Framing them as either good or bad (or would vs would not take) misses the point though. What we're discussing isn't a binary thing, it's an elasticity curve from something like (effort required / lucrativeness of job) to the willingness of the candidate to complete the assignment.
On that basis, I'd say there definitely are cases in which I would spend 50 hours on a take home assignment with even a 1% chance of success. There are also cases where I wouldn't spend five minutes on a take home assignment with a 100% chance of success.
If you're going to give take home assignments, the effort required should be in line with how amazing a job you're offering. If you're offering people $1mm/yr to work on <that thing you love>, putting a huge filter out in front is probably in your best interest. If you're trying to get someone to move to SF for $70k a year...
As a side note: if you're going to have a take home test, have the decency to make the scenario interesting and the answer illuminating.