The Job Market Is Repricing Judgement, and That Is the Whole Signal

In May 2026, entry-level job postings on Indeed were down 7.5% against the previous year while senior-level postings were up 14.7%. Recent graduate unemployment reached 5.7%, higher than the rate for the workforce as a whole for the first time in decades. At companies that adopted generative AI, entry-level hiring fell by as much as 80% quarter on quarter.

Read together these describe a market that is not contracting so much as changing what it is willing to pay for.

The word researchers have settled on

Seniorization. Employers have been rewriting junior job postings to require things that used to appear only in senior ones: judgement, stakeholder management, the ability to decide under ambiguity.

That is a strange thing to ask of a graduate, and the strangeness is the point. The tasks that a junior role existed to cover, the drafting, the summarising, the first pass at the analysis, are the tasks that got cheap. What remains in the role is the part nobody has worked out how to buy: knowing whether the output is right, and being accountable for saying so.

What is actually being priced

Not intelligence, and not effort. Something narrower, which is calibrated judgement inside a specific domain.

“If the junior rungs are the mechanism by which people acquire calibrated judgement, and the junior rungs are being removed, then the supply gets tighter over time rather than looser.”

Calibrated is the operative word. It means having been wrong enough times in one context to know the shape of your own error. That cannot be taught in a course and it cannot be acquired quickly, which is precisely why it is holding its price while everything adjacent to it falls.

This is also why the shortage is structural rather than cyclical. If the junior rungs are the mechanism by which people acquire calibrated judgement, and the junior rungs are being removed, then the supply of the thing being priced upward gets tighter over time rather than looser.

If the junior rungs are the mechanism by which people acquire calibrated judgement, and the junior rungs are being removed, then the supply gets tighter over time rather than looser.

The trap in reading this as good news

If you have fifteen years behind you, the temptation is to conclude your position is safe. The data supports something more limited than that.

Postings listing AI skills rose 144% in the year to April 2026 against 7% growth in postings overall. That is a skills-listing number and not a judgement number, and conflating the two is how experienced professionals end up completing a certificate they did not need. The premium is on judgement applied to a domain. The listing growth is on tool familiarity, which is genuinely useful and is not scarce.

The second limit is more uncomfortable. An employer paying a premium for judgement is buying it as an input to their business at a salary. The same judgement, applied to a problem you own, is not capped that way. The repricing is real in both directions, and only one of them has a ceiling set by a compensation band.

What follows from it

The practical read for someone mid-career is that the asset has appreciated and the best place to hold it may not be where you currently hold it.

That is not an argument for leaving a job. It is an argument for noticing that the scarce input in the market is the one you spent a decade accumulating, and that the number of ways to convert it into income has gone up at the same time as the tools for doing so got cheap.

Start by writing down the judgements you make that a competent newcomer could not, and be specific. Those sentences are the inventory. Everything else is a question of what you build on top of them.

See entry-level roles disappearing for the earlier version of this trend, and what AI cannot automate for where the line currently sits. How the system works covers converting the inventory into something you own.