Show HN: Mediator.ai – Using Nash bargaining and LLMs to systematize fairness
What it is
Nash bargaining is a mathematical framework that finds the single 'fairest' outcome when two parties negotiate. Think of it as plotting all possible deals on a graph, then using geometry to find the one point where neither party can gain without the other losing more. The catch: each person needs to quantify exactly how much they value each option—historically impossible. LLMs now do this translation.
Why it matters
Most mediation feels like splitting the difference or following conventions. This systematizes fairness using proven math instead of mediator intuition. If you're negotiating anything with clear tradeoffs—prenups, divorces, business partnerships—this approach removes the 'who has the better lawyer' dynamic. You describe what matters in plain English; the system does the math.
Key details
- •Founder was inspired by his own prenup negotiation 8 years ago—felt arbitrary despite hiring a mediator
- •Nash bargaining solution (1950s) finds provably optimal splits but required utility functions no one could create
- •LLMs convert natural language preferences into mathematical utility functions automatically
- •Currently deployed for prenups and divorce negotiations at mediator.ai
- •The approach surfaces which tradeoffs actually maximize joint welfare, not just 50/50 splits