With agent loops automating execution, the highest-value human skill becomes designing the environment and rules for the AI. This involves writing the strategy document (like 'program.md'), defining success metrics, and constructing the evaluation function. Your job is no longer to do the work, but to architect the system in which the work gets done.

As AI agents take over task execution, the primary role of human knowledge workers evolves. Instead of being the "doers," humans become the "architects" who design, model, and orchestrate the workflows that both human and AI teammates follow. This places a premium on systems thinking and process design skills.

The most advanced use of AI has evolved beyond manual prompting. The new discipline involves designing automated, recurring systems called 'loops' that manage agents to perform complex work. This represents a fundamental shift in knowledge work from directly performing tasks to architecting and overseeing autonomous systems that execute those tasks.

The new paradigm for knowledge workers isn't about using AI as a tool, but as a team of digital employees. The worker's role evolves into that of a manager, assigning tasks and reviewing the output of autonomous AI agents, similar to managing freelancers.

Instead of repeatedly performing tasks, knowledge workers will train AI agents by creating "evals"—data sets that teach the AI how to handle specific workflows. This fundamental shift means the economy will transition from paying for human execution to paying for human training data.

Knowledge work will shift from performing repetitive tasks to teaching AI agents how to do them. Workers will identify agent mistakes and turn them into reinforcement learning (RL) environments, creating a high-leverage, fixed-cost asset similar to software.

The future of knowledge work isn't about humans performing tasks, but about training an AI agent to perform them once. This is a structurally more efficient model, amortizing the initial training effort over the agent's entire lifecycle, which will create a new job category centered on agent management and training.

Iterative AI agent loops, like Andre Karpathy's Auto Research, are not just another tool but a new foundational building block of work. Similar to how spreadsheets or email became ubiquitous across all roles and industries, these loops will be a core component of how knowledge work is performed, fundamentally changing process and productivity.

The most significant shift in knowledge work is the new ability for non-engineers to build and use code via AI. Professionals can now create software-based solutions, like automated analytics dashboards, fundamentally changing their job scope from performing tasks to building engines that perform tasks.