Contact
READING PROGRESS:0%
AXIOM

Every notes app sells relief from the first tax and quietly hands you the second.

José Sabogal woodblock print cover for Amauta Magazine, Volume 1 (1926)

José Sabogal, woodblock cover print (1926). Carving out heavy forms by hand, long after the preliminary sketch was laid down.

SOURCE: José Sabogal, Amauta Covers (1926), Public Domain ReviewARCHIVE PLATE ↗

Capture has two taxes.

The first tax happens at the exact millisecond a thought enters your head: which app do I open? Which folder does this go in? Should this be a reminder, a calendar block, or a draft note?

Every modern operating system and note-taking tool has spent the last ten years shaving that first tax down to the bone. We have lock screen widgets, Apple Watch voice complications, Action Buttons, and universal share sheets. Capturing a thought takes two taps.

The second tax comes later. And nobody in software talks about it.

The second tax is triage debt: every single bucket you fill must eventually be emptied by hand, one item at a time, and copied into the system where the work actually gets done.

THE PRODUCTIVITY TAX PIPELINE:
[Fleeting Thought]
       │
       ▼ (Tax 1: Reduced by tech to 2 seconds)
[Apple Notes / Voice Memos / Reminders]  <── "THE JUNK DRAWER"
       │
       ▼ (Tax 2: Compounding human triage debt)
[Sunday Afternoon Manual Triage: Copy / Paste / Classify / Move]
       │
       ▼
[Jira / Trello / Monday / Obsidian / Calendar]

The Junk Drawer Paradox

Apple Notes, Google Keep, and iOS Reminders solved the first tax by turning themselves into universal junk drawers.

You throw a phone number, a URL, a contractor estimate, and a grocery item into a blank note. You feel organized for thirty seconds. But nothing inside a junk drawer gets done on its own.

Three weeks later, you have 400 unorganized notes titled “New Note (14)”. To make any of it useful, you have to spend an entire Sunday afternoon sitting at a desk, opening each note, reading what past-you wrote, deciding what it means, and manually copying the text into your project management board or calendar.

Every capture without automated triage is just a future decision you deferred. And deferred decisions collect compound interest.

The Bucket Is the Problem

Productivity startups keep trying to solve this by building “the one unified bucket.”

“Put your tasks, docs, calendar, and team chat in one single workspace!”

It never works. Real operational work doesn’t live in one bucket:

  • Personal projects belong in simple Kanban boards.
  • Long-term engineering thinking belongs in local markdown vaults.
  • High-touch work tasks belong in enterprise trackers.
  • Family logistics belong in shared calendars.

The answer is not forcing all human activity into one proprietary database. The answer is inverting the relationship between capture and filing:

The Old Model: The human files; the system stores.

The Inverted Model: The human speaks; the system files; the human only corrects.

The Inversion: Zero-Decision Capture

When you sit down to architect a modern capture engine (like Inlet), the goal is zero decisions before the thought is saved.

You speak a sentence into your phone: “Follow up with the electrical contractor on Tuesday about the 200-amp panel permit, and save an idea about two-gate keychain permissions for the blog.”

You don’t pick a destination. You don’t choose a tag. The audio is written locally to disk in under two seconds.

A background agent reads the transcript, decomposes the compound statement into two discrete outcomes, and executes:

  1. Outcome 1 (Task): Created in Trello under HOME // Backlog with a due date of next Tuesday.
  2. Outcome 2 (Idea): Formatted into a standard markdown template with frontmatter and filed to your Obsidian vault under 20_Ideas/.

The human never opened a junk drawer. The thought was routed straight to the operational surface where it can actually be acted upon.

The Human as Auditor

Does the agent ever misfile? Of course it does.

Honest Caveats

  • Context Blindness: A background agent is great at recognizing that “call Bob about fence boards” is a task. It is terrible at detecting emotional subtext, sarcastic remarks, or half-formed philosophical fragments. Deep, ambiguous thinking still requires a human sitting down with an open journal or blank markdown file.
  • Routing Drift: As your board names, project folders, and categories change over time, background classifiers will misfile cards until you update their system prompt definitions. You trade weekend manual triage for occasional prompt tuning.

Frictionless capture without automated routing doesn’t make you productive. It just builds a bigger landfill. Stop organizing the junk drawer. Put an agent on the door.

STAY IN THE LOOP