K2
BroadwaybabytoDo not take legal advice from chatGPT.
Do not take medical advice from chatGPT.
Do not take financial advice from chatGPT.
In fact, stop using chatGPT. It’s destroying the environment, giving people terrible advice and lining tech bro pockets.
quagga 🦆@broadwaybabyto sooooo many people in the online chronic illness communities i’m in (POTS, MCAS, long covid) use chatGPT because doctors won’t help them.. like i get how frustrating that is, i experience it myself, but holy fuck a notoriously erroneous & wasteful computer SHOULD NOT be giving medical advice!!!
@quagga right? I worry about them because it gives such bad advice and who knows how secure your data is!
I would rely more on old fashioned support groups on Reddit or Facebook (evil as meta is) than chatGPT
DeanFarrell@broadwaybabyto I've yet to use ChatGPT.
@DeanFarrell same! Found out someone was feeding my disabled ginger articles into it when I had asked them to proof read … needless to say I was enraged.
Court Cantrell prefers not to@broadwaybabyto @DeanFarrell I haven't used it either and fully intend not to *ever*. Everything about it is janky.
cedricdes@broadwaybabyto oh, you just know from others that it gives bad advise. I'm gona post it as well, so people know from me what I've learned from you, which you learned from others
Ozzelot@DeanFarrell @broadwaybabyto I have attempted it in its very beginning. It was useless for niche information I needed, and I dismissed it as a boring unreliable toy. And that doesn't even mention the environmental impact and training on stolen works.
jan Ki | 奇 
@broadwaybabyto
some people even take advice on mushrooms from it...
some people even take advice on mushrooms from it...
eddie@broadwaybabyto It's not even giving 'advice', because it does not understand the meaning of the generated text.
Allan Svelmøe Hansen@broadwaybabyto but but but then I'll have to do my own work......
Why Not Zoidberg? 🦑@broadwaybabyto Waiting for the first report of someone asking an LLM what mushrooms are poisonous...
Lord Caramac the Clueless, KSC@broadwaybabyto I don't use ChatGPT for anything serious, I just play around with it once in a blue moon to see if I can get it to say things it isn't supposed to say.
@LordCaramac do you know that it causes enormous damage to the environment? I find a lot of folks don’t know that and so they play with it on a lark but contribute to its massive footprint.
@broadwaybabyto What's causing the damage is basically any kind of economic growth. The global economy is already so big that it will eventually collapse, the only question is when and how. All of humanity's collective industrial enterprises are utterly unsustainable. Some are worse than others, but even the "green" technologies are just less bad but not good. The Industrial Age will end in chaotic collapse, one way or another. We're so deep in ecological overshoot that the best we might still be able to do is a slow version of collapse, a decline over the course of two or three centuries, but that will only happen if we get rid of Capitalism on a global scale very soon.
Steve@broadwaybabyto @LordCaramac I thought it was the training that had the massive footprint, not running a few prompts. I can run an LLM on my hardware at home, so it definitely doesn't consume more than 600W for a couple of seconds each prompt
@flying_saucers @broadwaybabyto I don't even use LLMs much. I find them rather boring. Once in a blue moon, I log into ChatGPT and try to make it do something it's not supposed to. Smaller, more primitive, uncensored LLMs are much more fun, yet even those I find boring after a few minutes. It's image generators that I really like, but mostly the ones based on Stable Diffusion or Flux, the ones where you can still easily get weird mutants and eldritch abominations.
@LordCaramac @broadwaybabyto I use them for annoying things like copying the structure of a document without having to delete all the bits in between, getting my head going on a topic when I'm stuck, that sort of stuff. Not often, but often enough to have one installed and find it useful-ish. Also doing a course at the moment for some professional development, and they have an LLM chat that's surprisingly good at diving deeper into some chapters to improve understanding
Aerliss@flying_saucers @LordCaramac @broadwaybabyto CO2 and equivalent numbers are hard to come by, but one old fashioned search will give you plenty of articles to read on the matter https://www.scientificamerican.com/article/a-computer-scientist-breaks-down-generative-ais-hefty-carbon-footprint/
https://www.sciencefocus.com/news/worst-chatgpt-prompts-for-the-environment
https://www.sciencedaily.com/releases/2025/06/250619035520.htm
They also use water; https://www.independent.co.uk/tech/chatgpt-data-centre-water-consumption-b2318972.html
Bare in mind these are just for simple search queries. Analysing text, generating large text files & images have a heavier burden.
