Update social post and demo voiceover scripts

docs/demo-script.md

@@ -1,63 +1,43 @@
-# Demo Recording Script
+# Demo Video Voiceover Script
 
-Use this script to record a short demo video for judges. Keep it between 45 and 90 seconds if possible.
+Use this script for an AI voiceover. Target length: 75-90 seconds.
 
-## Goal
+```text
+Welcome to LifeChoice Simulator.
 
-Show the app working end to end so judges can evaluate it even if API limits, model latency, or live runtime issues affect the Space during review.
+This is an LLM-powered interactive future game for the Thousand Token Wood track.
 
-## Recording Flow
+Instead of asking an AI for advice, you enter a real fork in the road and play through one possible future.
 
-1. Open the live Space:
-   https://huggingface.co/spaces/build-small-hackathon/LifeChoice-Simulator
+Here I start with MTech versus Software Job. I choose the path I want to enter, add the pressure that makes the decision real, and select a mentor voice to follow me through the simulation.
 
-2. Start on the setup screen.
-   Use the default dilemma:
-   `MTech vs Software Job`
+The app now creates an eight-node alternate future. A 7B model generates adaptive scenarios based on my dilemma, my chosen path, and the choices I already made.
 
-3. Keep the selected path as:
-   `MTech`
+But this is not just a chat transcript. Behind the scenes, a game engine tracks five changing metrics: financial security, fulfillment, validation, stress, and family satisfaction.
 
-4. Use the calibration prompt:
-   `My family needs income soon, but a professor is already interested in my research.`
+Every choice also creates facts, obligations, and closed options. That means earlier decisions can return later as consequences.
 
-5. Select the persona:
-   `Mentor`
+The world reacts too. The background can become stable, thriving, or struggling. The character can shift between neutral, stressed, and confident.
 
-6. Click `Enter this future`.
-   Point out that the app is built around a 7B model for adaptive future scenarios, with deterministic state logic keeping the simulation grounded.
+As I commit choices, the future keeps changing. The model dreams up the next scenario, while the simulation remembers what I did.
 
-7. Show the first scenario card, source label, metric panel, persona bubble, world state, and character sprite.
+At the end, LifeChoice generates a final report from the causal ledger. It does not predict my life or tell me the correct answer. It shows how tradeoffs compounded inside this playable future.
 
-8. Pick a choice and click `Commit choice`.
-   Mention that every choice updates deterministic metrics and creates durable consequences.
+LifeChoice Simulator is part text-adventure, part alternate-life RPG, and part reflective decision game.
 
-9. Make two or three more choices.
-   Show that the world state, metrics, persona text, and later scenes respond to earlier decisions.
+Built with Gradio, Hugging Face Spaces, Qwen 2.5 7B, and Codex as the implementation partner.
+```
 
-10. Continue to the final report if time allows.
-    Show that the report is grounded in the causal ledger.
+## Visual Beat Guide
 
-## Suggested Voiceover
-
-This is LifeChoice Simulator, a causal future simulator for difficult decisions.
-
-Instead of giving one-shot advice, the app lets me enter a path and experience a short sequence of consequences. I choose MTech over a Software Job, add the real pressure behind the decision, and pick a mentor voice.
-
-The adaptive scenario experience is powered by a 7B model. The simulation engine owns the actual state around it: metrics, facts, obligations, closed options, safety rules, validation, and the final report.
-
-Every choice changes the world. The metric panel updates, the environment can move between stable, thriving, and struggling, and the character can become neutral, stressed, or confident.
-
-The important part is that earlier decisions return later. The app keeps a causal ledger, so the final report reflects what happened inside this simulation instead of pretending to predict real life.
-
-LifeChoice is built for reflection, not advice. It helps users see tradeoffs more clearly before making high-stakes decisions with real people in the loop.
-
-## Capture Checklist
-
-- Setup screen is visible.
-- First scene is visible.
-- At least one choice is committed.
-- Metrics visibly change.
-- World or character state is visible.
-- A cascade or final report is shown if time permits.
-- The video ends on the Space URL or README.
+| Voiceover Moment | Suggested Screen |
+|---|---|
+| Welcome / track intro | App landing/setup screen |
+| MTech vs Software Job | Dilemma input and path selector |
+| Mentor voice | Persona selector |
+| Eight-node future | First scenario card |
+| Five metrics | Live world-state panel |
+| Facts and obligations | Commit one or two choices |
+| World reacts | Show world badge and character sprite |
+| Final report | End-state report screen |
+| Built with stack | README or Space header |