In a new piece for Platformer, a writer describes a follow-up to their earlier experiment in workplace automation. After spending six months attempting to automate their own job, they gave an AI model called Claude Fable 5 a larger assignment: replacing their boss, Casey. The article's title frames the question directly: how much of a boss's job can AI actually do?
The experiment moves the conversation about AI and work from individual productivity to the realm of management and editorial judgment. It asks whether a model can handle the responsibilities that typically come with overseeing others, such as making decisions, giving feedback, and setting direction. The available excerpt does not include the outcome, so the piece is best understood as a premise rather than a verdict.
Even without results, the experiment is significant for what it represents. Generative AI is increasingly being tested not just as a tool for completing tasks, but as a potential occupant of roles that involve authority and accountability. For society, that raises questions about which parts of work remain distinctly human, and whether the move to automate management will change how workplaces are structured.