# Stuck in 1995: paper minds in a sensor world, real cost

Canonical URL: https://themodernhermit.blog/paper-minds-sensor-world-real-cost/

Published: 2025-10-22T22:45:53+00:00

Updated: 2026-11-04T22:43:03+00:00

Some bosses still treat the factory like a filing cabinet with a power socket.

*This article has been refreshed because the old problem has not gone away. AI, robots and sensor systems are moving fast, but too many factories still copy live data onto boards, emails and spreadsheets as if duplication is a production strategy.*

I keep seeing the same factory ritual. The machine has the data. The software has the data. The screen has the data. Then someone still writes the same thing on a board, sends it in an email, puts it into Excel, and later everyone wonders why the numbers don't match. This is not control. This is duplication with a hi-vis vest. I am not against visual boards. A clean board on the shop floor can help people see what is happening quickly. The problem starts when the board becomes a second system, then a third system, then a museum of manual copying. If a fault is already logged by the machine, why does someone need to write it again by hand? If the product change is already in the system, why does the line need a paper note, an email chain and a spreadsheet with formulas nobody dares to touch?
When the same fault comes back again and again, the answer is often another checker with a clipboard. Not a sensor. Not a camera. Not a simple screen. Not a clean data pipe. A human being, standing there, catching what the process should have caught before. That is not smart production. That is using people as cheap sensors with shoes. Companies say sensors cost money. Screens cost money. Gateways cost money. Proper software support costs money. Fair enough. But mistakes also cost money. Downtime costs money. Waste costs money. Rework costs money. Burnt-out workers cost money, even if the spreadsheet prefers not to see them. The old way often looks cheaper only because the damage is hidden inside people's time, stress and patience. I understand caution. Not every line needs a robot arm. Not every task needs AI. But if a repeated task can be counted, copied, checked and corrected, then the first question should be simple: can we remove it, automate it, or make it visible in one place? Humans should solve problems. Humans should improve work. Humans should notice what machines cannot. They should not spend half a shift proving that paper still exists.
Maybe the future of manufacturing does not begin with AGI. Maybe it begins with the courage to stop writing the same number twice.

**Related sense check**: this same old-factory problem becomes even uglier in an AGI world, where advanced companies talk about robots and trillions while slower companies still ask workers to copy software data onto a whiteboard. Read the wider version here: [AGI Dreams, Spreadsheet Factories And Safety Glasses Rule](https://themodernhermit.blog/agi-dreams-spreadsheet-factories-safety-glasses/).

---

Scene 1: **Fault response: add paperwork**
*Production line. A small sensor with a price tag sits on a table. Beside it, two new hires stand with clipboards.*
Dialogue:
Finance: "Too expensive."
Worker: "So are faults."
Manager: "Hire checkers."

Scene 2: **Duplicating the past**
*Boardroom. A whiteboard is crammed with hand ticks. A laptop on the table shows the same tasks already closed.*
Dialogue:
Manager: "Great progress."
Worker: "It's already done."
Laptop: "I was here first."

Scene 3: **Heritage solutions**
*Finance meeting. A slide shows downtime costs. Someone points quietly at a sensor option while the chair looks away.*
Dialogue:
Engineer: "Fit the sensor?"
Chair: "We prefer heritage."
Printer: "I'm back."

---

---

Count what breaks, how often it breaks, and how much it costs.
If you can write it twice, you can automate it once.

**Verdict**: *Toss.*
**Toss**handwritten duplication, orphan spreadsheets and paper check rituals.
**Keep**visual boards only if they sync to live data. Keep humans for judgement, repair and improvement, not for pretending to be sensors with shoes.

---
