Automating My Life with AI

The Memo

AI was supposed to help us live better. At work, it's delivering. In personal life? Everything useful sits behind another subscription, and hardware costs have gone brutal — capable local inference rigs now run five figures.

This year I decided to treat AI subscriptions the way companies treat R&D spend — a deliberate investment in my own capability.

I'm not using off-the-shelf agent frameworks. I want full control over tool definitions, permission boundaries, and execution paths. My health data, routines, and personal notes are sensitive. I'd rather build a precise, lightweight system that speaks natively to my existing tools than adapt my life to someone else's skill marketplace.

The next 6 months: I'm building production-grade personal agents to automate content creation, health tracking, meal planning, and daily orchestration. I'll publish honest progress updates as I go.

This is my public memo to myself. December 2026 — I'll come back and review what actually moved the needle.

Experiments

Automations I've shipped so far.

I publish updates on X. Follow along →