I Built a Survival Sim for AGI Collapse. 80% Fail. Need Your Feedback.

I’m not an AI researcher — I’m a builder.

After watching Yudkowsky, Bostrom, and Bengio, I kept thinking:  
**"People won’t act until they *feel* the risk."**

So I built **Survive99** — a decision-based sim where you face 5 real AGI collapse scenarios (job takeover, family triage, policy vote, etc.).

**80% of players fail the job category.**  
**No one passes all 5.**

It’s not a game. It’s a **stress test** for the 1% survival odds.

**Play it here (5 free, no signup):**  
[survive99.com](https://survive99.com)

---

**What I need from you (alignment folks):**

1. **Is this the right message?** Too doom? Not doom enough?  
2. **Are the scenarios accurate?** (Based on public warnings — but I’m not in the labs.)  
3. **What’s missing?** Should I add:  
   - Deception detection?  
   - Lab leak response?  
   - Pause advocacy tools?  
4. **Pass → verified badge in AI safety Discords** — does this help filter signal from noise?

This isn’t about me. It’s about **getting the alarm to the people who can still pull the brake.**

Feedback welcome. Let’s make it right.