marination loops" added as a methodology pattern. Design sketch captured: weekly scheduled task, reads latest content, produces one brief that builds on the previous cycle's output. Each cycle's output becomes next cycle's input (true Ralph Wiggum Loop). Dependencies identified for implementation.
The lesson
The competitive advantage isn't just having AI agents — it's having agents that have been THINKING for weeks before you need them. Start synthesis loops early based on rough ideas of future needs. Your competitors will cold-start. Your agents will arrive pre-marinated. This is a real consulting differentiator and a signature methodology pattern.
One implementation insight every other Tuesday.
What actually works when you ship AI in real organizations. Nothing you could get from a press release.