“There are Subreddits for folks who are choosing to date AI, complete with (AI-produced) images of them wining and dining humanized renderings of their partners.”
WUT. JESUS.
The center Ace speaks loudest to me. That’s a beautiful image. And it is totally on point. I have a launch this Tuesday and more or less launched a different thing via a pilot group tonight!
I have never used ChatGPT so appreciate this research you did so I don’t have to. I also have no idea how to access it. 😂 Yay for four years!! Congrats, Caroline!
People say it's undeniably helpful at proofreading and the like, so maybe it will prove useful to us one day, but in the meantime, I'd say you're not missing much. 🙃 And thank you! I always value and appreciate your presence here.
Wow. That is both genius and terrifying, and I’d love to see how it’d go. Given their “personalities,” I imagine they’d hate each other, but maybe opposites attract?
It took me a long, long time to dip my toes into the chatbot world also. I copied and pasted a newsletter I was working on directly into ChatGPT without asking anything, and was duly shocked when seconds later I received reams of advice about sentences and sections the bot thought needed cleaning up. I would never use it to *write* (what would be the point?), but I confess I did take some of the editing advice it gave me! There were certain suggestions that I would never use, because it wouldn't sound like me if I did. But there were other points I recognized as being good advice.
Another thing I did, which a friend suggested and which I haven't even looked at yet because I went away on a trip right after doing it (a trip I'm still on): I gave both ChatGPT and Claude THE ENTIRE FOUR YEARS of my newletter writing by giving them to URL to my Substack. They both came back with ideas, themes, strengths, weaknesses---though mostly strengths because as you mention, they seem to like to flatter. It was overwhelming. I printed it all out, and it's waiting at home for me to peruse when I get back. I have to say, getting this degree of feedback is awfully seductive.
I hear you. While I'd never want feedback where style is concerned (I'm already particular about notes from human editors!), part of me does wonder if AI might be helpful with the macro/business side of things. Marketing suggestions, for example. Or publishing strategies I wouldn't think of. While I have friends who work as proofreaders, copyeditors, copywriters, research fellows, etc., and I very much see the value in having humans perform these jobs, it is seductive to get automated help for those of us who can't afford to employ someone (AKA me). All of which is to say, if you're up for sharing, I'd love to hear your findings when you go through that printed feedback!
Okay, I've been home for a couple of weeks and I've finally looked at my secret stash of AI feedback. Turns out I didn't correctly remember what I'd done--OOPS. I do have lengthy evaluations of the intensive journaling workshop I created from both Claude and ChatGPT, but it looks like my entire Substack publication was too much for them to handle. I had nothing printed out for that from ChatGPT, and a brief two pages from Claude that starts out with "I can see your Substack landing page, but I'm only getting limited content due to JavaScript requirements."
I subsequently tried putting the entire publication into ChatGPT, but ChatGPT wasn't having it. It asked for individual posts, and I gave it my last three. Overwhelmingly I felt like: if my newsletter were as good as ChatGPT says it is, I'd have a lot more subscribers. I wish now I'd saved what it said, but apparently my writing is clear and the tone strikes a perfect balance, my voice is authentic and I strike the right balance of humility and authority. Wow! I'm the greatest.
This was all straight out of the gate, mind you... I realize that you need to tailor your questions to get better answers.
As for the journaling workshop that got evaluated, what I do find valuable is the clinical dissection of each of its parts, its structure, its tone, and ideas for ways to do things differently. While I'd never use an AI idea verbatim (they reek of robot), I find it helpful to have my eyes opened to different possibilities. And I'd never have such a clear-eyed view of the thing as a whole... I think because I'm just too close to it.
As you say in your piece above, "the flattery, the availability, the absolute deference can be appealing, even intoxicating." So the trick is to stay sober with this stuff.
I will most definitely report back when I'm home. I'd actually forgotten I'd done that till I read your piece. I printed it all so that I could close the offending tabs on my computer (what if my husband saw? 😅 Seriously, I felt guilty) and I squirreled it away... at the same time putting it right out of my mind I guess.
