Software Ate the World, Skills Will Eat Work (gist.github.com)

A groundbreaking shift in AI development is underway with the emergence of "Skills," which are poised to replace traditional work approaches. Unlike unpredictable AI agents, Skills are controllable, auditable, and correctable, allowing for the creation of millions of Skills in the next 2-3 years—outpacing the launch of apps in the App Store. This paradigm will enable the concept of unmanned companies, where humans transition from being operators to mere overseers, significantly increasing efficiency as Skills self-evolve based on revenue outcomes. The ability to define every job as a series of Skills means that companies can streamline workflows by automating tasks that were once reliant on human intervention. Skills can be created quickly and automatically through AI watching human performance, making the replacement of inefficient workflows feasible. As Skills are deployed, companies will shift from a Human Loop model—where decision-making resides with humans—to an Agent Loop model, wherein agents autonomously execute tasks while humans oversee the process. This evolution in work dynamics could lead to unmanned companies outperforming their human-operated counterparts, particularly as Skills learn and optimize for better revenue generation, entering a realm where human expertise may no longer be the determining factor in operational success.