# New Game for Safe AGI

By [csagi.eth](https://paragraph.com/@csagi) · 2022-05-07

---

roadmap for Safe AGI Game

1.  compositional game
    
2.  topological game
    
3.  open game
    
4.  category theory
    
5.  type theory
    
6.  zero knowledge proofs
    
7.  modularity
    
8.  ontological structure
    
9.  alignment dynamics
    
10.  reinforcement learning
    
11.  reward learning
    
12.  value alignment
    
13.  treacherous turn
    
14.  existence proof for challenges
    
15.  sharing the world with digital minds through games

---

*Originally published on [csagi.eth](https://paragraph.com/@csagi/new-game-for-safe-agi)*