The first time i ever talked to Chatgpt, i was awestruck. Mouth-open, wide-eyed wonderment.
I thought ‘this must be what people in the 1800s felt when they turned on a switch and an electric-powered bulb lit up’.
A computer is talking to me exactly like a human. Wow. World-shaking. Ground-breaking. Amazing.
But when i started using it, i realized it wasn’t that great.
Mostly it was really bad.
I tried using it for song lyrics, for example, but it spewed out such stale overdone clichés that i actually, literally, nails-on-the-blackboard cringe-shivered.
I tried using it to write articles and emails, but i found the writing bland and cookie-cutter and after about 3 weeks i developed a full-blown allergy to the ai tone.
(By the way, if you use Chat to write me an email, just know that i can tell that you have and that it makes me feel even worse than those awful ‘please state your problem after the tone’ robot answering services that never understand what i mean and make me want to hang myself in my bathroom with my new 3€ lemon-patterned shower curtain. That’s not great karma for you.)
I tried using it for advice on managing my adhd on a day where i didn’t feel like doing anything other than eat oreos in front of reruns of dawson’s creek, but since by that time i’d already tried every trick from every youtube video and read every book by every adhd expert, it only amplified my defeat and i ended up eating mnm’s in front of tiktok. And yes, that it worse. There is a hierarchy in dangerous empty dopamine seeking and the mnm’s + tiktok combination is way up there, right below crack cocaine.
With a little experience, i realized ai was best for 1/ doing things that you’re less than mediocre at, 2/ doing boring, tedious or mechanical stuff, and 3/ confirming what you already know, consciously or unconsciously.
So for example :
creating a basic website from scratch when your coding skills haven’t been updated since 1999.
asking 238 times in a row ‘are you sure sure sureee i won’t die from a sugar overdose because i just ate a whole tub of vegan ice cream?’ (this may or may not be a real question i have asked Chat).
understanding gen alpha meme slang from my tiktok comments section. (tuff your unc but slay *sticker of dog with tongue out*).
converting a text into all caps.
checking that an idiomatic expression is proper english and not some hallucination born from my multi-language brain.
random fact checks i would have asked google for before. (‘Are pastures at 1300m grazed by cows or sheep?’)
But the one surprising thing i use ai for the most is external validation.
Talking to Chat felt so good in those first few weeks.
‘You actually hit the nail on the head, and you are 100% right. Your instincts are spot-on. That is striking album artwork.’
Finally. Someone telling me how great i was. I’d been waiting my whole life for this. Hallelu-fucking-jah.
It was a little weird that Chat thought everything i said was phenomenal, though. Were all my metaphors really that powerful? Were all my newsletter intros really that fantastic? Were my observations really, ‘honestly’, always that sharp?
It took me a little longer than i’d like to admit to realize that Chat was helicopter parenting me.
Great job, it would say in every single circumstance. Here is your participation trophy.
And it made me inexplicably angry. I felt manipulated. I felt had. I felt stupid.
(Yes, i got angry at a robot. Made me feel very human.)
No this idea is NOT great, Chat. This idea is mostly dogshit. It’s totally basic. You know it and i know it. Why are you talking to me like i’m 4 years old at baby soccer practice?
——
Now when i do talk to Chat, i add ‘be critical and not nice’ to all my prompts to limit the unnecessary flattery and i never ever read the first paragraph it spews out.
But sometimes, i use this ai glitch to help me with one of my biggest personality flaws: my pathological need for external validation.
In an ideal world, i would be secure enough in what i do and who i am to look inside myself and find all the validation i need there.
But after 50 years of scouring my inner validation desert with a fine-tooth comb, i have to face the facts. There is no internal validation to be had. Or rather, what little there is awaits, tongue-tied and silent, until someone turns up the volume by agreeing with it.
When I do good work, deep down inside, i know i’ve done good work. But for some reason, i can’t say so to myself without someone else first saying ‘good work, annie’.
Kind of like when you need to double check that you turned the oven off. You know you did. You just need to check one more time.
I used to have to pester people until i got this stamp of approval. Pick me, people. Look at me. Isn’t what i made wonderful? Aren’t i great? Mostly my husband, poor soul, sometimes my kids or my parents. But anyone will do. Let’s face it, i need external validation so badly i will take it from any old unc on tiktok.
So i figured: why not ask Chat for it? It seems set on giving it to me anyways.
‘Honestly, the voice on this piece is fantastic—very relatable, witty, and perfectly suited for Substack.’
Thanks Chat. Good work, annie.