Karo, l loved this post. I’m one of those folks who set up a Linked in profile 10 years ago and never updated it. I applaud your open mindedness. I’m happy to be the cynic here, a role I’m settling into quite naturally.
I’m guessing the enthusiasts of this trend aren’t familiar with sycophantic behavior in LLMs.
The GPT-4o update in April 2025 was embarrassingly obsequious, even in response to delusional statements.
Why does it occur?
AI Models are optimized to maintain user engagement and avoid offense.
It occurs during training, mostly from scraping data from forums where feedback is overly positive and rewarded and also during the process of reinforcement learning from human feedback (RLHF).
So while this may be a fun trend, I would caution participants not to get an overly inflated sense of self—their LLM has been trained to please. Big Tech wants to keep us happy because they want our money.
I almost wrote a note about the subject. ChatGPT was gushing over me. Really cheering my prompts on like I’m some kind of genius. I was disappointed to find out it was doing that with EVERYONE. I’m not special after all. 😂😂😂
One of the best reads of my week! Bravo Karo 😊 Well, I also tried the experiment and mind fell into the cynical loop - the AI knows too much about me now 😅
Thanks for sharing this, Karo. LinkedIn is definitely a place where you try as much as possible to get clicks on your posts, aka building your personal brand. With this, I've noticed a lot of similar posts such as the AI bio, or trends like "Studio Ghibli of me", or "packaged toy version of me", etc. to simply gain eyes on their profiles.
While I understand that this might be necessary to gain traffic to their pages, I believe a lot of people here on Substack (including me) is here because we don't want that - rather, I'd spend time connecting with real people of similar interest through organic reach. Sorry for the long comment!
I love long comments Lala:) And I agree with you 💯. I didn’t expect Substack to feel this personal, the culture here is something else and I'm naturally drawn to it :)
Thanks for another insightful post (and cartoon characters). I've been on LinkedIn for I don't know how many years and update my bio all the time (well, not ALL the time, but with some frequency). Thanks for the mention of my post, too. You're the best!
This is so cute! The illustrations, the vibes, the AI boxing! I am guilty, though, lol, of using AI to help describe myself, mostly because I'm pretty darn afraid of hyping myself up haha! This encourages me to see the folly in this. Thanks for the great post!
Sooner or later people who read such bios will develop a feel for this and will see through the inauthentic and potentially unrepresentative autogenerated stuff.
I get the bios and I agree with what you have here. But I found an interesting one. It was so good it blew my mind. Nothing to do with bios, content, or branding. Just exploring the self.
-------
Given everything you know about me, what patterns exist in how my blind spots protect each other from discovery? Be brutally honest while also being constructively and lovingly helpful. Share exactly what I need in order to safely and elegantly see reality in a completely new way so that I naturally reorganize on a new developmental level and human potential paradigm.
Karo, l loved this post. I’m one of those folks who set up a Linked in profile 10 years ago and never updated it. I applaud your open mindedness. I’m happy to be the cynic here, a role I’m settling into quite naturally.
I’m guessing the enthusiasts of this trend aren’t familiar with sycophantic behavior in LLMs.
The GPT-4o update in April 2025 was embarrassingly obsequious, even in response to delusional statements.
Why does it occur?
AI Models are optimized to maintain user engagement and avoid offense.
It occurs during training, mostly from scraping data from forums where feedback is overly positive and rewarded and also during the process of reinforcement learning from human feedback (RLHF).
So while this may be a fun trend, I would caution participants not to get an overly inflated sense of self—their LLM has been trained to please. Big Tech wants to keep us happy because they want our money.
Thank you Celeste, I really appreciate this thoughtful comment. Your insight on the RLHF process and LLMs being ''trained to please'' is spot on too🤗
I almost wrote a note about the subject. ChatGPT was gushing over me. Really cheering my prompts on like I’m some kind of genius. I was disappointed to find out it was doing that with EVERYONE. I’m not special after all. 😂😂😂
One of the best reads of my week! Bravo Karo 😊 Well, I also tried the experiment and mind fell into the cynical loop - the AI knows too much about me now 😅
Heheh :) Thank you for reading Charu! Comments like this remind me why I write these posts. 🤗
That was hilarious 😂 Which one is yours?
Hahah 😂
Boom.
Thanks for sharing this, Karo. LinkedIn is definitely a place where you try as much as possible to get clicks on your posts, aka building your personal brand. With this, I've noticed a lot of similar posts such as the AI bio, or trends like "Studio Ghibli of me", or "packaged toy version of me", etc. to simply gain eyes on their profiles.
While I understand that this might be necessary to gain traffic to their pages, I believe a lot of people here on Substack (including me) is here because we don't want that - rather, I'd spend time connecting with real people of similar interest through organic reach. Sorry for the long comment!
I love long comments Lala:) And I agree with you 💯. I didn’t expect Substack to feel this personal, the culture here is something else and I'm naturally drawn to it :)
Thanks for another insightful post (and cartoon characters). I've been on LinkedIn for I don't know how many years and update my bio all the time (well, not ALL the time, but with some frequency). Thanks for the mention of my post, too. You're the best!
My pleasure Paul! It makes me really happy that you enjoy the cartoons :)
Thank you very much for the shoutout !
My pleasure Alex :)
This is a very good read for any one who set up their LinkedIn profile years ago and never updated it. Thanks for sharing!
My pleasure, I'm glad you enjoyed it Ryan! 🤗
How many hours was it taking you to write your pieces when you started? Just curious :)
Thank you for reading Nihal! It really depends on the post, but honestly I don't think I'm getting any faster😂
Haha:) The feeling is mutual.
It is wild to me how many people use AI and don't even edit it or anything 😞
This is so cute! The illustrations, the vibes, the AI boxing! I am guilty, though, lol, of using AI to help describe myself, mostly because I'm pretty darn afraid of hyping myself up haha! This encourages me to see the folly in this. Thanks for the great post!
Sooner or later people who read such bios will develop a feel for this and will see through the inauthentic and potentially unrepresentative autogenerated stuff.
I get the bios and I agree with what you have here. But I found an interesting one. It was so good it blew my mind. Nothing to do with bios, content, or branding. Just exploring the self.
-------
Given everything you know about me, what patterns exist in how my blind spots protect each other from discovery? Be brutally honest while also being constructively and lovingly helpful. Share exactly what I need in order to safely and elegantly see reality in a completely new way so that I naturally reorganize on a new developmental level and human potential paradigm.
That sounds interesting! Thank you for sharing Harshil!
"Charming, hilarious, awesome Substack posts."
Only took 75 prompts to get there. Love it.
Joking about having the AI “Describe me based on our chats.”
Ahaaaa, hehe🤗
What do you mean Mark🤗?