Training cutoff
The date after which the model has read nothing, and the topics it has read least about.
Every model has a date past which it has seen nothing, and a distribution: it has read an enormous amount about popular topics and much less about your industry in your country. Ask about the thin end and the answer degrades smoothly instead of announcing that it has.
Prices, rules, rankings, ad policies and who runs a company all go stale, and nothing in the answer marks which part is old. For anything that changes, supply the current source yourself.
Find your way around
Learn it in the course, copy it from a playbook, see it work in a case study.