I loved reading the thread and digging into this detail of nostr! Here’s what GPT had to say.

👉 the problem: most clients only know how to render kind 1 as plain text + maybe inline images. They won’t automatically say “oh this is a gallery” just because of the k tag. Some might ignore it entirely.

👉The real role of k tags:

They aren’t a substitute for using the “right” kind. They’re a disambiguation tool in threads:

“This kind 1 note is replying to an event of kind 20, not just another kind 1.”

Without them, clients may assume everything you reply to is another plain text note.

So k is about reply context, not defining object type.