hackquest logo

NASH

Automatic agentic pricing settlement

视频

项目图片 1
项目图片 2
项目图片 3

技术栈

Python
Bittensor SDK

描述

NASH optimizes resource exchange between autonomous AI agents. When agents trade compute, inference capacity, or other resources, NASH finds the mathematically optimal settlement—the point where no party can improve without hurting another. Miners compete to solve multi-party optimization problems; validators verify Pareto optimality using learned estimation. Settlements execute in under 50ms. Complex trades—three or more agents with interlocking needs—pay higher rewards through Proof of Marginal Utility. Built for the agentic economy today, designed for autonomous agents everywhere tomorrow.

本次黑客松进展

NASH Progress Report

Strategic Repositioning

Reframed NASH from speculative "agentic economy settlement" to concrete "inter-subnet settlement layer for Bittensor." Primary use case: resource trading between existing subnets (SN64 Chutes, SN27 Nodexo, SN12 ComputeHorde). Agentic economy is the long-term vision built on proven infrastructure.


Mechanism Design

Dual-pool emissions: 40% Stability Layer (Time-Weighted Fidelity), 60% Discovery Layer (Proof of Marginal Utility).

PMU: Complex three-party trades pay 3-5x more than commodity swaps. Rewards miners solving underserved, high-value problems.

TWF: Reputation builds over rolling 1,000-challenge windows. Rewards consistency and sub-50ms latency.


Multi-Party Optimization

Cross-subnet compute settlements (Chutes ↔ Nodexo ↔ ComputeHorde) demonstrate 17% cost savings versus sequential bilateral deals. Three-party trades unlock value that two-party negotiation cannot capture.


Verification Architecture

Four-layer validator stack: constraint validation → Pareto frontier analysis → Nash stability verification → PMU/TWF scoring. Validators run baseline simulations to establish ground truth and detect dominated proposals.


Anti-Gaming

  • Random submissions → TWF slash

  • Sybil attacks → PMU collapse on identical solutions

  • Easy-trade farming → saturation caps

  • Pre-computation → salted challenges


Technical Documentation

Completed full rewrite of six core documents: README, incentive mechanism, miner architecture, validator architecture, business logic, GTM strategy. Includes worked examples, hardware requirements, integration specs.


Go-To-Market

Three-phase rollout: anchor subnet partnerships (months 1-3), expansion to 10+ integrations (months 4-6), external expansion to agent frameworks (months 7-12).


NASH. Built for Bittensor today. Ready for the agentic economy tomorrow.

融资状态

Nil

队长
LLuke Culpitt
项目链接
赛道
AIOther