Looking into it a little more, the official Jetpack support article says it only includes "the content of any public post." So, not like it's siphoning any private data, just like it's following your RSS feed.

Automattic's docs also now say they're winding down the Firehose service. Since it was around before the rise of LLMs, and the shutdown is really new, I wonder if that means the LLM use cases turned out to cause new problems for the company...