For those wondering what it's for: it's basically NumPy + a JIT compiler with standard Haskell syntax (you mostly just need to change the type signatures, not the code).
It can vectorize, parallelize on the CPU, or offload to the GPU automatically.
I've spent a lot of time wrapping my head around monads; whenever I thought I "got it," I would come across some exotic monad that completely blew my mind. The best way to understand them is not to rely on analogies but just follow the rules—everybody says that, but it took me a while to truly realize it.
Pretty cool to have a first-class tracing mechanism. Obviously... it's a monad! Haskell has had a MonadTrace monad for a long time, that can be switched on or off depending on your environment.
To add to sibling comment, if you have streaming data you have to update the whole index every time with r/kd trees whereas with H3 you just compute the bin, O(1) instead of O(log n).
Not rocket science but different tradeoffs, that’s what engineering is all about.
The big reason is that H3 is data independant. You put your data in predefined bins and then join on them, whereas kd/r trees depend on the data and building the trees may become prohibitive or very hard (especially in distributed systems).