COOKED
The core deliverables — API references, user manuals, release notes, help-center articles, SOPs — are exactly the structured text generation LLMs do at usable quality from source code, specs, and Jira tickets. What resists is the upstream work: extracting undocumented knowledge from engineers who won't write it down, deciding what the documentation set should be, and owning regulated deliverables in medical device, pharma, and aerospace where traceability and validation matter. No license protects the role, and doc teams are already being consolidated into fewer, more senior information architects supervising generated drafts.
Part 2020 shock, part continued decline in the years since.
Median pay $72,850 → $90,390 -0.7% in real terms
This line is counted by the Bureau of Labor Statistics — the one figure on this page that isn't a judgement of ours. Headcount moves on demand, offshoring, demographics and the business cycle, and automation is one term among several, often not the loudest.
So a falling line is not evidence that AI did it, and a rising one is not evidence that it won't. Both happen in this register: some occupations resist automation and shrink anyway, others are highly automatable and keep growing. The marked year is 2020.
BLS projection, 2024–2034
+0.9%
Percentage only. The projection counts a different population from the 45,500 above — it includes self-employed workers, which for this occupation is most of them, so the two headcounts are not comparable.
Exposed, but growing
AI can already do a lot of these tasks, and the BLS still expects +0.9% more of these jobs by 2034. Demand for the output is growing faster than the work is being automated away — the mechanism BLS gives for software developers, and the combination people most often misread as an error.
Different clocks. The score is what current AI could do to this work today. The projection is how many of these jobs will exist in 2034. Everything between the two — how fast employers actually adopt, whether demand grows in the meantime — is why they can point opposite ways without either being wrong.
~4,500 openings a year on average, including replacing people who leave.
Grant WriterLexicographerManual WriterReport WriterMedical WriterClinical WriterHandbook WriterProposal WriterContracts WriterNarrative WriterTechnical WriterScientific WriterTechnician WriterEngineering WriterDocument SpecialistDocumentation WriterProposal CoordinatorRequirements AnalystSpecification WriterHealth Science WriterInformation DeveloperDocumentation DesignerTechnical CommunicatorHealth Technical Writer
Holding it up: judgment & accountability . Weakest point: liability shield .
Core tasks are already automatable Turning a merged pull request, a Swagger spec, or a Jira changelog into a release note or endpoint reference is a text-transformation job a model does in one pass, and the 6 rather than 2 is bought entirely by the interviewing — cornering the engineer who shipped an undocumented flag, watching a beta user fail the install, and reverse-engineering behavior the spec got wrong.
Fully desk- and screen-based The 3 covers the rare days you leave the screen — hardware teardowns for an installation guide, photographing a service panel, walking a factory line to check that the SOP matches what operators actually do — but the deliverable is written in a docs-as-code toolchain and the physical trip is optional for most of the corpus.
No licence, no signature requirement There is no credential gating the title; the STC certifications are resume decoration, and on FDA 21 CFR 820 device labeling or DO-178C aerospace documentation the signatures on the traceability record belong to regulatory affairs, quality, and the DER — you draft, they approve and carry the exposure.
Meaningful discretion The 7 sits above pure procedure because you decide the information architecture: whether something is a tutorial or a how-to, when a deprecation notice goes loud, how to word a warning in an installation step where getting it wrong injures someone — but those calls run through docs review, legal for regulated labeling, and engineering sign-off before they ship.
The verdict above describes this occupation as a whole. Almost nobody does the typical version of a job — tick what's actually in your week and see how your own mix sits.
Your task mix speaks to task resistance (6/20 here) — how much of the day's work current AI already does. That is the dimension the boxes above are about.
It cannot move the other three. Liability shield (2/20) is whether the law requires a licensed human to sign. Trust premium (5/20) is whether buyers specifically pay for a person. Judgment and accountability (7/20) is whether the role exists to own consequential calls. Those are facts about the occupation's standing, not about which tasks are in your week — a paralegal who does only trial exhibits still holds no licence. Together they are 14 of this occupation's 23 points (61%).
Embodiment (3/20) is also a property of the work rather than the worker, but we don't tag individual tasks as physical or not, so the picker can't tell you anything about it. That's a limit of this tool, not a claim.
Did we get the list right? Tell us what's missing — the tasks are written from the outside, and you're reading this from the inside.
No occupation passed every test: close enough to technical writers on skills and subject matter, at least 10 points more resistant, no big jump in training, no new licence, no pay cut, and not shrinking on its own. That happens for 223 of the 654 occupations here that aren't SAFE, and it is worth stating plainly rather than leaving the section off.
