The new meaning of expertise.

For much of history an expert was someone who knew things other people could not easily find out. The physician knew the remedies, the lawyer knew the statutes, the engineer knew the tables. Access to that knowledge has been widening for decades, and intelligent tools widen it further. A reasonable person might ask what is left.

More than we can tell

The philosopher and chemist Michael Polanyi offered a starting point in a single sentence: "we can know more than we can tell." His example was simple. A person can recognize a face among thousands and cannot say how. Much skilled knowledge is like this. It lives in perception and practice, and it resists being written down.

This tacit knowledge is what separates an expert from a well-informed beginner. Both can look up the same facts. The expert sees which facts matter here.

What experts do with it

They notice. An experienced clinician registers that something about a patient is off before any test confirms it. An experienced editor feels where an argument goes slack.

They frame. Experts spend more effort deciding what kind of problem they face. Novices tend to start solving at once.

They know the edges. Expertise includes an accurate sense of where one's own knowledge ends, which is what makes an expert's confidence worth something.

They answer for the result. Professional expertise has always carried responsibility. Advice from a source that bears no consequences is a different thing from advice someone must stand behind.

Information tells you what is generally true. Expertise tells you what is true of this case.

The problem of how experts are made

If expertise is tacit, it cannot be acquired by reading. It is built through years of practice with feedback, usually beginning with routine work that looks dull from outside: the standard cases, the first drafts, the basic analyses. That routine work is exactly what capable tools now do well.

This creates a real dilemma. Removing routine work from juniors makes sense for efficiency and may remove the ground on which they would have become seniors. Nobody yet has a settled answer. Promising directions include deliberate practice on realistic cases, closer work alongside experienced people, and using tools to give faster feedback instead of replacing the attempt.

A firmer definition

Expertise, on this account, is not the possession of information. It is trained perception, sound framing, calibrated confidence, and accountability, all in a particular field. None of those is made obsolete by a tool that knows the facts. All of them have to be grown, and the growing needs protecting.