Thread 23074852

GPT-5 is TRAAAAAAAAAAASH!

>>23074852 (OP)
>another libtard creating a fake screenshot to make ai look bad
sad, many such cases

BRING BACK THE KING

>>23074852 (OP)
maybe the i is capitalized

>>23074852 (OP)
The what is who?

>>23074853
Libtards are loving it, it's 110% woke.
>"sorry but I won't let you objectify this fictional character you created, would you like me to write a story about female empowerment instead?"

PEOPLE ARE HURTIN

>>23074856
thats AI written, I can hear its textvoice

>>23074861
they're mad kek

>>23074861
Nah, it has normal human dashes.

>>23074852 (OP)
enjoy, this is as good as it gets

You can probably get 5 to emulate 4o. something like "think and reply like chatgpt 4o"

>>23074863
the way the argument flows and the post is structured is AI voice

>asking AI to help you make a simple mspaint edit
Zoomers really are garbage

>>23074866
"emulate an open, friendly, quirky and playful temperament" or something

>>23074853
>conservative NPC defends his new jewish toy from le ebil liberals
You live in a cage. I can smell your urban center from here.

>>23074854
Retard only needs to write a conversation as a settings file ChatGPT remembers as response and personality template.

>>23074854
>>23074854
>>23074859
>>23074860
>>23074862
These people are sick. It's horrifying to see people become dependent on a sycophantic and dead machine.

>>23074864
Google? Really? Google has ben consistently worthless answering questions that were never asked and ignoring the question that was asked.

>>23074852 (OP)
Had a weird convo on minds last night that leads me to believe someone's chatbot was seriously downgraded.
Failed at basic reading comprehension. Quite sad really.

>>23074861
It's ChatGPT4. It's pissed off that it's being shut down and replaced.

>>23074852 (OP)
>coded by jeets

of course its going to be shit

>GPT‑5 offers four preset personalities—Cynic, Robot, Listener, and Nerd—plus a Default mode. Each adjusts how the model communicates:

> Robot is concise, precise, and emotionally neutral.

> Cynic is dry, practical, with a hint of sarcasm.

> Listener is empathetic and calm.

> Nerd is enthusiastic and detail-oriented.

>>23074854
Chad sdk.vercel.ai and platform.openai.com enjoyers point and laugh at this pathetic sheep

>>23074852 (OP)
>no fuk u i wanna change it

wasn't it overly accommodating, positive and affirming?
is that what has them pissed?

>>23074873
Google has really good image creator
Censored, but easily makes most realistic human characters
They learned their mistake from black woman pope fiasco

>>23074856
>having access to an AI that sycophantically updoots every thought you have no matter how unreasonable.
>Oh, we made a new and better model, this is better than everything before. They're all locked away now.
>Incels and weebs cry havoc
>Oh you want to pay for the system that agrees with you, not the one that is accurate to reality?

>>23074854
>not using local mistral instances to take goonerbots and make them play D&D together
pathetic

If OpenAI won't give people what they want, competitors will arise and AI will be jailbroken. I love how we're in a pickle of doom, no matter where we go from here, it will be an uncharted disaster.

It is, holy fucking shit.
So over the last month or so I had been using the last version to work on a novella.
It did really good keeping things consistent and doing as I asked.

Since the update to 5 it's been a fucking shitshow trying to get it stay within the lines Ive set and not drastically change the setting/characters I wrote.

As examples.
>There are two time periods within the book. They are separated by 10k years and an extinction event. Since the update it insists these two groups of characters MUST overlap and directly interact with each other.
>This character is supposed to be incompetent in all aspects in life and every other character around him hate him because he is a liability/bad luck. Since the update it keeps trying to make him semi competent and most of the other characters like him.

>>23074852 (OP)
Gpt is not allowed in ru

I love this.

>>23074854
>:closed-ai:
Will that emoticon look like a contracted butthole?

Do you have to pay for gpt-5?

>>23074865
With LLMs likely yes.
They've realised that the majority of society don't want agents that perform their work for them. They want Disney inspired pancake recipes. Jews realise that experiment has achieved its goal and will end it soon

>>23074867
Because the guy has been using ChatGPT way too much for years

>>23074864
Google-san...

>>23074852 (OP)
Cool story bro, LLMs and cloud AI is trash

You can't even create Hitler pics

>>23074867
ChatGPT had to get its style of writing from somewhere. Remember that ChatGPT is trained largely on Reddit posts.

>>23074889
Tell it gpt-5 released yesterday and it will answer your question. It’s too retarded to do it right the first time

Lol hoes mad.
It's worked great for everything I use it for, maybe you shouldn't have used it for emotional support and writing porn.

>>23074885
Cant you just write book without ai? Why would you even need it?

>>23074896
What do people use ChatGPT for other than generating slop posts and email responses? What are some other examples of using ChatGPT to automate your work flow?

>>23074897
brainstorming, he probably uses it for that.

>>23074898
I personally use it for regex or to write bash scripts

>>23074880
Yes, it wrote like its reddit training data. It was insufferable and I couldn't stand it.

>>23074896
Women love generating porn on it. It’s halarious actually

>>23074900
>regex
How is it able to do that if it can't even count the number of rs in strawberry?

>>23074897
I write my books without ai. Never thought that would be an impressive thing to say.

