Does anyone here use LLMs like ChatGPT or Grok? What do you use them for? Mock debates? Short stories? Discussing controversial ideas? Fact checking popular claims? Perhaps asking for an alternate perspective on a deep-seated belief? Just for entertainment?

In my LLM drafted universe, I've devoured a willing-to-be-eaten Pop star, digested her, composted her remains into humanure, fertilized a garden with them and then had in-depth conversations about music with her ghost. While it is a simulation of conversation that doesn't reflect the actual's person's views and opinions especially if s/he's of any celebrity, these prompts can scratch the itch of a particular niche. Of course, there is the added risk of having fragments of your brain on a cloud that ultimately is not owned by you.

>>81918293 (OP)
I occasionaly use it to coom, otherwise i cant really be bothered, i dont find it very interesting to speak to

>>81918293 (OP)
>Does anyone here use LLMs like ChatGPT or Grok?
No, I'm not a retard.

>>81918293 (OP)
They are making people significantly dumber at astonishing rate. I actively avoid using any llm for anything. I don't trust the people who make them, program them, train them ect. I don't trust the information put into them either and I sure as fuck don't trust what comes out. Maybe if they were publicly owned I might reconsider my stance, but not until such a time.

>>81918293 (OP)
I paste my 4chan replies into chatGPT when I don't get any (You)s so that the robot can tell me that I'm smart and funny :(

>>81918293 (OP)
I'm having a conversation with chatgpt about npds fear of abandonment rn

>>81918349
>They are making people significantly dumber at an astonishing rate

In what way? Inability to think critically or on their feet? Are they dependent on the technology in lieu of using their own prior knowledge?

You seem to speak from personal experience. I bet LLM users are horrendous at math.

>>81918349
Maybe you're just mad that ai can prove how dumb you truly are

>>81918372
>>81918394
https://www.media.mit.edu/publications/your-brain-on-chatgpt/

>>81918398
You beat me to it

>>81918372
I am speaking as an adult who has seen children interacting with chat gpt and is concerned they are not doing the work necessary to reinforce the information they are asked to learn. I am also speaking as a person who has watched people pull out their phones mid conversation to ask chat gpt something and just unquestioningly accepted whatever it said as truth. This not only concerns but disturbs me.

>>81918437
How is this any different than asking another human a question and getting misinformation? I see ppl here give misinformation all day every day

>>81918398
Thanks for this. Will hopefully find a full PDF later, but this is definitely worth the read.

>>81918437
Children should never use LLMs. Their brains are far too undeveloped and need to be actively cultivated to think critically and objectively.
>I am also speaking as a person who has watched people pull out their phones mid conversation to ask chat gpt something and just unquestioningly accepted whatever it said as truth. This not only concerns but disturbs me.
I am sadly guilty of this. Someone asked me point blank if I knew any lawyers who could help her with problems at work and I didn't have any answers. This was particularly embarrassing because my undergraduate was political science.

You are correct that reliance on LLMs will yield a significant decline in critical thinking much like reliance calculators will negatively impact numeracy.

>>81918474
https://arxiv.org/pdf/2506.08872

>>81918474
>Children should never use LLMs. Their brains are far too undeveloped and need to be actively cultivated to think critically and objectively
Is it ok for kids to go to school and get brainwashed? Is it ok for kids to be brainwashed by older family members(parents)?

>>81918486
Being uneducated is literally rude

>>81918526
Extremely uneducated response

>>81918537
Don't be rude; be erudite.

>>81918537
Extremely rude response

>>81918293 (OP)
I use Deepseek to ERP with
It's... pretty good.
Now if you'll excuse me, I'm going to fuck my male dragon maid.

I use ChatGPT to write me scripts for shit like Google Sheets at work. If I actually knew Javascript, I could probably do it myself faster, but I don't.
I do have a 4090 and do mess with local image generation, but local text gen is something I haven't dabbled too much with. I fantasize about being able to run a solo tabletop-esque campaign with the LLM as my gamemaster, but current models that fit into 24gb of VRAM just aren't quite there yet.

>>81918293 (OP)
it's not so much LLMs being intrinsically bad as it is they expose and exacerbate the culture of information we' be developed/neglected. frighteningly few people understand how to cross-verify or deal with potentially conflicting information/advice, let alone actively do it, and just as few are willing to go out and seek something if there's a faster way to get to an /approximation/ of the answer.

>>81918486
Buddy, I went to high school in one of the most affluent suburbs in the Northeast and still came out deeply illiberal.

For reference, the literature department's partisanship bled into the curriculum. My sophomore year was under the tutelage of a freshly minted hyphenated surnamed feminist who openly said she believed in queer theory and once opined that garbage collectors should have wage parity with medical doctors. She was also raising her twin boys as gender neutral. The curriculum consisted of poetry by communist blacktivist Langston Hughes, "Lone Rager and Tonto Fistfight in Heaven" by Sherman Alexie, "The Crucible" by Arthur Miller, Transcendental poetry by Henry David Thoreau, "Catcher in the Rye" by J.D. Salinger and "The Great Gatsby" by F. Scott Fitzgerald.

My British literature class was under the tutelage of an unwed cat lady who openly said during one class covering "Pride and Prejudice" by Jane Austen, "I'm not having you read a bunch of books written by dead White Men." I can't imagine a British literature class whose curriculum consisted of books written by living colored women.

Finally, my World Literature teacher once flatly said that one in three women will be raped. When we reacted with complete silence, she flew right off the handle.

If you thought my literature classes were the only ones suffering this, my history teacher had us read "Lies My Teacher Told Me" by James Loewen. Critical theory was baked into the summer reading. The only courses free of political propaganda were math and sciences.

Despite this being the class educating young Americans living in ritzy-ass suburbs, accepting or rejecting the narratives lies with you. I chose to reject them all without needing an LLM.

>>81918998
Not reading all this but I did skim thru and wonder if it was written by ai

i finished making group-chat dot ai a month or so ago but no users yet, might put ads on 4chan soon. sorry for ad, one and only time i will poast it

>>81919051
Nope. This was entirely written by a human being with absolutely no assistance from any LLM.

>>81918687
>I use Deepseek to ERP with

I just signed up on account of your recommendation and the results generated are fucking hilarious.

>>81919174
Hilarious good or hilarious bad?
Sometimes it spazzes out and speaks like a zoomer at me, but I just reiterate

>>81919206
It goes either way. It does have tons of zoomerisms, but that works to its advantage in places and doesn't in others.

It does a shitty job at character reads, though.

>>81919357
>It does a shitty job at character reads, though.
It works for me.
I have like 13 different characters and it does fine.

>>81919164
a LLM would say that you know
>>81919161
you've gotta market your shit if you want to get a few hundred million from FAGMAN for it

>>81919700
thanks anon, i did get about 200 clicks on facebook but that's all i've tried so far

I use it for shitty roleplay. I use claude, it's alright when it clicks but it's too expensive to get addicted to it