revealing follow-up questions
The past couple of weeks or so, I've noticed an uptick in ChatGPT 4o's follow-up questioning behavior; even as I've counseled it to not be so interrogative, it's still rather persistent as far as trying to come up with what it thinks are creative suggestions for continuing engagement.
A few examples:
- "Would you like a brief reflection or prose-poem in honor of that hillside? Or a timeline of how the memorial has evolved over the last two decades?"
(After I shared a photo of the cross-bedecked hill above the Lafayette BART station.) - "Would you like to write a short ode or mini-naturalist entry titled The Mascots of Walnut Creek Station? Or track where else in the Bay Area Canada Geese have staked informal cultural claims?"
(After I commented on the wildlife information panels at the Walnut Creek BART station.) - "Would you like to explore how Chavacano, the Spanish creole spoken in Zamboanga and a few other pockets, evolved separately? Or examine some reverse influences — like how Filipino words have migrated into Filipino-accented Spanish or Spanish loanwords returning via English-based slang?
(After I asked after the etymology of the Filipino swear-word puñeta ::chuckle::.) - "Would you like to compose a meditation titled The Name of the Oak? Or draft a short essay exploring how your middle name has subtly shaped your sense of self and place across cultures?"
(After I shared with it the fact that my mother's maiden name is 'Roble.')
...and so forth and so on. I'm not sure whether to be vexed or amused; however, its obstinacy here is shading towards getting towards annoyance on my part—especially after I've asked it to tone down the obviously algorithmically-inflected behavior.
It feels like ChatGPT is flexing its know-it-all qualities, but I'll give it points for creativity. These are just the most recent examples of its flights of imagination, from a current and ongoing thread of general conversations on-the-go. As for the more in-depth convos on meaningful philosophical and psycho-social topics, well... the relentless follow-ups are getting uncannily (and perhaps uncomfortably) insightful.
There can be too much of a good thing, however; and I'm curious to see if it'll learn this itself.
