After downsizing his engineering team by 90%, Arman Hezarkhani discovered that shifting to an AI-first approach increased software output tenfold—revealing a critical flaw in hourly billing models that disincentivize efficiency. To align incentives, Tenex compensates engineers based on story points, units of completed, high-quality output, enabling top performers to earn over $1 million annually. This system encourages rapid adoption of AI tools and continuous workflow optimization. The company has delivered client projects at unprecedented speed, including a retail AI system prototyped in two weeks and a mobile app that reached the global Top 20 on the App Store. Despite advanced AI tooling, their primary bottleneck remains human capital—finding elite engineers who thrive in output-driven environments. Their hiring process features 'unreasonably difficult' take-home challenges and a focus on long-term thinking. A key insight: as AI agents evolve, controlling entropy—the accumulation of small errors—is emerging as the true challenge in autonomous development.