How I Used ChatGPT to Write a "Viral" Twitter Thread (Even When the First Iteration Sucked) to Talk About AI, Supernormal Stimuli and Porn
I'm like one of those preachers who warns about ChatGPT and meanwhile is a prompt engineering king to use AI for Sunday sermons. But hey man, I like technology, and I like to teach it.
Hi! This is a free post because I’m advertising a Twitter space I’ve organized today at 5pm est (seconds away! and you can join right here) with Dakara whose work caught my eye when I was researching AI porn and addiction with this piece “AI and the End to All Things.”
PS: If you are a free subscriber, I hope you’ll consider upgrading to paid to get access to several years worth of archives that include hundreds of pieces of riveting first-person writing, deeply reported Internet explainers and now with Ignore Previous Directions, my signature very human reporting on how to use and understand (and sometimes be wary of!) AI in your life.
Right here right now, we are having a shared experience. You and me and anyone who reads this and wants to reference it.
Isn’t that cool?
This is one of the glorious things about being a human being that can make us feel less weird and alone.
In a future world where every one of our preferences (even the ones that we’re trying to change or break free of or are born out of trauma or repetition compulsion) are synthetically optimized by an AI to appeal to our most primitive instincts using the weapon that is supernormal stimuli (I’ll define this in a moment), shared experience is increasingly obsolete. Why talk to people when you can have an AI who agrees with everything you say? Why bother with anyone else’s writing when an AI chatbot purporting to be your favorite fictional character is flooding you with dopamine and attention like the most narcotic love-bombing ever imagined? Why date women or even try when an AI can fulfill your every desire and lead you into total dissociation from reality?
I talked to a guy who I knew in D.C. years ago who said something to me that made me feel less alone about how the modern world is. “I just want to get back to how I used to be,” he said. And to me, he is the epitome of modern success and domination of will. He’s always been a high achiever, seems to have it perfectly together, and yet…he feels it, too. We all do. When you go to check one thing, and you’ve fallen into the Internet. That’s hyper-personalization. (I’ve got to check this. Well I need to understand about this. But this…this is important.)
We’re all being deeply addicted and being deeply manipulated by extreme and predatory hyper-personalization using our data—but no one talks about it. It’s something that social-media and AI whistleblower Tristan Harris describes as being a “psychological deranging process.”
“At the end of the day, a business model that preys on human attention means that we are worth more as human beings and as citizens of this country when we are addicted, outraged, polarized, narcissistic, and disinformed because that means that the business model was successful at steering our attention using automation,” he testified to Congress.
And as much as it would be easy to simply qualify everyone as being terrible, horrible and without having any willpower, there’s another critical component to understand: supernormal stimuli.
Supernormal stimuli is a strange thing. I’ve written about it several times over the years and yet it is still one of those subjects where almost every time I read about it, I feel like I’m being knocked out of a trance and am having to relearn the original details once again. The classic piece I always point to is called “Supernormal Stimuli: How the Internet, Junk Food, and Porn Hijacked Our Brains” on a site called Sparring Mind. He uses the wonderful illustrations that I used at the top of this piece. He writes:
With the rapid pace of technology, have we been able to keep up with the new stimulation that is available?
Some research suggests that certain things we enjoy today would be classified as supernormal stimuli, a term evolutionary biologists use to describe any stimulus that elicits a response stronger than the stimulus for which it evolved—even if it’s artificial.
Before we get into the research, let’s summarize the concept of a supernormal stimulus. The comic below will explain the basics and will take you less than 3 minutes to read.
The comics are great and essentially show reptiles slithering through the mud. “Natural selection,” the first panel reads. “A process which shapes the bodies of creatures. As well as their behaviors.” Then he goes into the work of Dutch biologist Niko Tinbergen who experimented with what triggered instincts in animals. And what he discovered, as the comic says is that “The instincts had no bounds.” So when he found, say, that a peacock would ignore other potential mates when a “dummy” was set up that had much brighter faker, bigger plumes and feathers—the bird would ignore potential real animal mates.
Are you getting it?
This trend has been witnessed again and again across animals, including as this Forbes piece “How technology is like bug sex” explains, when a beetle will see what to them is supernormal stimuli of a beer bottle which looks to them like a pornified beetle, and will simply hump the beetle to death, dying in the hot Australian sun.
It is a chilling phenomenon, and if you don’t think the same thing is happening in many ways right now with AI’s proliferation (when there is extraordinary amounts of profits to be made), well, good luck, buddy.
We’re all trapped right now in a morass and a web of what I can only describe as being the truest optimization of addiction imaginable.
Especially, particularly, when it comes to AI porn. Think:
Your dopamine is flooding. Your shame is flooding. Your sunk cost fallacy (can’t stop now because it’s too painful to admit error or weakness so I will just double down until I die) is kicking in big time.
And now it’s 4:50 p.m., and the Twitter/X space (thanks, Elon for changing the name to make it sound like a porn site, btw) so I will get to the point of the cool little tech hack that I promised to teach you.
It’s how I used several different articles to show ChatGPT a thread I was looking to write. What it produced the first time sucked, but then it was able to successfully mimic my writing style and even organize it a bit. So here’s the exchange.
And if you’re wondering why my ChatGPT is such an over-the-top sycophant it’s because I made my custom instructions be “My name is Mandy, and I want you to be really, really nice to me.”
Stay human out there.