Writer’s Brief

The Hook

I ran a code audit on my own personal task system and the reviewers all said the same thing: stop building.

Main Points to Cover

  1. The meta loop problem - when your task system becomes the biggest consumer of its own output (32% of all tasks were about improving the task system)
  2. The “operating system” realization - capture, management, execution, output layers. You’re not building a to-do app. You’re building an OS. And that changes everything about when to stop.
  3. The freeze as the hardest feature - choosing to not build anything for a week when the system is exciting and growing fast

The Angle

Personal discovery. You build something, it works, it grows, and then the growth itself becomes the problem. The system that was supposed to help you ship things becomes the thing you ship. The audit catches it. The panel of imaginary experts all independently say “freeze it.” And you do.

Target Reader

Anyone building personal tools with AI (Claude Code, Cursor, etc.) who’s hit that moment where the tooling feels more productive than the work. Solo devs, indie hackers, creative technologists.

Tone Notes

Self-aware, a little sheepish about getting caught. Not preachy - this isn’t “don’t build tools.” It’s “notice when the tool becomes the project.” The humor is in the specifics: tasks.json grew 345% in 3 days. The biggest tag was “meta.” Casey Neistat asking “what have you shipped that isn’t the system itself?”

Raw Material / Moments to Write From

These are notes for Alex to write from. Not prose.

  • tasks.json grew from 141KB to 628KB in three days. 345% growth. That’s not a task list, that’s a database.
  • 181 done tasks sitting in the active file. Like keeping every completed sticky note on your desk.
  • 32% of all tasks were meta - about improving the system itself. The system was its own biggest client.
  • Asked Casey Neistat (fictional) to review. His question: “what have you shipped that isn’t the system itself?” Ouch.
  • David Allen (fictional GTD guy): “You have seven capture inboxes. GTD has one.”
  • Donella Meadows (fictional systems thinker) identified “Loop 4: The Growth Loop” - when a task feels hard, you build infrastructure. That infrastructure generates ideas. Those ideas become tasks. No natural governor.
  • Sahil Lavingia (fictional): “I run a $10M/year business with fewer moving parts than this.”
  • The OS layers model - Layers 0-2 (infrastructure, capture, management) are healthy. Layer 3 (execution) is underused. Layer 4 (output) is getting crowded out by layers 1-2 generating work about themselves.
  • When asked “what shipped this week that someone outside your system saw?” - Alex’s answer: “I’m not actually sure because there’s no real easy way to see what is done.” Then caught himself wanting to build a view for that. During the freeze. The irony.
  • “I feel like I’m actually less ADHD with a tool like this” - MC might actually be working as intended, just growing too fast
  • Morning sweep: “I’m just ignoring that every time it comes through.” The thing he built to organize his mornings is itself getting ignored.
  • The freeze decision was unanimous across all fictional reviewers, independently. Systems thinker, GTD, two creators. All said stop.
  • None of the 5 recommendations from the first audit (Feb 24) had been implemented by the second audit (Feb 27). Too busy building new features to fix what was already flagged.
  • Archive report revealed the heaviest task had 38 thread messages (writing cohort 2 email). The lightest had 0. Some tasks are projects pretending to be tasks.
  • The meta-task governor: m key used to create tasks. Now it creates log entries. Same capture, different commitment level. “Your notebook is allowed to be messy. Your to-do list should only have things you’re actually going to do.”
  • Alex on Chief of Staff vs Mission Control: “I need to figure out how to teach it.” The tool you built for yourself vs the tool you can teach others. Different design pressures.