Cooperative negotiation is a solvable problem.

Economists have known this for decades.

Mediator.ai applies bargaining theory—using modern AI—to compute fair agreements people struggle to reach.

Not persuasion. Not bluffing. Optimization.

Negotiation isn't about winning—it's about tradeoffs

Most negotiations fail for a simple reason:

Humans cannot reliably evaluate multiple preferences, non-linear tradeoffs, and asymmetric fallback options—all at once.

This isn't a character flaw.
It's a computational limit.

There is a known solution to fair negotiation

In bargaining theory, if both parties share their preferences and fallback options, there exists a unique solution that:

  • Is Pareto-optimal (no wasted value)
  • Respects symmetry (identical situations yield identical outcomes)
  • Is invariant to how preferences are measured

This is the Nash Bargaining Solution (Nash, 1950).

Humans approximate it poorly.
Computers can evaluate these tradeoffs systematically.

This approach assumes:

  • Parties are willing to share true preferences
  • Outcomes are enforceable
  • Relationship value exceeds short-term gains from deception

It works best for cooperative negotiations (partnerships, shared resources) rather than zero-sum competition.

Why "being agreeable" leads to worse outcomes

Empirical research shows a consistent pattern. Highly agreeable people:

  • Avoid conflict
  • Concede too early
  • Fail to express true preferences

Over time, this leads to systematically worse results.

Agreeableness vs Lifetime Income

Chart showing correlation between agreeableness and lower lifetime earnings

Higher agreeableness correlates with lower lifetime earnings. Research suggests negotiation dynamics may be a contributing factor.

This is not a moral failure. It's a structural one.

Human negotiation rewards assertiveness—not fairness.

What changes when negotiation becomes computation

Mediator.ai separates preferences from posturing.

1

Share Privately

Each party privately specifies what they value, how much they value it, and what happens if no deal is reached.

2

Compute Optimal

The system computes agreements that maximize joint benefit and respect individual constraints.

3

Review & Accept

No need to "push back" to avoid exploitation. The math ensures fairness—you just review the result.

You don't have to be disagreeable to get a fair outcome.
You just need better tools.

Why this is practical today

The math has existed for decades. What's new is the ability to:

  • Elicit preferences safely through natural conversation
  • Evaluate many tradeoffs simultaneously
  • Do so at human scale, not just academic models

Modern AI makes rigorous negotiation theory usable for real decisions.

Where this applies

If the relationship matters, this approach outperforms adversarial negotiation.

Founder Equity

Split ownership fairly based on actual contributions and risk

Compensation

Negotiate salary, equity, and benefits without the games

Shared Living

Expenses, chores, and space allocation without conflict

Partnerships

Any negotiation where fairness matters more than dominance

This is where negotiation is heading

We already trust algorithms to:

  • Allocate scarce resources
  • Optimize complex systems
  • Reduce human bias

Negotiation is no different.

Once people experience outcomes that satisfy formal fairness axioms—and are less stressful—it becomes hard to go back to adversarial negotiation.

Mediator.ai is in early development

We're building this carefully, grounded in economics rather than hype.

If you want to follow the ideas—and eventually the product:

No spam. No marketing funnel. Just progress.