Almost every conversation about choosing a degree runs on two pieces of advice that contradict each other and are both unreliable. There is a third question that is more useful than either, and it is not about you finding your calling.

The question is usually framed as a choice between two bad pieces of advice: follow your passion, or pick something employable. Both fail for the same reason. They treat a subject as a container of content rather than a way of working.

What “follow your passion” gets wrong

Passion is a result far more often than it is a cause. Most people become absorbed in things they have become good at, and they become good at things they were willing to be bad at for a while.

Asking an eighteen-year-old to identify a pre-existing passion mostly selects for what they have already been exposed to. That is a small sample, and it is heavily shaped by which teachers they happened to get.

What “pick something employable” gets wrong

Employability advice is always a description of the recent past. It reflects the market that existed when the advice was formed, and it arrives with a lag roughly equal to the length of a degree.

It also flattens enormous variation. Two graduates of the same subject can end up in completely different positions depending on what they can actually do, which is rarely determined by the subject name on the certificate.

A better question

Instead of what you love or what pays, ask: what kind of thinking do I want to get better at? Subjects are training in particular mental operations, and the operations are what transfer.

• Some train you to build formal models and reason within them — economics, engineering, mathematics.
• Some train you to interpret evidence that does not settle the question — history, law, much of the social sciences.
• Some train you to design and test under uncertainty — the experimental sciences, medicine, parts of computing.
• Some train you to read closely and write precisely — literature, philosophy, languages.

Most careers are built out of two or three of these, in a mix nobody announced in advance. Choosing by operation rather than by job title gives you something that survives the market shifting under you.

Where AI actually changes the calculation

The honest position is that nobody knows the shape of this in ten years, and anyone claiming otherwise is guessing with confidence. But one pattern is already visible and worth planning around.

These tools are strongest at producing plausible first drafts of well-defined tasks: summarising, restructuring, generating boilerplate, translating between formats. They are weakest at deciding which task is worth doing, judging whether an output is actually right, and taking responsibility for a decision.

The practical implication is not “avoid subjects that involve writing code or text.” It is that the value in most fields is shifting from producing the artefact toward specifying, checking and owning it. Those are judgement skills, and judgement is built by doing the underlying work yourself long enough to recognise when something is wrong.

Which is an argument for learning a field properly rather than for choosing a different field.

Three things worth doing before you commit

• Read a second-year reading list, not a prospectus. The prospectus is marketing. The reading list is the actual daily experience of the subject. If it bores you now, that is information.
• Find out what the assessment looks like. A subject assessed by continuous problem sets is a different life from one assessed by two long essays a term, regardless of the topic.
• Talk to someone two years in, not someone who finished ten years ago. Curricula change, and nostalgia is a poor guide.

The decision is more reversible than it feels

Choosing a subject feels permanent at seventeen and mostly is not. Conversion routes exist, postgraduate study redirects people constantly, and a large share of graduates work outside the field named on their degree.

That is not a reason to choose carelessly. It is a reason to choose without panic, on a basis that is actually about you, and then to do the thing properly — which matters far more than which thing you picked.