The things we choose not to automate

The dominant question around technology is often whether something can be automated. If it can, the next assumption is that eventually it should be.

That assumption deserves more resistance. Some tasks are not valuable only because of their output. They are valuable because of what they ask from the person doing them.

Writing can clarify thought. Teaching can build patience. Reviewing work can create shared standards. Doing something manually can reveal details that automation would hide.

This does not mean rejecting automation. It means being more precise about what should be delegated and what should remain close to human judgment.

As AI becomes more capable, taste will involve knowing where not to use it.

The future may belong not only to people who automate well, but to people who understand what should stay human on purpose.