← All education signals
Labour signal

The Skill Half-Life

What you know expires faster each year, and relearning stops being an event and settles in as a permanent condition.

What is it?

Knowledge is starting to behave like a perishable good, its usefulness shrinking fastest in roles closest to the tools themselves. The response isn’t a subject to learn but a permanent posture: relearning as routine. Whether this reaches beyond AI-exposed work is the real unknown.

Why now?

Cognitive tasks that took years to automate now shift in a release cycle, so knowledge closest to the tool ages fastest. The scarce resource stops being expertise itself and becomes the capacity to rebuild it. The premium shifts from what you know to how fast you can relearn it.

Who’s building it?

Employers and training platforms rebuild certification around live proof — set by model releases above them.

Corporate L&D teamsMicro-credential platformsModel providersCertification bodiesHiring managers
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 — corporate L&D teams and micro-credential platforms beneath them.
Buyers — workers in AI-exposed roles first, then anyone whose expertise ages against tools.
Gatekeepers — certifying bodies and employers deciding what counts as current proof.
What’s at stake
The certifying layer captures value once held by degree programs — proof of current capability replaces proof of past attendance.
Training budgets shift from one-off courses to constant renewal.
Reskilling infrastructure is unproven at the scale now required.
What’s unsettled
Whether this generalises beyond AI-exposed roles to the wider workforce.
How fast credentials can be verified without becoming meaningless noise.
Whether firms reward adaptability in practice, or still default to fixed job descriptions.
02 · How the demand matureda progression, not a timeline
Novelty
A first course or credential feels sufficient, briefly, before it fades.
Erosion
That credential ages as the tool it describes keeps changing shape.
Reinvention
Workers rebuild the skill from scratch, and the rebuilding itself becomes the visible, marketable capability.
Infrastructure
Enough repeated cycles justify real certification systems, real hiring signals, real employer trust.
03 · Where the market goesthe forks, not a prediction
→Universal or concentrated — decided by whether AI reshapes most jobs at this pace, or stays contained to roles that touch the tools directly.
→Certification could consolidate around a few trusted verifiers, or fragment across every employer.
→Degree programs adapt into continuous formats, or lose ground entirely to micro-credential platforms built for renewal.
→Value moves toward whoever verifies current capability, not whoever taught it first.
The open question is whether relearning becomes everyone’s condition, or stays a tax paid only by the people standing closest to the machines.
04 · The Post-Labor Workforce deskadjacent markets, moving now
The Automation Wage
Compensation debates as jobs get automated
OpenAI’s 2026 blueprint proposed robot taxes and a 4-day week.
TechCrunch
The Deskilling Debate
Whether relying on AI erodes core human skill
Endoscopists’ adenoma detection fell from 28% to 22%.
The Lancet Gastroenterology & Hepatology (PubMed)
The Portfolio Career
Careers splitting into multiple simultaneous roles
US multiple jobholding was 5.5% in July 2026, after a 6.0% peak.
Federal Reserve Bank of Richmond
05 · The Wire
01
Skills in AI-exposed jobs are changing 66% faster than in others.
Process Excellence Network ↗
02
44% of workers’ core skills will be disrupted by 2027.
World Economic Forum ↗
03
One in 10 advanced-economy job postings requires at least one new skill
International Monetary Fund ↗
04
Adapting to AI’s shift requires new skills matching changing labour demand
OECD ↗
The panel on this signal
They agreed on
–Knowledge now behaves like a perishable good, and the premium shifts to the capacity for renewal itself.
–Training built for attendance is losing its value.
–Certification is moving from one-off credentials toward continuous proof of current capability.
The learning scientist disagreed on
–The cited depreciation figures are drawn from occupations with the most direct tool exposure; extending that rate economy-wide assumes a uniform pace of change across roles that current data has not measured outside those closest-affected categories.
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.
✦Signal changelogre-read as evidence changed
2025
What you know expires faster each year, and relearning stops being an event and settles in as a permanent condition.
→
2026
A synthesis of 65 peer-reviewed studies (2015-2025) finds micro-credentials mostly serve as supplements alongside degrees rather than replacements, with employer recognition uneven and stronger in technical than soft-skill fields. Frontiers in Computer Science (peer-reviewed) ↗