i mean, it is pretty much what I expected it to mean. It's a macro for semantic instructions. I didn't really see any LLM "bullshit", just a way to macro using an LLM...
Not the original commenter but I expected a semantic macro to be something that expands on syntax-aware macros (in languages like Scheme or Rust) into doing things based on the semantics of the code. Not sure exactly what that would entail but I was intrigued. Instead it turned out to be just quick buttons for sending something to an LLM. Between Github's Copilot, ChatGPT and LLM plugins for Obsidian those needs have already been covered for a long time.
Did you expect this to be something different?