This is a shameless plug for some AI integrations I am working on atm., not really datahike related (yet). Feedback welcome https://github.com/simm-is/repl-mcp
I do think that the whole coding workflow could actually be done against datahike and not the file system once coding assistants execute it. I am thinking about different ways to use Clojure's memory model to speed up explorative coding that way.