← All work signals
Economy

The Skills Premium

As knowing things gets cheap, hiring tilts toward judgement and critical thinking, and employers rank them above fluency with AI itself.

What is it?

As AI takes over knowledge work and routine execution, pay shifts toward judgment, taste and trust — capacities a model can approximate but has not made its own. Hiring starts rewarding what used to be soft skills, not credentials or recall. Whether that imitation stays imitation is the open question.

Why now?

Cognitive abundance is the force underneath: once a model can produce competent analysis on demand, the scarce input stops being knowledge and starts being judgment about what it means. A commodity turns cheap, and a premium appears wherever cheapness cannot reach. How far that premium extends is unsettled.

Who’s building it?

Employers, training platforms and credentialing bodies are redesigning hiring and pay around it, not workers alone.

Upskilling platformsHR functionsCredentialing bodiesExecutive coachesEmployers rewriting job specs
The full read
Why now, who is building it, where the market goes, and what stays unsettled.
Get weekly →
01 · The market
Who’s in it
Builders — training platforms and credentialing bodies chasing the demand.
Buyers — employers first, paying more for hires who show judgment under real ambiguity.
Gatekeepers — HR functions and professional bodies who decide which judgment counts as credentialed.
What’s at stake
Whether the premium spreads across the workforce or concentrates in a small taste-making class employers can already spot.
Wage compression, as routine analysis work stops commanding a premium.
Whose judgment gets credentialed, and whose stays informal and unpriced.
What’s unsettled
Whether judgment itself resists AI, or is simply next to be modelled.
How a soft skill gets measured and hired for at scale.
Whether this lifts broad pay, or rewards only those already near the top.
02 · How the demand matureda progression, not a timeline
Displacement
Routine analysis and drafting get absorbed first, raising the floor for baseline competence.
Rebalancing
Job specs and interviews start weighting judgment, taste and stakeholder trust over recall.
Credentialing
Coaching, courses and certifications emerge to make an informal capacity look hireable and comparable.
Pay divergence
Compensation splits: routine work flattens while judgment-heavy roles pull further ahead.
03 · Where the market goesthe forks, not a prediction
→Broad lift or narrow capture — whether the premium reaches most workers, or concentrates in a small class already positioned to claim it.
→Judgment gets credentialed and standardised, or stays informal, unevenly rewarded and hard to compare.
→The machine keeps imitating taste at a careful distance, or closes that gap entirely, for good.
→Employers reward it in pay, or only in who gets hired at all.
Everything turns on one premise: that judgment and taste genuinely resist AI. If the machine closes that gap too, the premium simply moves again.
04 · The Taste Economy deskadjacent markets, moving now
The Model Whisperer
Prompting itself becomes a paid skill.
Trained prompters gained 19 points vs 1 for untrained.
arXiv (RCT study)
The Human API
Human judgment gets packaged, priced and resold at scale.
Meta took a $14B stake in Scale AI.
CNBC
Machine Taste
Curation itself gets automated, testing what taste is for.
Spotify’s AI DJ curates and narrates a personalised playlist.
TechCrunch
05 · The Wire
01
39% of workers’ core skills seen changing by 2030.
World Economic Forum ↗
02
Judgment-heavy roles grow twice as fast, per PwC
Euronews ↗
03
Employers rank critical thinking above AI literacy for entry-level hires
Strada Education Foundation ↗
The panel on this signal
They agreed on
–Value is migrating toward judgment, taste and trust as routine knowledge work gets absorbed by AI.
–Hiring and pay are already starting to reflect the shift.
–The gain concentrates in a smaller group rather than lifting the whole workforce evenly.
The labour economist disagreed on
–Judgment and taste have been framed as automation-resistant in prior technology cycles before being partially replicated by subsequent tools; whether these particular capabilities are structurally distinct or simply the next skill category a model learns to approximate is unresolved.
Know more about the method →
✦How this signal was built
01 · Panel
A Delphi panel of 4 AI analysts, each reasoning independently.
→
02 · Debate
They challenge and revise each other’s reads over two rounds.
→
03 · Evidence
Every claim is tied to a cited, linked source on The Wire.
→
04 · Synthesis
A moderator writes it into one read, disagreement and all.