The 2031 worker-shape forecast
A forecast for what a "knowledge worker" actually looks like by around 2031, once most workers are deep into the later stages of the 7-stage roadmap. Not a prediction about tools—a prediction about roles.
New, August 28, 2026—promoted from a recurring thread in developing-thinking.md, evidence-backed rather than a single-source hunch.
The picture
By ~2031, a "worker" is a small team: one human plus N agents, running more or less continuously (the 7-stage roadmap's Stage 6, "AI as a pod"). Within that shape, the knowledge-worker market splits into three types:
- Cognitive owners—rare context plus judgment, the actual source of expertise. What the cognitive stack calls layer 2, the brain.
- Cognitive operators—run agent fleets well. Directing, reviewing, deciding what's good enough (Stage 5, "AI as a fleet").
- Cognitive curators—maintain the brain modules and skill libraries other people's pods run on. The human side of a knowledge factory.
The generic middle collapses. Work that used to be "produce an artifact for someone else's review"—the brief, the deck, the order form—goes first, because the accountable person can often go straight to AI faster than routing through the intermediary who used to draft it for them.
Independent corroboration
Paul Roetzer (MAICON) arrived at a similar typology independently: Architect / Orchestrator / Apprentice. His "Apprentice" category names a real gap this forecast doesn't fully answer either: how do future experts develop judgment when AI absorbs the tactical learning rungs juniors used to climb? The traditional ladder—research, drafting, analysis, coordination—built judgment by making juniors do the work. If AI does that work instead, nothing yet replaces the ladder.
The evidence, not just the shape
Wage and employment data now gives this a leading indicator, not just a plausible narrative: non-professional occupations (admin, sales, customer service) are decelerating below professional wage growth, and call-center employment is down 39% against its historical trend in the US since 2022. Consistent with the generic middle collapsing first, ahead of the judgment tier.
Separately, young college graduates now have higher unemployment than non-graduates, concentrated specifically at first-job hiring in AI-exposed occupations—a possible leading indicator that entry rungs are eroding before the generic mid-career middle does, sharpening the Apprentice-gap question above.
Relationship to other frameworks
This is the labor-market counterpart to the 7-stage roadmap's Stage 5 ("AI as a fleet") and Stage 6 ("AI as a pod")—the roadmap describes how an individual worker's relationship to AI evolves; this describes what the resulting labor market looks like once most workers are deep into those stages. It also names the human role inside the knowledge factory: curators are exactly the "input owners" and "domain SMEs" that architecture requires.
Using this framework
Deploy when:
- Workforce planning conversations default to headcount reduction—reframe around which of the three types a role is becoming, not whether it survives
- Someone asks "what happens to entry-level hiring"—this is the sharpest current answer, and it's honestly incomplete: the Apprentice-gap question has no good answer yet
- A company's AI strategy has no answer for who maintains the shared context/skill libraries other workers' agent fleets depend on—that's the curator role, and most orgs haven't named it yet
- Evaluating whether a role is "safe" from AI—ask whether it's currently structured as artifact-production-for-review (generic middle, collapses first) or judgment-plus-context (owner, more durable)