LLM to GTD
November 11, 2025
72 hours from "never heard of it" to production.
In last week's newsletter, I launched Decision Analysis. The response? Crickets.
I think that you've gotten the wrong impression about me. I think that in all fairness, I should explain to you exactly what it is that I do.
For instance, here's what happened this weekend when I performed a Decision Analysis on myself.
Friday, November 7th 5:00-5:30pm: Turns out I need project management. Did automated research on project management platforms. Then asked: "Which one should I use?"
Criteria: open-source, API-first, Kanban support, self-hostable.
Answer: Vikunja. (Personal plan: €4/month, 2-week free trial).
[Dinner]
7:30-8:45pm: Asked: "How do I build an API wrapper for Vikunja?"
Used my existing spawn-experiments research into AI coding methodologies. Several iterations later: working Python wrapper with automatic creation of projects, tasks, labels, views, buckets, plus export script.
Round-trip automation complete.
[Movie]
11:30pm-12:15am: Tests, fixes, documentation. Populated metadata with GTD framework.
Saturday, November 8th 12:30-1pm: Project hierarchy and kanban lifecycles.
Sunday, November 9th 6-7:30pm: Recurring tasks, habit tracking, command-line tools.
10:30pm-midnight: Competitive research on GTD and AI-powered productivity tools.
Monday, November 10th 9am-1pm: Refinements. Added new domains: coding sprints, writing projects, event prep.
Tuesday, November 11th 7:30-10am: Natural language interface. Speaking opportunities tracker. Content pipeline system.
10am-11:30am: Wrote this newsletter.
What I'm Tracking Here's what's in the system right now:
Upcoming events - Speaking opportunities at conferences and meetups. Proposal deadlines. Talk preparation. Travel planning. Content pipeline - Newsletter ideas. Drafts in progress. Blog posts planned. Social media scheduled. Proposed spending - Tools and services to evaluate. Infrastructure decisions. Q1 2026 budget allocation. Development sprints - QRCards feature roadmap. Decision Analysis product features. Research queue priorities. API wrapper enhancements. Applications portfolio - 17 active applications tracked. Tasks automatically organized by project hierarchy. Source documents linked. Labels for categorization.
Everything automatically goes to the correct project, correct hierarchy level, with appropriate labels and Kanban buckets.
The AI is the UI. The app is how I look at it.
I tell Claude: "Add to newsletter: how I became my own first customer"
Done.
Claude extracts the details, calculates the publication date, generates a structured HTML description, assigns it to the Ideas bucket in the Newsletter project, and creates the task.
I can say:
"Add Northwest Event Show 2026 to my talks" → Task created in Talks project, Ideas bucket, proposal deadline auto-calculated 3 months before the event, full HTML description with next steps "Track this content idea: why simulation beats spreadsheets" → Newsletter project, Ideas bucket, due next Monday "Create task: review elevator simulation results" → Operations project, no bucket assignment, standard due date
The AI knows the domain rules. I don't have to remember them.
The Velocity Let me repeat the timeline because it's absurd:
Friday evening: Never heard of Vikunja Saturday-Sunday: Built the entire LLM-to-GTD system Monday: Using it in production Tuesday morning: Writing about it
72 hours from zero to production.
That's what I do.
Schedule a free call: app.ivantohelpyou.com/decision-analysis