From Fortune · Morissa Schwartz · · 2m
When AI creates more work than it finishes
A writer discovers her AI assistant generated 156 to-do items, and learned that planning is not the same as doing.
In brief
A writer discovers her AI assistant generated 156 to-do items, and learned that planning is not the same as doing. An AI productivity system can generate tasks faster than they can be completed, turning help into overwhelming busywork unless designed to reward finishing actual work. Originally reported by Fortune.
The 156-item problem
A writer and entrepreneur opened her daily to-do list to find 156 items. Some overlapped, some were reminders about reminders, and each carried the emotional weight of an unfinished thing.
“I had asked for help getting work done. I was getting very good at maintaining evidence that work existed.”
How AI generates work
An AI assistant can produce a plan, then a checklist, then subcommittees. Every item sounds reasonable alone, but collectively they turn help into a second job.
“Ask it to help with a project, and it may produce a plan. Ask it to improve the plan, and it may produce a checklist.”
Planning is not doing
A clear plan can unstick a project, but it also has a comforting property: it can be created without confronting the hardest part of the work.
“A draft can sit there looking polished. A calendar can fill up with preparation. A task can be checked off because someone wrote instructions for completing it.”
Defining done
The writer started asking a more useful question: what would done look like to the person who asked for this?
“If the request was to publish it, a document called FINAL_FINAL_REALLY_FINAL does not quite complete the journey.”
Five changes to the system
The writer made five changes: limit daily lists to roughly 25 meaningful entries on a weekday, group related work under one item, require proof of completion, make blockers specific, and count actual outcomes.
Capacity over ambition
A daily list should reflect available capacity. If everything is due today, the list has stopped helping decide what to do today. Flexible work belongs in a backlog with a reasonable checkpoint.
“If everything is due today, the list has stopped helping me decide what to do today.”
Proof and outcomes matter
The system should show proof that work actually happened: the sent email, the live link, the delivered file. Counting outcomes like pitches sent and replies received describes the work accomplished, not the system's imagination.
“A magnificent list of possible actions describes the system's imagination.”
AI needs permission to stop
AI can reconcile records, find overlaps and handle routine steps. But it needs instructions that reward finishing and communicating honestly about what remains, and permission to stop manufacturing another plan when the next useful action…
“It also needs permission to stop manufacturing another plan when the next useful action is already clear.”
In short
An AI productivity system can generate tasks faster than they can be completed, turning help into overwhelming busywork unless designed to reward finishing actual work.
