@thheller So with the Service Worker hack I got blocking semantics working and added a derefable future and ported injest's auto-parallelizing =>> thread last macro to it and I'm seeing between 2 to 5 times performance increases over a single thread for heavy work loads, even while serializing the work across all the workers. Got a new update showing the latest https://clojureverse.org/t/in-mesh-one-step-closer-to-threads-on-the-web/9069/2?u=john_newman. You might also like the yield thing.