The rules I’ve set up for one tab, “Job Search Blueprint,” leaked into “Reading List & Newsletter,” meaning, an answer was prompted in the Reading List tab that would’ve meant I got some list of pieces to read, but instead, the answer was in regards to a completely different tab (which has different rules, different expectations, different memories, etc.). There was nothing that prompted a different answer than the one that should logically follow from those particular rules in one conversation, however, the prompt, “let’s get the pieces for today” was completely misinterpreted (with no real reason) to mean “generate a list of jobs based on previous conversations,” etc. It followed the rules of my Job tab, reciting everything we’d been through in one tab, almost as though the memories of the Reading List conversation were overwritten completely.