The skills-based organization is mostly theatre
Everyone is selling it. Few can show it working.
Where work, technology and people get rebuilt.
The Hype Index · Week of 22 May 2026
"Skills-based organization"
Mostly a taxonomy purchase. The hard parts — manager assessment, comp decoupling, internal mobility — remain untouched at the companies announcing the transition.
Latest
Recently published
-
Inclusion doesn't survive being bolted on to AI
Seramount's framework is a useful corrective. The data behind it is harder to ignore.
-
Organisations are buying AI faster than people can learn it
The tools are deployed. The time to use them is not.
-
Test la connection
A short test of the Notion → site bridge
-
The four-day week is mostly a recruiting tool
Treat it as one and the math works out. Sell it as productivity and it collapses.
-
The nine-box grid, line by line
The classic talent ritual, dismantled. What it claims to measure, what it actually measures, and why HR keeps using it anyway.
Editor's Picks
What to read first
- 01The skills-based organization is mostly theatre
Everyone is selling it. Few can show it working.
- 02The nine-box grid, line by line
The classic talent ritual, dismantled. What it claims to measure, what it actually measures, and why HR keeps using it anyway.
- 03Why the HR tech stack keeps re-consolidating
Every five years someone announces the suite is dead. They are wrong each time, for the same reason.
The Debate
Should AI make hiring decisions?
Case for
Structured AI scoring can reduce individual interviewer bias and surface candidates whose résumés do not match conventional patterns — making the funnel more consistent than ad-hoc human screening.
Case against
Algorithmic systems trained on historical hiring data reproduce — and at scale amplify — the discriminatory patterns of the past. The problem is the training data, and there is no neutral training data.
The Theory
Use AI for sourcing and structured first-pass screens. Keep humans on the decision and on the audit. The danger is not that the model is wrong — it is that the model is fast.
The Number
73%
Of managers report using a generative AI tool in their workflow at least weekly
Sample figure Sample figure — pending verified source ↗
Sections
The four pillars
The Canon
A working bibliography for the publication.
-
On the conditions modern knowledge work systematically prevents — and what people lose when they cannot concentrate.
-
On the proliferation of work that those doing it cannot themselves justify, and what its persistence reveals about the modern firm.
-
On unconscious bias, the limits of awareness as a fix, and what actually moves the numbers inside organisations.
-
1956. On the social ethic of mid-century corporate America — still the cleanest mirror for what we kept and what we replaced.
Members
Become a member of The Work Theory
Get the long-form theory pieces, the full archive of debates and teardowns, and the weekly briefing — supporting independent editorial work on the future of work.
The Brief — weekly
The Work Theory, in your inbox.
Independent writing on the future of work, delivered when it is worth reading. Never on a schedule for its own sake.