What the work involves

Frontier labs need economists to tell them where a model's reasoning is actually wrong — not merely unpolished. Day to day, that means writing research-grade prompts drawn from your own subject area, producing the reference answer yourself, and grading model output against it. Expect a mix of applied micro and macro: identification strategy critiques, national accounts and price index mechanics, interpreting regression tables, monetary policy transmission, welfare analysis, and reading a model's attempt at an empirical design for the assumption it quietly skipped.

Much of the value sits in the rationale, not the label. A grade that says "incorrect" is nearly useless to a lab; a grade that says the model applied a difference-in-differences estimator where parallel trends is violated by the policy's own anticipation effects is training signal. Expect to write more prose than you might assume, and to have it reviewed.

What the platform screens for

  • Verifiable institutional history. At least a year in economics research at a recognised institution — a central bank, an IFI, a research department, or a graduate program. Named teams, papers, or working-paper series carry more weight than titles.
  • Depth that survives follow-up. The AI-led interview will pick one thing you claimed and push on it two or three layers. Breadth claimed across all of economics tends to collapse; a clearly bounded specialism does not.
  • Grading judgment. Can you separate a wrong answer from an answer that is right by a route you dislike, and can you say why in writing.
  • Honest availability. Ten hours weekly, sustained across four weeks minimum.

Logistics

Fully remote and asynchronous, with no fixed shifts — tasks are pulled from a queue and delivered against turnaround windows rather than a calendar. Pay in the $120–150/hr range has been observed for this brief; rates are set per project and per candidate, not guaranteed. Contributors who deliver clean, well-argued work are routinely routed to follow-on projects, which is where the engagement tends to become ongoing rather than a one-off month.