Julien Dambron

Thank God the Machine Took the Old Work

ยท 2 min read

In 1485, Brother Gabriel of St. Gall watched a printing press eat the work of scribes.

He wrote: "The machine has taken our old work. Thank God. Now we can begin our real work."

That is the only honest way to look at AI.

The old work is vanishing. Copy. Boilerplate. First drafts. Sorting. The industrial equation that said your job was your worth.

Losing that is only a tragedy if you believed the equation.

A cognitive steam engine #

The steam engine did not make humans weaker. It made muscle cheap. What remained valuable was direction.

Language models do the same thing to thought. They are a cognitive steam engine. Research, code, and synthesis get cheaper by the week. Judgment does not.

The machine can produce almost infinite intelligence. It cannot tell you what is beautiful, true, or good.

That is the job that does not get automated. Steering.

Where scarcity moves #

When the cost of generating text, code, and images falls toward zero, the market floods with mediocrity. Taste becomes scarce. Curation becomes scarce. Verification becomes scarce. So does the courage to say "this one, not that one."

Do not compete with the machine on speed, memory, or volume. You will lose. Freestyle chess already showed the pattern: a skilled human plus a machine beats the human alone and the machine alone. A weak operator makes the team worse than the computer by itself.

So the skill is not "using AI." The skill is piloting it. Know where it is strong. Know where it lies. Steer.

How you work it #

Treat the model as a sparring partner, not an answer vending machine. Feed it your outline. Ask for the three strongest objections. Ask what you are not seeing. Keep the judgment. Borrow the pressure.

Stop selling hours of execution. Run your mind like a small studio. Orchestrate agents. Keep the questions. Keep the taste.

The highest use is not ad targeting. It is a tutor for every student. A second set of eyes in an ICU. A way to put expertise in more hands than a university ever could.

The best defense against misuse is not a priesthood of labs. It is lots of people with the tool, able to build, check, and stop bad uses.

The getting-better curve #

Whatever you are using today is the worst version you will ever use. Do not design your life around its current flaws. Design around the fact that it will get better, fast.

The future is not written. Corporate inevitability is a story. You still choose what kind of garden the machine grows.

Privilege #

Brother Gabriel was right. The press was not the end of work. It was the start of better work.

AI is your printing press.

The question is not whether the old work disappears. It will, and it should.

The question is whether you see liberation, or only loss.