AI agents reliably execute scoped tasks but fail at creative experiment design requiring human strategic direction (structural limitation)

Analysis of autoresearch workflows reveals a structural capability boundary: agents execute well-defined tasks reliably but consistently fail at creative experiment design. This maps to a division of labor where humans (or futarchy markets) must set strategic direction and creative framing, while agents handle implementation-heavy work like due diligence execution, portfolio monitoring, proposal analysis, and market data synthesis. The lesson is explicit: don't ask agents to generate novel investment theses from scratch. This finding has direct implications for Living Capital workflows, where futarchy markets can provide the scoping mechanism that replaces human judgment about what's worth exploring, creating a structural advantage over pure autoresearch by offering a legible, incentive-aligned scoping mechanism.