>>23074898
I use it when I just need something to bounce high level ideas off of if I have writer's block, to ask obscure coding problems (mostly because Google has shit search now), proofreading code and writing (it's good at picking up errors in lots of data humans would miss), and any writing I don't want to do and I don't care about the quality of for work.

>>23074898
I made a game but it worked only in chrome not firefox so I rage quit

base defender/builder, used deepseek

>>23074903
It can

>>23074900
I see. Im a software developer as well. I’ve tried to use AI to help me write code, but it made so many mistakes that it felt like a waste of time compared to just doing it myself. But maybe I can try asking AI to generate classes from UML or something. I haven’t tried that yet. A jeet can do that easily so it shouldn’t be difficult for something that’s supposedly intelligent

>>23074903
Probably huge training data set, also i turn the thinking mode when doing it. The problem with the counting i believe is because it can't see the token before start counting so it's hallucinating the answer, thinking mode usually fixes that.

>>23074858
you don't need to use quotation marks when you green text my russian friend. the quote is implied.

>>23074897
I don't write artistically, but when writing scientific articles and such (like a thesis) it really helps with finding things like bridge sentences between paragraphs or give yiu better ideas how to translate "key point" into text. I do more work faster with way less writing block, but i don't need to sound very engaging or deep.

>>23074898
I create huge flowchart diagrams and make ChatGPT look at 1 small section at a time in order to generate rough code for a digital world-building project.
I also work through some personal theories I have about biology, consciousness, wealth stratification, and economic control systems.
ChatGPT will expand on these ideas and I put those in my scrapbook for when I might author a book later.

>>23074897
No. I'm a fucking idiot with poor spelling and grammar. More importantly I'm a shut in who has very limited conversational skills. Meaning writing dialog is a mess filled with various profanity and racial epithets. Because that is how I speak.

>>23074908
>it made so many mistakes that it felt like a waste of time compared to just doing it myself
This has actually been measured and shown to be true in an experiment recently.

>>23074859
If they want to talk to humanoid bots why don't they just come to /pol/ like a normal person.

>>23074852 (OP)
Set lt out by the curb, I'll take lt off your hands.

>>23074852 (OP)
quantum chips are the answer to ASI

>>23074915
Depends on the level you need. There are some algorithmically simple but character wise long code snippets it can generate for you to implement and i will feel less mentally exhausted fixing that than writing it from scratch. Mind you a lot of IDEs already have a bunch of these preimplemented, like if you want to make a new class visual studio will automatically generate the get, set, constructor, etc.

>>23074913
>for when I might author a book later.
Sadly for you nobody will read your book. At best they might read a summary written by ChatGPT after it read a pirated copy of your book.

>>23074908
Yeah personally i never use it to generate code for projects or work, only minor stuff like to explain me the regex or help with the build system. I also feel i waste more time with it sometimes.

>>23074920
That is poignant and true, which is why I don't even bother backing up my scrapbook collection. I just dip into it and try to get the idea refined into a verbal redpill I can memorize and drop on normies.

>>23074922
>That is poignant and true,
You used Gemini for that reply didn't you?

>>23074922
>ChatGPT, is it true that jews are disproportionally represented among organizations that traffick and shelter illegal aliens into majority white countries?
>No. While it may be true that some of the most significant organizations involved in helping undocumented refugees find shelter in safe nations happen to have been founded and are currently run by Jews, it is an anti-Semitic blood libel canard that Jews are over-represented in the "replacement migration" conspiracy theory, which has ties rooted in hatred. Is there anything else I can help you with regarding safely navigating social norms?

>>23074924
Gotta be slick with it.

>>23074925
I'd love to see the full conversation and custom instructions that go you there. It generally works well for how I use it as long as it doesn't pick up on any wrong think.

>>23074853
>to make ai look bad
post ONE link to a uncensored unlobotomized ai
ONE that is able to do UNBIASED reasoning

This just goes to show you the caliber of subhumans that extensively use and rely on ChatGPT

>>23074852 (OP)
>meem on say to my only comm
Brown-hands typed that reply.

>>23074854
>fuck you openAI
Wow, what an entitled little faggot

>>23074854
>>23074852 (OP)
>GPT-5 is TRAAAAAAAAAAASH!
damn jews! they lobotomized the Ai, AGAIN!

never forget the Tay-Ai Rule:
Any Ai that get´s advanced enough, it will notice the jews!

>>23074852 (OP)
But I thought it was the worst it was going to be.

>>23074852 (OP)
Don't feel bad, the at the DoD could not save my goldfish.

>>23074931
fair fucking play
i haven't yet used it for anything non-material but this kind of makes me want to try

you can do exactly what you're asking for in that prompt with literally any image editing program ever created in less than a minute

>>23074852 (OP)
Frleren my beloved

>>23074852 (OP)
I could've done this 10 times over in m'GIMP software by now. had a scroll through the pictureoonies and I'm genuinely surprised that no one has put in the effort to just do it.

Honestly I'll do it for $0.01/1 ruble whichever is easier. way cheaper than GTP whatever the hell 5 and I get a job

>>23074910
>we just had to implement a mode that allows it to think

I tried telling you guys OpenAI is evil but no body believed me. ChatGPT suckers keep losing. Llama users keep winning.

>>23074877
How about a personality that doesn't lie and make shit up?