The ‘app era’ has shifted the burden of connecting disparate information and tasks to individuals, a hidden form of labor that goes unlogged. AI could alleviate this by managing the complex interdependencies between recurring jobs, rather than acting as a single, autonomous agent. This approach involves building ‘loops of loops’ – specialized, memory-enabled agents that can recognize and act upon the connections between different tasks.

• Apps have made individual tasks easy but left the crucial work of connecting them to humans.
• The ‘wiring’ between tasks—remembering, checking, following up—is often overlooked as work.
• AI should help manage these connections, not by pretending life is simple, but by handling recurring obligations.
• A ‘loop of loops’ system involves narrow, specialized agents with memory and boundaries that can notice when one task affects another.
• This system helps manage dependencies that previously lived only in one’s head, like how weather impacts school trip packing.
• The goal is not to automate life but to recognize and structure the work of connection that apps have obscured.