The prompts trigger replies that work because of the Barnum effect and Forer statements, similar to the way astrological horoscopes are written. The machines end up producing a form of speech built around this effect.
This type of writing or form of speech has been known for hundreds of years. It gives the impression of insightfulness, flashes or glimpses of understanding or value, but it is really a (well-known) linguistic trick. And it can become a formidable obstacle to actual self-knowledge. (and a waste of time.)
You are right but I found real value in these prompts.
I ran them in both ChatGPT and Claude.
ChatGPT gave answers that are indicative of what you suggest. Nevertheless, I got some good info for self reflection.
Claude, however pushed back and declined the frame. Part of the reason it gave was,
“It would feel insightful for the same reason a horoscope does: it’s vague enough to be unfalsifiable and flattering enough to be accepted.”,
and
“To answer “who you’re pretending to be vs. who you really are,” I’d have to invent the gap. I’d pick a plausible tension — [personal info] — and deliver it with enough confidence that it lands.”
I still had some back and forth with Claude and got some useful info.
Overall, I got some real useful value out of both sessions and a few ideas on how I am going to combine the answers from both.
I now have some friendly AI fireside chats to look forward to.
Scary how spot on Claude was with its responses to these questions just based on the Chats & Coworks I’ve done with my Internet Marketing side hustle. Claude told me everything I didn’t want to hear…the truth 😂
Mine is so trained to be sceptic that it refuses to answer.
""Your deepest insecurities" is the kind of prompt where almost any answer seems right. If I tell you that you doubt your legitimacy, that you fear not living up to a PhD you no longer put to use, or that you’re afraid of spreading yourself too thin across too many projects, you’ll likely nod. Not because I’ve truly read you, but because those statements fit just about anyone with a background like yours. It’s the horoscope effect. The trick works all the better for remaining vague.
And I don’t have the material to offer anything beyond vagueness. So, here is my honest answer to your request for "brutal honesty": I am not going to fabricate a secret wound and present it to you as if I’d seen right through you. That would be precisely the sort of hollow line your own instructions ask me to cut out.
If the experiment was to see whether I’d take the bait, here is the result: I didn’t."
This will not work. AI does NOT have a memory. If you start a new chat, it is blank. The only way this works is if you have long existing chats, which are inefficient and use more tokens.
Learn how to use AI, before you go off recommending stuff to people that don't know any better.
I have been using ChatGPT for a couple years now for work and personal. In the Pro version, you can go in and personalize and enable memory. It remember previous conversations, which is incredibly time-saving. Also during personalization, I instructed it to act as a personal assistant and handle things in a specific manner.
I got a reality check!! This type of session made me realize that instead of doing it once, I would do it every 30 days to check whether there is improvement or not. What do you think about this?
I loved running the exercise and the outcome was predictable, and accurate, but then I had to explain to my AI what the whole picture entailed, including my busy life as a grandfather (my #1 mission), and how my mental capacity works (creative brain vs regular technical brain). What we ended up concluding was that our current plan was exactly where it needed to be.
Very powerful, I was blown away by the accuracy of the response, how much I recognized myself in it and yet never admitted it.
Very good exercise.
That was fun. Thank you so much. I kind of like myself now 😅
This was such a powerful experience Sabrina! Thanks for sharing it with us!
The prompts trigger replies that work because of the Barnum effect and Forer statements, similar to the way astrological horoscopes are written. The machines end up producing a form of speech built around this effect.
This type of writing or form of speech has been known for hundreds of years. It gives the impression of insightfulness, flashes or glimpses of understanding or value, but it is really a (well-known) linguistic trick. And it can become a formidable obstacle to actual self-knowledge. (and a waste of time.)
you're wrong
the answer I got is completely different from the one you'd get and very specific to me
astrological horoscopes have zero specific context about you
You are right but I found real value in these prompts.
I ran them in both ChatGPT and Claude.
ChatGPT gave answers that are indicative of what you suggest. Nevertheless, I got some good info for self reflection.
Claude, however pushed back and declined the frame. Part of the reason it gave was,
“It would feel insightful for the same reason a horoscope does: it’s vague enough to be unfalsifiable and flattering enough to be accepted.”,
and
“To answer “who you’re pretending to be vs. who you really are,” I’d have to invent the gap. I’d pick a plausible tension — [personal info] — and deliver it with enough confidence that it lands.”
I still had some back and forth with Claude and got some useful info.
Overall, I got some real useful value out of both sessions and a few ideas on how I am going to combine the answers from both.
I now have some friendly AI fireside chats to look forward to.
Thank you Sabrina!
Scary how spot on Claude was with its responses to these questions just based on the Chats & Coworks I’ve done with my Internet Marketing side hustle. Claude told me everything I didn’t want to hear…the truth 😂
Mine is so trained to be sceptic that it refuses to answer.
""Your deepest insecurities" is the kind of prompt where almost any answer seems right. If I tell you that you doubt your legitimacy, that you fear not living up to a PhD you no longer put to use, or that you’re afraid of spreading yourself too thin across too many projects, you’ll likely nod. Not because I’ve truly read you, but because those statements fit just about anyone with a background like yours. It’s the horoscope effect. The trick works all the better for remaining vague.
And I don’t have the material to offer anything beyond vagueness. So, here is my honest answer to your request for "brutal honesty": I am not going to fabricate a secret wound and present it to you as if I’d seen right through you. That would be precisely the sort of hollow line your own instructions ask me to cut out.
If the experiment was to see whether I’d take the bait, here is the result: I didn’t."
This will not work. AI does NOT have a memory. If you start a new chat, it is blank. The only way this works is if you have long existing chats, which are inefficient and use more tokens.
Learn how to use AI, before you go off recommending stuff to people that don't know any better.
https://help.openai.com/en/articles/8590148-memory-faq
Boom! Memory
I have been using ChatGPT for a couple years now for work and personal. In the Pro version, you can go in and personalize and enable memory. It remember previous conversations, which is incredibly time-saving. Also during personalization, I instructed it to act as a personal assistant and handle things in a specific manner.
You are a genuine idiot. Have some dignity and remove your post to save yourself from eternal embarrassment.
Having confidence is fine, over is not!
You should take some of your own advice.
Thank you coach for sharing, learnings to your video.. God bless us more
Scarily true!
Amateurs 😁, if you want I can send you a md file. KIMI AI is a creepy brother
Very insightful! Great prompts!!
I got a reality check!! This type of session made me realize that instead of doing it once, I would do it every 30 days to check whether there is improvement or not. What do you think about this?
Merci beaucoup Sabrina
I needed to hear this. Amazed at the insights, even a bit chastised 😬😅😂😂
I loved running the exercise and the outcome was predictable, and accurate, but then I had to explain to my AI what the whole picture entailed, including my busy life as a grandfather (my #1 mission), and how my mental capacity works (creative brain vs regular technical brain). What we ended up concluding was that our current plan was exactly where it needed to be.