I’m tempted to do this too because so many writers swear it’s effective and because I am terrified of being left behind / not using all the tools available to me to improve. But I keep resisting because I don’t want to feed the LLMs any more free training material. maybe I’m naive and the fact that my writing’s online already means it’s been fed to the machine (despite me toggling the substack “opt out” button). Would love both of your thoughts on this quandary!
Hi Isabel! I didn't have the feeling of being left behind, per se. And I guess I don't feel like I'm giving the machine anything it doesn't already have. But I *was* starting to feel like I was a fool not to try it out and see what all the fuss was about. In general AI scares the shit out of me, and it's making people dumber by the minute. I don't think using it for editing is helping me "improve", exactly... if I were truly improving, I'd see those areas that need tightening up before ChatGPT had to point them out to me. And maybe I *would* eventually make those edits myself if I worked on my piece long enough, but ChatGPT showed them to me in the blink of an eye. That's the irresistable part, for me. I feel sure I will sin again.
So much to comment on…in this eerie and evolving world.
But most important is this:
“This past week marks four years of writing this newsletter. ❤️ My heartfelt thanks for being here,”
I’m sure I speak for many of us when saying: ‘we’re glad to be here.’ 🙏
Thank you so much, James! I’m so glad to have met you, and so value your warm and thoughtful presence here.
Hear, hear.
Or definitely, definitely.
“There are Subreddits for folks who are choosing to date AI, complete with (AI-produced) images of them wining and dining humanized renderings of their partners.”
WUT. JESUS.
The center Ace speaks loudest to me. That’s a beautiful image. And it is totally on point. I have a launch this Tuesday and more or less launched a different thing via a pilot group tonight!
I LOVE that image. One of my favorite decks. It sounds like you have a lot of exciting things at play!
SO happy you are real (and have been for 4 years) and not a chatbot. ❤️
Thank you. ❤️ Human through and through! And grateful to have met you.
I have never used ChatGPT so appreciate this research you did so I don’t have to. I also have no idea how to access it. 😂 Yay for four years!! Congrats, Caroline!
People say it's undeniably helpful at proofreading and the like, so maybe it will prove useful to us one day, but in the meantime, I'd say you're not missing much. 🙃 And thank you! I always value and appreciate your presence here.
You made me laugh with the Colonial Williamsburg quip!
Putting aside the water usage I wonder if it's possible to have ChatGPT and Gemini pretend to go out on a date with each other.
Wow. That is both genius and terrifying, and I’d love to see how it’d go. Given their “personalities,” I imagine they’d hate each other, but maybe opposites attract?
Huge congratulations on 4 years! Here's to many more of real human writing. 💖
Caroline, congratulations! Thank you for your wonderful newsletter. ❤ 🤗
Thank you, Yvonne! ❤️
Just thought I would pop in and say I don't know how to access chat gpt or any other AI either. You are not alone.
Also, I have learned that it is arrogant and condescending to assume that everyone knows what I/we do....learned the hard way....
Je geeft me vertrouwen en schrijft prettig. Heb ervan geleerd en wil graag de volgende nieuwsbrief ontvangen.
Groet, Gustaaf
Thank you, Gustaaf! I'm glad you enjoyed it and hope you'll read the next one, too. :)
Wonderful newsletter as always. Thank you Caroline :)
Thank you so much, Clem.
'Things take the time they take. Don't worry' are just exactly the words I needed to hear today ❤
wow, bringing back Tamagotchi memories! And I love those interpretations of the same card, fascinating!
Loved every second of this!! 💛 Such a good read, Caroline ✨
Thank you, Meg! ☺️
It took me a long, long time to dip my toes into the chatbot world also. I copied and pasted a newsletter I was working on directly into ChatGPT without asking anything, and was duly shocked when seconds later I received reams of advice about sentences and sections the bot thought needed cleaning up. I would never use it to *write* (what would be the point?), but I confess I did take some of the editing advice it gave me! There were certain suggestions that I would never use, because it wouldn't sound like me if I did. But there were other points I recognized as being good advice.
