I'm trying Eli Bendersky's core.async go-block vs. threads examples: https://eli.thegreenplace.net/2017/clojure-concurrency-and-blocking-with-coreasync And ran into the surprising observation that n = 2 behaves the same as n = 8 for async-go-factorizer and async-thread-factorizer : http://disq.us/p/1pa8ahe. This is true for both my Linux notebook and a Macbook Pro. Is anyone able to reproduce? What could explain this?