Evidence
Measured outcomes or structured forecasts from a linked source. We name the population, date, and important limitation.
Neither hype
nor retreat.
A field guide to AI, robotics and work—for people starting out, changing course or leading teams.
Why grey?
The loudest technology stories are binary: acceleration or collapse, automation or immunity, human or machine. Work is not.
A tool’s capability does not by itself determine who gets hired, who earns more or whose workload rises. Adoption, demand, costs, institutions and bargaining power also matter. We look at that wider picture, then suggest small ways to test what it means for your work.
Our recommendations cannot guarantee job security. Personal adaptation is only part of the response; employer training, working conditions and public policy matter too.
Editorial method
Clear advice. Traceable evidence.
Practical steps are our recommendations. Research claims carry numbered citations; expandable source notes show findings, methods and limitations. We distinguish measured evidence, early signals and our own judgment without putting a label before every paragraph.
Signal
An early pattern worth watching: useful for attention, too young or narrow for a settled conclusion.
Judgment
Our synthesis and recommendation. It should be testable in the world, but it is ours—not something a citation can decide.
Standards
What we owe the reader.
- Date the evidence.Every briefing has a publication date and source-review date. Source notes distinguish when research was published from the period it studied.
- Prefer primary material.Research papers, public statistics, and reporting organizations come before summaries.
- Keep categories honest.Exposure is not displacement. A forecast is not an outcome. A benchmark is not a job.
- Name source incentives.Vendor research can be valuable; we label it and avoid treating it as neutral.
- Show the counterevidence.We include null and negative results, uncertainty and newer follow-ups. A selected review is not a claim to cover every paper.
- State the geography.Our employment projections are U.S. national figures; other studies cover global exposure, Denmark or particular firms. Check local pay, demand and entry requirements before making a career decision.
- Label the tools.Our task planner uses transparent editorial rules, not a validated risk model. Workflow examples are illustrations, not measured case studies. Automatically collected feed items are separate from reviewed sources.
- End in useful agency.Recommendations should create a next move without pretending everyone has equal room to maneuver.