Every few weeks the internet re-litigates whether the future belongs to generalists or specialists, and the argument always ends in the same shrug: be a hybrid, stay adaptable, it depends. It is true and it is useless, because nobody attaches a number to breadth. Breadth is measurable. We measure it on every occupation we track.
Here is the measure. For each occupation, count how many other fields its people could cross into at 40 percent skill coverage or better, its out-degree: the number of doors its current skill set already opens. That count is what “generalist” means once you stop using it as a personality type. It is optionality, and it is not evenly distributed.
Read that again, because it is the whole argument. Optionality is bimodal. A large group of professions, 60 of 153, is boxed in at the strict coverage bar we set: their skill bundle does not cover 40 percent of any other single occupation’s demands. A smaller group ranges across the map. Most people sit in the middle, one door open. The generalist advantage is real, but it is a minority position, and it is a property of your skills, not your temperament.
Who is wide, and who is boxed in
| Occupation | Fields reachable | Shape |
|---|---|---|
| Operations manager | 15 | Generalist bundle |
| Business analyst | 15 | Generalist bundle |
| Sales engineer | 14 | Bridge role, technical plus commercial |
| Data scientist | 11 | Portable analytical core |
| Software engineer | 9 | Portable technical core |
| Copywriter, paralegal, creative director | 0 | Deep, specific, boxed in |
The wide occupations share a shape. Operations manager, business analyst, sales engineer, project manager: these are bundles of coordination, analysis, and communication, the skills that recur across the most fields. The narrow ones, copywriter, paralegal, creative director, are deep and specific, and their depth is exactly what does not cover another occupation’s demand list. This is the transferable-skills finding seen from the other direction: the skills that travel are the general ones, so the people who hold more of them travel further.
Why breadth is winning this particular decade
The reason the generalist question feels urgent now is AI, and here the data and the discourse agree. As automated tools absorb narrow, well-defined execution, the value that stays human is the part that connects things: judgment across domains, the decision about what to build, the translation between a model’s output and a business’s need. Those are breadth skills by definition. A wide out-degree is not only more exits in a crisis; it is a hedge against your specific deep skill being the next one commoditized.
That does not make specialization a mistake. Depth is what lets a person guide the tools rather than compete with them, and the highest pay still concentrates in scarce, deep, often licensed skills, as our piece on the adjacency premium showed. The honest synthesis is the one the think-pieces gesture at without evidence: hold a deep core so you are worth hiring, and enough general range that your core is not your only door. The difference is that now you can measure the range instead of guessing at it.
How to read your own number
Count your doors before you assume you have options, and before you assume you have none. A specialist reading a zero on the strict bar is not doomed; it means the nearest crossings sit below 40 percent coverage and will take real skill-building, which is worth knowing precisely rather than vaguely. A generalist reading a twelve is not safe; it means the market can route you many places, none of which is guaranteed to pay more, per the gravity finding. The instrument on the front page draws your doors from your own skills, and the route pages price the specific ones. Breadth stops being a debate the moment it becomes a count.
PivotHop pipeline, July 2026 run: 77,443 live job postings across 153 occupations. Breadth, or out-degree, is the number of occupations whose demanded skills the origin’s profile covers at 40 percent or more, computed over the top 20 skills per occupation from posting text. Distribution across the 153: 60 occupations reach zero fields at that bar, 27 reach one, 37 reach two to four, 29 reach five or more; median one, mean 2.3. The 40 percent bar is deliberately strict; lower it and every count rises, but the shape, a wide floor of boxed-in fields and a thin ceiling of broad ones, holds. Run your own count on the front-page instrument.