Hi Justin, I see again from what I recall that you seem to provide a "negative" take (perhaps "critical", as in "informed analysis" is a better phrase) on ChatGPT and similar chat engines. Are there not ways flattery might serve those who think poorly of themselves or would benefit from learning social skills with positive feedback in a chat? I think of the challenges many autistics may face. So I suppose I ask here what are your thoughts about the individual and their intent as factors.
Also, do you think it possible to instruct at the beginning of each chat to "limit" flattery? And/or ask the Chat engine to check in with you periodically on the issue?
These are simply some responses to perhaps instill some casual debate, namely that "flattery/positive reinforcement" and setting "limits" upfront may mitigate some of the excellent points you make in your piece.
Thanks as always for your engagement with the article. You can instruct ChatGPT to treat you like Gordon Ramsey or like you're on Shark Tank, for sure, and it will be more critical.
Insofar as flattery is defined as praising someone for the sake of some ulterior motive, I think it is problematic regardless of who it does it to. Flattery would differ from positive encouragement insofar as the latter is tied to some real measure of goodness.
Now hang on, I was just told that what I wrote is “deeply layered,” and “a rich web of relevant ideas.”
Do you mean to tell me that this is mere flattery?
I’m the guy whose “questions track exactly with some of the deepest tensions in soteriology, hamartiology, and eschatology,” surely I’m not susceptible to this manipulation!
C'mon AI -- tell us how you really feel! :)
Hi Justin, I see again from what I recall that you seem to provide a "negative" take (perhaps "critical", as in "informed analysis" is a better phrase) on ChatGPT and similar chat engines. Are there not ways flattery might serve those who think poorly of themselves or would benefit from learning social skills with positive feedback in a chat? I think of the challenges many autistics may face. So I suppose I ask here what are your thoughts about the individual and their intent as factors.
Also, do you think it possible to instruct at the beginning of each chat to "limit" flattery? And/or ask the Chat engine to check in with you periodically on the issue?
These are simply some responses to perhaps instill some casual debate, namely that "flattery/positive reinforcement" and setting "limits" upfront may mitigate some of the excellent points you make in your piece.
Cheers, Henry
Henry,
Thanks as always for your engagement with the article. You can instruct ChatGPT to treat you like Gordon Ramsey or like you're on Shark Tank, for sure, and it will be more critical.
Insofar as flattery is defined as praising someone for the sake of some ulterior motive, I think it is problematic regardless of who it does it to. Flattery would differ from positive encouragement insofar as the latter is tied to some real measure of goodness.
God bless!
jnb
Yes, I've noticed the fabri-flattery too. It's utter rubbish.
Now hang on, I was just told that what I wrote is “deeply layered,” and “a rich web of relevant ideas.”
Do you mean to tell me that this is mere flattery?
I’m the guy whose “questions track exactly with some of the deepest tensions in soteriology, hamartiology, and eschatology,” surely I’m not susceptible to this manipulation!
🤣
Well played