Another thing I did, which a friend suggested and which I haven't even looked at yet because I went away on a trip right after doing it (a trip I'm still on): I gave both ChatGPT and Claude THE ENTIRE FOUR YEARS of my newletter writing by giving them to URL to my Substack. They both came back with ideas, themes, strengths, weaknesses---though mostly strengths because as you mention, they seem to like to flatter. It was overwhelming. I printed it all out, and it's waiting at home for me to peruse when I get back. I have to say, getting this degree of feedback is awfully seductive.
I hear you. While I'd never want feedback where style is concerned (I'm already particular about notes from human editors!), part of me does wonder if AI might be helpful with the macro/business side of things. Marketing suggestions, for example. Or publishing strategies I wouldn't think of. While I have friends who work as proofreaders, copyeditors, copywriters, research fellows, etc., and I very much see the value in having humans perform these jobs, it is seductive to get automated help for those of us who can't afford to employ someone (AKA me). All of which is to say, if you're up for sharing, I'd love to hear your findings when you go through that printed feedback!
Okay, I've been home for a couple of weeks and I've finally looked at my secret stash of AI feedback. Turns out I didn't correctly remember what I'd done--OOPS. I do have lengthy evaluations of the intensive journaling workshop I created from both Claude and ChatGPT, but it looks like my entire Substack publication was too much for them to handle. I had nothing printed out for that from ChatGPT, and a brief two pages from Claude that starts out with "I can see your Substack landing page, but I'm only getting limited content due to JavaScript requirements."
I subsequently tried putting the entire publication into ChatGPT, but ChatGPT wasn't having it. It asked for individual posts, and I gave it my last three. Overwhelmingly I felt like: if my newsletter were as good as ChatGPT says it is, I'd have a lot more subscribers. I wish now I'd saved what it said, but apparently my writing is clear and the tone strikes a perfect balance, my voice is authentic and I strike the right balance of humility and authority. Wow! I'm the greatest.
This was all straight out of the gate, mind you... I realize that you need to tailor your questions to get better answers.
As for the journaling workshop that got evaluated, what I do find valuable is the clinical dissection of each of its parts, its structure, its tone, and ideas for ways to do things differently. While I'd never use an AI idea verbatim (they reek of robot), I find it helpful to have my eyes opened to different possibilities. And I'd never have such a clear-eyed view of the thing as a whole... I think because I'm just too close to it.
As you say in your piece above, "the flattery, the availability, the absolute deference can be appealing, even intoxicating." So the trick is to stay sober with this stuff.
I will most definitely report back when I'm home. I'd actually forgotten I'd done that till I read your piece. I printed it all so that I could close the offending tabs on my computer (what if my husband saw? 😅 Seriously, I felt guilty) and I squirreled it away... at the same time putting it right out of my mind I guess.
I’m tempted to do this too because so many writers swear it’s effective and because I am terrified of being left behind / not using all the tools available to me to improve. But I keep resisting because I don’t want to feed the LLMs any more free training material. maybe I’m naive and the fact that my writing’s online already means it’s been fed to the machine (despite me toggling the substack “opt out” button). Would love both of your thoughts on this quandary!
Hi Isabel! I didn't have the feeling of being left behind, per se. And I guess I don't feel like I'm giving the machine anything it doesn't already have. But I *was* starting to feel like I was a fool not to try it out and see what all the fuss was about. In general AI scares the shit out of me, and it's making people dumber by the minute. I don't think using it for editing is helping me "improve", exactly... if I were truly improving, I'd see those areas that need tightening up before ChatGPT had to point them out to me. And maybe I *would* eventually make those edits myself if I worked on my piece long enough, but ChatGPT showed them to me in the blink of an eye. That's the irresistable part, for me. I feel sure I will sin again.
hehe did I say you were sinning?? I think I’ve got a lot of limiting beliefs that are coming into focus here, actually :)
No! *You* didn't say I was sinning... I did! That's 100% my choice of words :)