Switch back to subprocessing for llama.cpp

This should resolve a number of memory leak and stability defects by allowing
us to isolate llama.cpp in a separate process and shutdown when idle, and
gracefully restart if it has problems.  This also serves as a first step to be
able to run multiple copies to support multiple models concurrently.

llm/llm_darwin_arm64.go

@@ -0,0 +1,8 @@
+package llm
+
+import (
+	"embed"
+)
+
+//go:embed build/darwin/arm64/*/bin/*
+var libEmbed embed.FS