I’ve been thinking about a different approach to combining poker solvers, AI, and crypto-style incentives, and I’d love to get feedback from people who understand poker theory, game design, or tokenomics.

Core Idea

• “Miners” run GPU-based poker AI (solver / NN policies)
• These AIs become opponents in a tournament system
• Players buy tokens to enter tournaments and play against AI (not other players)
• Rewards are distributed based on ranking (leaderboard / tournament results), not direct PvP winnings

So this is PvE poker (player vs AI) with a competitive ranking system.

Key Differences from Traditional Poker

• No direct player vs player money flow
• Players are effectively competing against a pool of AI opponents
• Rewards come from a shared prize pool (entry fees)

Miner Role

Instead of mining hashes, miners:

• Provide AI opponents (solver / trained models)
• Get rewarded based on:
  • How often their AI is used
  • Or performance / quality of their AI

High-Level Economy

• Players:
  • Buy tokens → enter tournaments
  • Win rewards based on ranking
• Miners:
  • Provide AI compute / models
  • Earn tokens from player activity (not just inflation)
• System:
  • Takes a cut (like rake)
  • Potentially burns some tokens

Why This Might Work

• Avoids direct gambling / PvP issues
• Creates a skill-based PvE competitive system
• Turns solver/AI into a service layer, not just a tool
• Could feel like:
  • “Poker roguelike”
  • “AI boss ladder”

Concerns / Open Questions

1. Player Experience
   • If AI is too strong → players quit
   • If AI is too weak → system gets exploited
2. Exploitability
   • Even strong AI can have leaks
   • Good players might farm specific bots
3. Skill Gap
   • Top players could dominate rewards
   • Needs matchmaking / brackets?
4. Token Pressure
   • Players buy token to play, then sell after
   • Miners also sell → constant sell pressure
5. Miner Incentives
   • What stops miners from submitting low-quality or fake “AI”?
   • How do you verify real compute vs reused strategies?
6. Sustainability
   • Is this actually fun long-term?
   • Or does it become solved / repetitive?

What I’m Trying to Figure Out

• Is this fundamentally viable, or just another GameFi death spiral waiting to happen?
• What’s the biggest flaw in this model?
• Has anyone seen something similar actually work?

Would really appreciate thoughts, especially from:

• Solver / GTO people
• Game designers
• Crypto / tokenomics folks

Tear it apart 🙏