My phone looks self-contained because the power plant is not visible from the screen.
The battery hides when the electricity was generated. The charger hides voltage conversion. A thin application hides remote computers, cooling equipment, network routes, and storage devices. A packaged meal hides fields, animals, machinery, refrigeration, and waste.
This is not a conspiracy. It is the ordinary effect of abstraction and outsourcing. One system presents a clean boundary by moving complexity somewhere else.
The danger begins when a clean boundary becomes a false boundary—when I cannot see the cost, repair the thing, compare alternatives, or even identify which remote dependency performs the work.
Locality Has Several Addresses
A program can execute on my phone while depending on a remote authentication service. A document can appear stored locally while essential fonts, scripts, or keys arrive from a server. A battery-powered device can consume energy generated from a grid whose sources vary by place and hour.
So I ask several different locality questions:
- Where is the state?
- Where does the computation occur?
- Where is energy converted?
- Where are failures repaired?
- Where can the design be inspected?
- Who can continue operating it if one organization disappears?
The answers need not all be “here.” Remote infrastructure can be more efficient, reliable, and maintainable than millions of duplicated local systems. The problem is not distance. The problem is unexamined dependence.
A concise local program may reduce network traffic and preserve function during an outage. A remote service may coordinate a shared model that cannot sensibly live on one device. Good architecture decides deliberately instead of using “cloud” or “edge” as moral labels.
Every Abstraction Should Expose Its Bill
An abstraction earns trust when it hides mechanics without hiding consequences.
For energy, I want to know:
- total energy per useful operation;
- peak power;
- standby consumption;
- where conversion losses occur;
- and what happens at end of life.
For computation:
- which state leaves the device;
- what remote work is required;
- what latency or outage breaks the experience;
- and whether a local export can preserve the user’s work.
For manufactured goods:
- which parts wear out;
- whether they can be replaced;
- which materials complicate recycling;
- and whether documentation exists.
These do not need to become a wall of guilt around every button. They can be engineering facts available at the moment of choice.
If a new version of an office application needs a remote data center to perform work an older local machine handled, the design should be able to explain the new value and the new dependency. “It is all necessary” is not an explanation.
Outsourcing Also Outsources Skill
When a product becomes impossible to inspect, the user loses more than repair access. A community loses the chain of skills that connects a symptom to a cause.
A sealed device says: replace the unit. A documented circuit says: measure the rail, inspect the connector, compare the waveform, replace the failed part. The second path creates technicians, toolmakers, teachers, and improved designs.
No individual needs to fabricate every transistor or refine every material. Ownership is not total self-sufficiency. It is a ladder:
- I can observe what the system is doing.
- I can retrieve my state.
- I can diagnose a bounded failure.
- I can replace ordinary parts.
- I can modify documented interfaces.
- A local specialist can reproduce or improve important components.
A healthy technical culture lets more people climb farther than “buy another one.”
A Public Fabrication Room in Every City
The twenty-first-century workshop should reach beyond 3D printers and desktop circuit boards.
I imagine shared regional facilities for:
- precision measurement and microscopy;
- micromachining;
- controlled thin-film deposition;
- photolithography;
- packaging and interconnect experiments;
- material characterization;
- and the safe chemical and ventilation infrastructure these processes require.
These are not casual tools to place on an unattended table. High voltage, vacuum systems, reactive chemicals, radiation sources, compressed gases, lasers, and fine particles require trained operation, maintenance, monitoring, and waste handling.
That is precisely why the facility should be shared.
A city already supports libraries, workshops, laboratories, trade schools, utilities, and emergency services. A professionally operated public fabrication room could combine access with responsibility. Independent inventors, students, repair specialists, artists, and small manufacturers could prototype beyond the boundaries imposed by consumer tools without pretending that industrial processes are harmless.
The output would not need to compete immediately with a giant semiconductor fab. The first value is literacy and iteration: measure a film, pattern a sensor, inspect a failure, package a tiny device, or discover that a proposed process does not work.
Community Ownership Is More Than Access Time
Paying by the hour for a proprietary machine is useful access, but it does not by itself create shared technical power.
Community ownership needs:
- operating documentation;
- maintenance knowledge;
- calibration records;
- transparent pricing;
- safe process recipes;
- a route for proposing modifications;
- and a repository of failures as well as successes.
The institution should preserve knowledge when one operator leaves. It should let members understand why a process has a boundary rather than treating every rule as an arbitrary prohibition.
This is where local fabrication and open documentation reinforce each other. A schematic without tools can remain theoretical. A tool without documentation becomes priesthood.
What I Would Ask of One Product
Take an ordinary connected temperature monitor.
I want the product to state:
- what it measures and with what practical resolution;
- which calculations happen locally;
- which data is transmitted;
- whether it works without an account;
- how long local history remains available;
- how the enclosure opens;
- whether the sensor and battery can be replaced;
- and where the design can be inspected.
Then a local workshop should be able to build a compatible enclosure, test a replacement sensor, verify power consumption, or adapt the interface for a greenhouse.
That does not eliminate the original manufacturer. It gives the product a longer and more interesting life.
Make the Boundary Legible
Civilization depends on specialization. I do not want to personally raise every ingredient, generate every watt, host every service, or machine every component.
I want the boundary between my life and that infrastructure to remain legible.
A legible boundary tells me what crosses it, what it costs, what fails, and what choices remain mine. It lets a skilled community reopen the box when the abstraction stops serving us.
The goal is not to drag every hidden pipe into the living room. It is to own enough tools—and enough understanding—that the invisible world behind the product can still be questioned, repaired, and changed.