This strategy seems like it would effectively choose the best candidates out of the pool labeled "desperate people".

It mostly categorically excludes anyone who has market desirability and doesn't have to put up with it.

I suppose it could work in parts of the country/world where the supply of engineers vastly exceeds the number of jobs, but it likely fails spectacularly in anything resembling a tech hub.

This strikes me as the stronger/more extreme version of take-home interview projects, which while being good at filtering for quality, also simultaneously selects only for people who are willing/forced to put up with it.

I suppose if your strategy is "hire the best of the desperates", this works for you.