Re: So they won't be using a
"Without knowing how many "persons of interest" were in the initial sample - which, pretty much by definition, they can't know - we still don't know yet how good the tool is."
If it was any good, you'd probably know it.