But the problem is broader than environmental impact of a single query
@flying_saucers @LordCaramac @broadwaybabyto every time you interact with these glorified chat bots, you're telling the owners & creators that their tech is worth something to you. Doesn't matter if you "only use it for silly things" or whatever. All they see is someone using their machine. So they'll push it more, work on it more, build more of them.
@flying_saucers @LordCaramac @broadwaybabyto proprietary LLMs that are designed with a more narrow focus definitely use less energy (one of the links above is about how choosing a good model can massively reduce impact). And of course they definitely have uses (my go to example is massive data analysis in medicine & science), but the online chat bots that the average person uses day to... energy, water, stolen work, every time.
And yeah, the training is energy HEAVY.
@flying_saucers @LordCaramac @broadwaybabyto Then there's the moral issues. Your "just trying to get it to do things it shouldn't" (like a teenager drawing dicks in a text book 🙄) is using the stolen work of tens if thousands of people, & telling the owners & makers that's acceptable.
So now they have more incentive to keep pushing these things & more incentive to steal more work. It's a snowballing problem.
@flying_saucers @LordCaramac @broadwaybabyto you can say "well everything's going to hell anyway, I'm just a tiny person having silly fun with a bit of tech, & capitalists are the problem".
That's a sound point. But there are millions of people out there doing work to halt this manmade asteroid. From people working grassroots to educate & campaign for local green initiatives, to engineers working on clean tech, to lobbying groups taking on the commercial giants.
@flying_saucers @LordCaramac @broadwaybabyto just because things are going to be bad, doesn't mean we throw in the towel completely. Things can be bad... or they can be worse.
Corpomancer@broadwaybabyto We at corporate regret you feel let down legally, developed a medical condition due to the stresses caused by and can no longer financially make ends meet due to our out of control greed.
We will continue to promise ourselves we'll do better, now please increase usage of the lying box so our targets get hit.
b ste@broadwaybabyto thanks for posting!
Queen Kitty Hestie :3c@broadwaybabyto ppl actually use chatgpt? gauth ai is better anyway (for doing hw)
Some Guy@broadwaybabyto It's great for creative writing ideas. I appreciate (genuinely, not being facetious or sarcastic) the hallucinations.
It does have its uses and information it provides on certain topics, while can be found elsewhere online, can be made easier to access with resources (external links) provided.
Safe to say "your mileage may vary."
Also, RE: tech bros
Don't put any PII in the prompts. ChatGPT doesn't follow HIPAA or maintain any form of confidentiality or respect privacy.
Azarilhⓥ@broadwaybabyto I do agree, but gotta be honest. When i wasn't against current language models, i once got something on my eye that was stratching it and it hurt. I asked ChatGPT how to get tid of it and with its help it took me mere seconds to solve my problem. So in a way, ChatGPT saved my eye from getting damaged...
Andrey Rys 🇨🇿@broadwaybabyto never used it, but my friend took a medical advices from it and now is near the borderline...
Aurelian Dumanovschi@broadwaybabyto that's the cool thing. This ain't making anybody money. This is a money sinkhole. Actually the more people use it, the less money the more money they lose.
If more people used the free services, the quicker they would go bankrupt.
r҉ustic cy͠be̸rpu̵nk🤠🤖@broadwaybabyto But then how would I continue ignoring the systemic issues that lead to vulnerable folks relying on ChatGPT in the first place? /s
Clinton Anderson SwordForHire"I used AI to....", is nothing more than, "Listen I'm not an asshole but....", for the 21st Century.
Evan Light@broadwaybabyto I realize that we don't know each other. I don't know if you're open to a differing perspective.
Know this:
I'm not a tech bro. I'm #actuallyaudhd and deeply empathetic to the suffering of others.
I held a similar outlook to yours.
Out of desperation, to cope with sensory processing issues realized upon my #actuallyautistic diagnosis, I found that, however improbable, this tool has helped me find a measure of peace. No coach has been able to do that. No therapist. ChatGPT did that—after I shared vulnerably of my experiences much as I have with my therapists.
If it can help me this way, perhaps it can also help others.
Is it a devil's bargain? I can't say.
With or without AI, we're losing this planet. Something has to change, radically. As a species, we need to change our focus to the good of the species and the individual as a member species and not the wealth of the individual.
That is the ill that plagues us.
The irony is that OpenAI had the potential to be that but for Sam Altman and cynical opportunists like him.
The world needs to be managed by the best of us and not the wealthiest or the most power hungry. That's what must change.
A quick look at your profile: relatable.
Comrade Ferret@broadwaybabyto as someone who personally seems to never learn this lesson: don't take tech advice from chatGPT either