enshAIttification — Agentic Web Platformisation Simulation

A Streamlit web application that runs a parametrized logistic choice model of adoption dynamics for the agentic web. The model contrasts open / commons-based signal processing (institutions loyal to members, e.g. data unions / FIDU-like stewards) with platformised, proprietary signal processing.

Quick start (local)

cd agentic-web-sim
python -m venv .venv
source .venv/bin/activate
pip install -r requirements.txt
streamlit run app/main.py --server.port=8501

Open http://localhost:8501

Docker

cd agentic-web-sim
docker compose up --build

The app listens on 8501. The compose file mounts ./app into the container for development convenience.

Project layout

• app/main.py — Streamlit UI
• app/model.py — Simulation loop (96 monthly steps over 8 years; calendar t in years)
• app/dynamics.py — State variables (Q, F, L, E, A, V, etc.)
• app/consumer_types.py — Consumer segments, utilities, shifted Bass arrivals
• app/scenarios.py — Named presets
• app/visualisation.py — Plotly figures
• .streamlit/config.toml — Theme and client defaults

Model summary

• Horizon: 8 years at monthly resolution (calendar time in equations is still years). The short horizon is intentional: agentic adoption moves fast; commons institutions move slowly—so k_F and early N_open matter under time pressure.
• Four consumer types (editable in the sidebar): mix weights, Bass p/q, peak timing, and utility α–ζ per segment; defaults match the original spec.
• Key lever: k_F — speed of commons institutional maturity F(t) in commons signal quality.
• Feedback: platform base quality, network effects, lock-in after dominance, enshittification past a share threshold (with a competitive brake when open adoption stays viable).

Licence

Use and modify for research and commentary; no warranty.