> “…small, genuine touches…not flattery. Internal tests show no rise in sycophancy…” While I’ve not experienced this new UX yet, I appreciate what the marketin

> “…small, genuine touches…not flattery. Internal tests show no rise in sycophancy…” While I’ve not experienced this new UX yet, I appreciate what the marketing team did in this tweet—-careful wording that adds to the Framing of artificial personalities (small moves, honest[1], not flattering, not sycophantic). I’m not using ChatGPT much, but I have used it as a study aid while reading a book on JS/React. When I was confused about something or encountered a gap in my understanding of the text I noticed the first few words from my Chat were doing work to tell me if I was on point, or in the realm, or set me up for correction (even if I’m “on point”, I do _continue reading_). I think of these small moves like _map orientation_. You can’t effectively use a map on the ground until you align the map to the territory. Do you see? I encounter artificial phone personalities and human CSRs frequently at work, and their sycophantic scripts enrage me because I’m trying to get something done—pay a bill, request human support. Adding emotional queues are either slowing me down or meant to manipulate my emotional reactions (generating the opposite reaction). [1]: one caveat, I do have a small problem with their use of the word “genuine”. I don’t believe artificial personality can possess “authenticity”, which is associated meaning of genuine. I don’t much like “honesty” either, but it’s closer to the point. I do appreciate a cue to how correct was my input.