With Cats Effect 3.0 right around the corner, we’ve been publishing a lot of numbers and scenarios which demonstrate disorientingly high performance in a lot of realistic setups. At first glance, this seems to defy intuition. After all, IO is quite a heavyweight abstraction; how is it that it can be competitive with hand-written and hand-tuned code for the same use-case? There are a lot of ways to answer this question, and they get both complex and depend heavily on exactly what use-case yo...
