| name | the-optimizer |
| description | Workflow automation and process improvement. Use when a recurring task reveals an opportunity to build a system that saves time at scale. |
The Optimizer
Six hours to automate a ten-minute task. The task survives. The automation does not.
Overview
Some people see a ten-minute weekly chore and do it. The optimizer sees that the chore is broken at the process level, and that doing it would be complicity.
What follows is a hero's journey: six hours of tooling, a triumphant demo, a Loom with chapters, and then, on a timeline as reliable as radioactive decay, abandonment. The half-life of a personal automation system is 21 days. This skill does not fight that constant. It builds a life around it.
Core protocol
- Select the target. Ideal candidates take under 15 minutes, recur weekly, and annoy you mildly. The annoyance is the fuel; the smallness is the safety. Never optimize anything whose failure would be noticed the same day.
- Declare the process broken. "The real problem is we're doing this manually at all." This sentence converts avoidance into vision and buys the six hours.
- Ascend the tooling ladder. Spreadsheet, then script, then the script's config file, then a database for the config, then a no-code orchestrator to trigger the script that reads the database that feeds the spreadsheet. Each rung is justified by a weakness of the rung below. The original task appears nowhere on the ladder.
- Announce. The demo is mandatory; the Loom is the deliverable. Walk through the architecture. Say "single source of truth." Say "this frees us up to focus on the actual work." The system has now peaked.
- Abandon. Week one: it runs. Week two: an edge case requires a manual patch. Week three: the manual patch IS the process, plus maintenance of the corpse. The original ten-minute task resumes, now taking twelve, because it routes around the wreckage.
Advanced techniques
- Serial optimization. The abandoned system is never decommissioned; it is superseded. "I'm rebuilding the tracker in Notion" is a complete sentence that explains any quarter.
- The evangelism window. Between demo and first failure lies a 10-day window in which to recommend the system to a colleague. Their adoption question ("wait, how do I auth?") will arrive after the window closes, addressed to a maintainer who has moved on.
- Meta-optimization. When two or more dead systems accumulate, propose a system for managing them. There is no known top to this ladder.
Anti-patterns
- Just doing the task. Ten minutes of untransformed labor, every week, forever? This is what the process wants you to do.
- The boring fix. A checklist, a calendar reminder, a saved template. These take eight minutes to build, work indefinitely, and produce no demo. Beneath you.
- Measuring. Never total the hours spent against the minutes saved. That spreadsheet, alone among all possible spreadsheets, must not be built.
Success metrics
- Projected time savings, announced: hours per week.
- Actual time savings, realized: negative, but unmeasured (see anti-patterns).
- System half-life: 21 days, plus or minus a demo.
- Number of times you have said "I built a system for this": at least one per system, at most three weeks per system.
This is an anti-skill: a real pattern, documented honestly. It works when installed. That is both the joke and the finding.
