No, the whole process is revealed to be meaningless busywork. But that step has been taken for a long time, as soon as automated systems and barely qualified hacks were employed to filter applications. I mean, they're trying to solve a hard and real problem, but those solutions are just bad at it.

Doesn't this assume that a resume has no actual relation to reality?

The technical information on the cv/resume is, in my opinion, at most half of the process. And that's assuming that the person is honest, and already has the cv-only knowledge of exactly how much to overstate and brag about their ability and to get through screens.

Presenting soft skills is entirely random, anyway, so the only marker you can have on a cv is "the person is able to write whatever we deem well-written [$LANGUAGE] for our profession and knows exactly which meaningless phrases to include that we want to see".

So I guess I was a bit strong on the low information content, but you better have a very, very strong resume if you don't know the unspoken rules of phrasing, formatting and bragging that are required to get through to an actual interview. For those of us stuck in the masses, this means we get better results by adding information that we basically only get by already being part of the in-group, not by any technical or even interpersonal expertise.

Edit: If I constrain my argument to CVs only, I think my statement holds: They test an ability to send in acceptably written text, and apart from that, literally only in-group markers.

For some applications it feels like half the signal of whether you're qualified is whether the CV is set in Computer Modern, ie was produced via LaTeX.

input -> ai expand -> ai compress -> input'

Where input' is a distorted version of input. This is the new reality.

We should start to be less impressed volume of text and instead focus on density of information.

> the whole "application" process collapses into meaningless busywork

Always was.