> I think this is mostly due to its multiprocess architecture vs just sharing in memory plans between threads
You can share stuff with a multiprocess architecture just fine (either through IPC or just plain shared memory + synchronization)
It's true that threads share memory by default, but processes can opt into sharing memory if they wish. And it appears that Postgres already makes use of shared memory for some things
PG shares data between process not code as far as I know and definetly not any plans. Sharing jitted code is not straight forward as pointers will be different per process.