May 6, 2026
Librarian + Analyst + Superfan
Tomorrow, I’m running a live workshop designed to show expertise-based business owners (e.g., coaches, consultants, advisors, and other thought-leader-y types) how to take the giant pile of content they have created over the years and “mine” it for output like:
- A catalog of existing themes
- A list of named/nameable frameworks
- Any gaps that could be filled with new content (e.g., their next book!)
Unfortunately, registration is closed at this point, but here’s why I’m bringing it up:
In preparation for the event, I’ve run myself through the process repeatedly to get a sense of:
- Timing (e.g., “How long does it take Sonnet 4.6 to answer a question on a project that contains 1200 files?” )
- Token limits (e.g., “Will attendees blow through their token limit on the first query if they use Opus 4.7 on 10MB of text files?”)
As a side effect of running myself through this, I’ve been blown away by what the LLM digs up and presents to me.
It’s like having a librarian + analyst + superfan all rolled into one.
And since it’s my content, I can immediately validate the output as good, bad, or irrelevant (and it’s mostly good).
Here’s the thing...
AI still stinks at authoring.
But that’s fine!
You can have it draft outlines and pull verbatim quotes and point out unexplored areas that you can then use to write yourself and continue growing your giant pile of content.
Yours,
—J