The Human Paperclip Maximizer: Goodhart's Law, Financial Optimization, and the Preconditions for AI Alignment
Connecting Goodhart's law and the paperclip maximizer to financial optimization, human flourishing, and the purposes that AI should serve.
Author: Hugo Latapie, Taijitu AI Inc.
Preprint | Version 10 | Published November 6, 2026
Abstract
Imagine an AI that makes so many paperclips it consumes resources needed for human life. Its measure of success has displaced the purpose that made it valuable. This failure shares an optimization structure with Goodhart's law, which arose from human monetary policy. The connection turns the question toward civilization itself. Markets became powerful by responding to dispersed knowledge and changing reality. Yet financial success can increasingly depend on accumulating claims over future production rather than expanding human possibilities. We examine that hypothesis through financial claims, asset revaluation, and housing access. The Declaration of Independence supplies a different account of purpose: life, liberty, and the pursuit of happiness. Enterprises can preserve market discovery while organizing first to avoid harm, then to create human benefit, and finally to earn returns that sustain their work. Perhaps solving the paperclip problem for machines begins with solving it for ourselves.
License
Creative Commons Attribution 4.0 International (CC BY 4.0).
Copyright © 2026 Taijitu AI Inc.