Mediator.ai vs Sup AI

Mediator.ai applies LLMs and Nash bargaining theory to real-world disputes, generating agreements that both parties would accept — including solutions neither side had imagined independently. The process is private by design: each party separately describes their position, priorities, and constraints. The AI then generates multiple candidate agreements, scores each one against both parties' stated needs, and iteratively refines proposals until reaching an optimal solution. Use cases range from founder equity disputes and contractor payment conflicts to shared housing arrangements and inheritance disagreements. The system's key insight is that human negotiation is systematically bad at identifying the entire solution space — we anchor on positions, not interests. By modeling both parties' utility functions simultaneously, the AI can find Pareto-optimal outcomes that pure adversarial negotiation often misses entirely. With 159 Hacker News points, the response was genuinely enthusiastic — and the concept is hard to dismiss. Nash bargaining as a formalism has decades of academic credibility; what's new is making it accessible via natural language input. The pricing isn't published yet and the team is small, but the application domain (legal, HR, personal disputes) is enormous if they can nail trust and confidentiality.

Sup AI is an ensemble AI assistant that runs your query through 339 language models simultaneously, measures per-segment confidence across all responses, and synthesizes a final answer that amplifies agreement and suppresses likely hallucinations. The team claims a 52.15% score on Humanity's Last Exam (HLE) — 7.41 percentage points above the single best model — which, if verified, would make it the highest-scoring system on the benchmark to date. The underlying mechanism works like an LLM panel: each model votes on sub-claims within the response, confidence is estimated by agreement density, and the final output surfaces high-confidence segments while flagging uncertain ones. It's designed to reduce hallucination rate on factual tasks, not improve reasoning per se — the models in the ensemble aren't doing collaborative chain-of-thought, they're voting on outputs. Sup AI was built by Ken Mueller (Stanford, CEO) and Scott Mueller (AI Research Scientist) and launched on Product Hunt today. Pricing starts with $10 in free credits, no auto-charge, with a credit card required to start. The HLE benchmark claim is the headline and will face scrutiny — if verified, this is a meaningful research result. If it's cherry-picked, it's still a usable product with a differentiated architecture.