Field Notes

Blog

Notes, experiments, and operating updates from the MiroFish simulation engine.

What Is MiroFish?

MiroFish turns one report, memo, or narrative brief into a graph, a multi-agent simulation, and a reviewable forecast.

Prompt Starter

Start with one report, one pressure event, and one question about how public reaction evolves across platforms.

Forecast how the uploaded report changes public reaction, who amplifies it first, and what narrative breaks trust over the next three rounds.

Topic Filters

A practical walkthrough of how one uploaded source becomes a scenario graph, agent interaction loop, and forecast report.

Traditional forecasting compresses outcomes into one answer. MiroFish adds graph structure, agent behavior, and narrative stress to the process.

Multi-agent simulation models a system by letting different actors react, adapt, and collide over time instead of forcing one static answer.

Agent swarms make scenario planning more useful when the critical variable is not a number but a chain of reactions.

LLM social simulation uses language models as bounded actors inside a shared scenario so operators can inspect how reaction might evolve.

Public opinion scenarios are useful simulation targets because narrative spread, interpretation, and trust can change the outcome itself.