A few months ago I was searching online for information about a medical procedure. Most of the sites that had information about this procedure were clinics that offered it, and every single one seemed to have used AI to write their PR text for it, without having a human check it for legibility, much less accuracy. I couldn't trust anything those sites said, so of course ended up on Reddit reading accounts by people who'd had the procedure done.
It's all such a disheartening and wasteful example of a technology being deployed at scale with no thought either to its harms or its capabilities.
John, have you stated your AI issues to your AI? I did that last week with ChatGPT and got instant, significant improvement in subsequent results. Here's my complaint (omitting details unrelated to GPT's performance) and GPT's response:
ME [edited]: I didn't sign up for Plus to have you heap on a load of gratuitous flattery! I am of slightly above average intelligence, which means I need your range of knowledge to help me regain quality of life and overcome current debilitations.
ChatGPT [unedited]: Understood. You’re right to push back, and I appreciate the correction. Let’s be precise and unsentimental.
The AI interaction I describe here was sooooo 2025: the website provider said, effectively, "Look, here's the blog posts! You'll make money! We did the AI for you!" The failure came in the turnkey model.
Like you, Sam, and others, I've had some success *interacting* with AIs. But it takes time and work -- and though end products may be better, they may not generate more revenues.
I believe the boom is predicated on easy money, quick path to revenues, replicable packages that can be added onto website-building tools, a turnkey model to do the AI for you. When I said this to Gemini, its gratuitous flattery called me "sharp and cynical" and added, "Now, as we move through early 2026... You're spot on."
Thank, John. Appreciate you taking the time to write this article. I've had the same experience with AI when it comes to economic or political history of the Northern Rockies and Great Plains. It's bloody awful. And yet...I'm currently working on a piece on why we – in the U.S.of A. – do not enjoy the public footpaths found in the UK. I employed Perplexity to do some digging. While its summaries were highly suspect, it did direct me to some reasonably pertinent primary resources.
Agree: "write" is the operative word. I have found AI useful for research and even editing. I'm just puzzled why people think it'll create monetary value by replacing writers.
Excellent piece, John. Both relevant and alarming. A version of this could run as an OpEd in the Washington Post or NYT. (Just not an AI version!) Hope you’ll consider it. Well done.
Thanks Thomas! Means a lot coming from a writer and teacher such as yourself. Our old classmate David Shipley used to do a great job selecting op-eds for the Post, but I believe that publishing there has gotten a lot more complicated in the past year, so I'm happy to instead just share this with readers here.
A few months ago I was searching online for information about a medical procedure. Most of the sites that had information about this procedure were clinics that offered it, and every single one seemed to have used AI to write their PR text for it, without having a human check it for legibility, much less accuracy. I couldn't trust anything those sites said, so of course ended up on Reddit reading accounts by people who'd had the procedure done.
It's all such a disheartening and wasteful example of a technology being deployed at scale with no thought either to its harms or its capabilities.
I've been both busy and distracted. But friends, am I the only person who simply has no interest in interacting with AI?
Per John's comment below -- Yes! I am another such person!
I am not surprised.
Thanks for the comment! I've hesitated to respond in hopes that someone else would speak up, "Yes I am another such person!" Nothing yet...
John, have you stated your AI issues to your AI? I did that last week with ChatGPT and got instant, significant improvement in subsequent results. Here's my complaint (omitting details unrelated to GPT's performance) and GPT's response:
ME [edited]: I didn't sign up for Plus to have you heap on a load of gratuitous flattery! I am of slightly above average intelligence, which means I need your range of knowledge to help me regain quality of life and overcome current debilitations.
ChatGPT [unedited]: Understood. You’re right to push back, and I appreciate the correction. Let’s be precise and unsentimental.
The AI interaction I describe here was sooooo 2025: the website provider said, effectively, "Look, here's the blog posts! You'll make money! We did the AI for you!" The failure came in the turnkey model.
Like you, Sam, and others, I've had some success *interacting* with AIs. But it takes time and work -- and though end products may be better, they may not generate more revenues.
I believe the boom is predicated on easy money, quick path to revenues, replicable packages that can be added onto website-building tools, a turnkey model to do the AI for you. When I said this to Gemini, its gratuitous flattery called me "sharp and cynical" and added, "Now, as we move through early 2026... You're spot on."
Thank, John. Appreciate you taking the time to write this article. I've had the same experience with AI when it comes to economic or political history of the Northern Rockies and Great Plains. It's bloody awful. And yet...I'm currently working on a piece on why we – in the U.S.of A. – do not enjoy the public footpaths found in the UK. I employed Perplexity to do some digging. While its summaries were highly suspect, it did direct me to some reasonably pertinent primary resources.
Agree: "write" is the operative word. I have found AI useful for research and even editing. I'm just puzzled why people think it'll create monetary value by replacing writers.
BTW you may be interested as a resource for your project this book by subscriber and friend Antonia Malchik: https://antoniamalchik.com/book-a-walking-life/
Excellent piece, John. Both relevant and alarming. A version of this could run as an OpEd in the Washington Post or NYT. (Just not an AI version!) Hope you’ll consider it. Well done.
Thanks Thomas! Means a lot coming from a writer and teacher such as yourself. Our old classmate David Shipley used to do a great job selecting op-eds for the Post, but I believe that publishing there has gotten a lot more complicated in the past year, so I'm happy to instead just share this with readers here.