Why, it would give you a traceback if your worker process collects it, as a reasonable worker would. Since workers are probably non-interactive, they can't exactly receive a Ctrl+C. The dispatcher process, OTOH, will. I suppose a pool should kill all processes on SystemExit.
When you think about more details, you grow more and more sophisticated solution that best suits your needs. But the point of the post is different: it's a dead simple way to start. Starting is often the hardest part.
Collecting tracebacks is the sort of plumbing you would expect a 'batteries included' library to do. I didn't expect the workers to receive the ctrl-C, I'd expect the main process to terminate, along with its children; none of that happens.
I think it's clear that these sorts of details belong in the library. Everybody could hack together their own traceback collector, segfault handlers etc., but it's difficult to get right and in a scripting language you should be able to forget about such things.
When you think about more details, you grow more and more sophisticated solution that best suits your needs. But the point of the post is different: it's a dead simple way to start. Starting is often the hardest part.