I predict organizations of agents will outperform pure task-based routers on price/performance.

Ethan Mollick predicts that multi-agent organizational structures will deliver better cost-to-performance ratios compared to simple task-based routing systems for AI applications. This suggests that more complex agent coordination and hierarchical arrangements may be more efficient than direct task assignment methods, despite potentially higher architectural complexity.

Source: Ethan Mollick (X) | 2026-07-01