◆ Field notes from the cutting table

Notes from the
cutting table.

Essays on data quality, archives, the human craft behind AI-ready datasets, and what happens when you take ordinary work seriously.

Issue · 01
2026
Featured
Apr 28, 2026 · 8 min read · Essay

Why manual data entry still matters — even in the age of AI.

Three years into the LLM boom, the businesses winning with AI are the ones who took data entry seriously when nobody else was. A look at where humans-in-the-loop still wins, what it costs to skip the layer, and three industries where the work is heating up.

Read the essay

Up next

Field note · 02

What 99.5% accuracy actually means.

Vendors throw around accuracy numbers like marketing copy. Here's what the math looks like at field, record, and document level — and the QA process that gets us there.

Quality · Operations
Field note · 03

The cost of bad data — a CFO's reckoning.

What duplicated customer records, mis-keyed claim amounts and broken catalog data really cost — line by line, on the P&L.

Finance · Risk
Field note · 04

Why we still believe in humans in the loop.

The boring, unfashionable answer to the question every AI-first company hits in year two: who reviews the model's mistakes, and at what scale?

AI · Operations

Get future essays in your inbox.

One thoughtful piece every 3–4 weeks. No marketing email, no nonsense, no link tracking.

Subscribe →