The usual reason is that exposure travels with the skill profile. The jobs most similar to yours tend to be exposed for the same reasons yours is, so the near neighbours don't clear the gap — and the ones that do are a different kind of work, not a transfer of what you already know. Read that as a limit of this method, not a verdict that you're stuck: it only compares whole occupations, and it cannot see specialisation, industry, or anything you'd bring that isn't in a federal skill survey.
Here is that claim on your own job rather than in the abstract. These are the three occupations closest to this one by skill and subject matter — the places the work would most naturally transfer — with what the register scores them:
That is the whole problem in three lines. The nearest work is not meaningfully safer, so there is no move here that trades a similar skill set for a better verdict. This is not us running out of ideas — it is what the neighbourhood looks like.
The moves above are yours to make. This is the other half: what would have to change in the world for the occupation itself to score higher. None of it is in any one person's gift, but it is where the floor actually comes from. Scores here are not a one-way ratchet. Only two of the five dimensions — task resistance and embodiment — track what machines can do. The other three track law, what buyers will pay for, and who is answerable, and those move in both directions, often in response to the same pressure AI creates. If every lever below landed, this occupation would score around 38/100 — EXPOSED.
Task-mix shift is genuinely two-tiered here: if generated drafts absorb API references, release notes, and help-center articles, the residual role is information architecture, docs-as-code toolchain ownership, taxonomy/metadata design for retrieval, and eliciting undocumented knowledge from engineers in interviews. A further concrete driver: enterprises requiring curated, version-accurate documentation corpora as grounding for their own customer-facing support bots — making corpus stewardship an ongoing judgment job rather than page production.
Regulated-documentation regimes attaching named-human accountability: FDA 21 CFR 820.30/IEC 62304 design-history-file and IEC 82079 instructions-for-use where a named person signs off, plus EU MDR technical documentation and the EU AI Act Annex IV technical documentation duty. If notified bodies or FDA inspectors begin requiring an identified human author/reviewer attestation for AI-generated IFUs, labeling, and validation protocols — and 21 CFR Part 11 e-signature records name that person — the medical-device/pharma/aerospace slice of the occupation gains a real signature requirement.
Formal ownership of documentation as a safety artifact: if aerospace (S1000D/ATA), pharma validation, or device firms designate a documentation owner who must justify what is documented, what warnings appear, and what traceability matrix links requirement to test — and that person is named in audit findings or recall investigations — the role owns consequential calls rather than executing tickets.
Narrow route only: procurement and audit contexts where a customer or regulator demands human-authored, human-verified documentation — e.g. defense/aerospace contract clauses barring unattested generative output in deliverables, or hospital purchasers requiring attested IFUs. No general consumer or developer-docs market pays for human authorship.
The limit. Even with the regulated-deliverable levers, the headcount story doesn't reverse: signature and architecture duties concentrate in a small senior tier per company, so scores can rise for the surviving role while the occupation's worker count keeps falling. Embodiment has no route.
| Washington-Arlington-Alexandria, DC-VA-MD-WV | 2,370 | $99,910 +11% |
| New York-Newark-Jersey City, NY-NJ | 2,170 | $111,050 +23% |
| Boston-Cambridge-Newton, MA-NH | 1,590 | $111,750 +24% |
| Los Angeles-Long Beach-Anaheim, CA | 1,460 | $104,000 +15% |
| Seattle-Tacoma-Bellevue, WA | 1,220 | $107,580 +19% |
| Dallas-Fort Worth-Arlington, TX | 1,030 | $86,520 -4% |
| San Francisco-Oakland-Fremont, CA | 1,020 | $135,470 +50% |
| San Jose-Sunnyvale-Santa Clara, CA | 920 | $140,700 +56% |
| San Jose-Sunnyvale-Santa Clara, CA | 920 | $140,700 +56% |
| San Francisco-Oakland-Fremont, CA | 1,020 | $135,470 +50% |
| Santa Rosa-Petaluma, CA | 40 | $129,590 +43% |
We have no reported case of a named organisation automating this occupation. Not one deployment, not one announcement.
That is worth saying out loud next to a score of 23. The verdict above is about what the work exposes — what current AI could do to these tasks. It is not a claim that anyone has done it. For this occupation those two things have come apart completely: the capability argument is on this page, and the evidence column is empty.
Has AI actually changed your work? One tap, anonymous, and the running tally is public. Nothing else is asked of you.
Rather than check back: get the digest and we'll tell you what changed — or watch a single occupation from its own page.