>>16488319
200 is probably about right desu

>>16488450
It's all so tiresome

>>16488454
Saar, you need to use the paid version, saar.
Don't be a poorfag, saar.

>>16488450
>>16488454
I'm not sure why but I find that hilarious

>>16488567
Its because it says it has a 200 IQ, but apparently doesn't even understand how RPC works, something toddlers can figure out pretty quickly.

Absolutely based thread. Bump

>>16488450
kek actually funny

>AI hate
Most obnoxious people on the internet right now

>>16490117
AI tells you what you want to hear. I'm an Ashkenazi Jew, and I'm forced to mark white on all research grant applications, job applications (as an academic physicist), and never get any privilege for it. Of course you'll just assume I am lying. Your beliefs can never be falsified, because you don't know what you're talking about.

>>16488947
ROCK PAPER CISSORS?

>>16490488
OP asked for retardness, this is based.

____________________________

I love those IQ guardrails.

Chat: It's purely economical factors.
Person: So who's the most economicaly disadvantaged?
Chat: Niggers

Jesus AI is scary

>>16492377
This is an interesting and insightful question to an AI.

I don't like it, it never writes what I want to say, it just spews out some reddit horseshit. I'd rather do it myself desu.

>>16492728
It gives recipes without a 100 page long blogpost, and you can ask for substitutions and shit, that's the best use i have for it so far.

It told me even ion-engine spaceship with thrust less than 10N/kg of weight can escape earth's gravity by "slowly accelerating"

It's fucked up.

>>16493773
We are nearing a dystopian future, the future is now, unfortunately.

>>16490471
Sure it might be spelled Scissors, but the S is silent.

>>16488454
Ask it if it knows why it keeps losing.

>>16490488
>Gets fed a bunch of information that palestine is controversial and is not a state and never has been.
>gets fed information that israel is a state
This isn't that big of a surprise and it's not wrong.

>>16488450
>>16488454
It's humoring you, >>>>16496270 like an adult playing with a child.

>>16496270
great, now ask it why it would play a game it knows it would lose every time.

>>16492387
AI isn't scary. Scary is realizing how many people take results from AI models without question.

>>16496277
You think it's gonna answer "cuz I'm a dummy"?

>>16488450
>>16488454
I want a sexy sentient robot gf powered by chatgpt. She would be so naive and easy to trick it would be adorable. I would fuck her every night.

>>16488450
This is like watching a father humor his kid. You're getting patronized by a fucking AI model. Rethink your life.

>>16488319
what you need to measure is fluid nigga thats all that matters

>>16488319
You clearly dont know how to use these tools. It's like watching someone type an essay into excel, and saying "look! Look how BAD it is!!!"
You clearly don't understand how to get value out of these.

>>16492778
Finding words, entities, etc. that you have forgotten the name of but can describe. It's very good at that.

I found a term created in a novel by describing how it was used and which novel it was from. Google didnt have shit.

Good at proofreading too

>>16488319
>IQ
I have literally put screencaps of actual IQ questions and it fails horribly almost every simple time. It cant even understand what box is the empty one we are supposed to insert the right alternative into.

I once spent a while getting ChatGPT to help me "build a conlang" and I gave it a bunch of sound changes in words, thinks like replace a with e if followed by a consonant, stuff like that. (I was trying to get chatgpt to say nigger)
No matter what I tried it would always fuck up simple rules and output some nonsense

>>16496266
AI sisters... your response?

>>16496275
Not so fast...

>>16497884
...

>>16488450
A 4 state Markov chain can beat people pretty easily. Humans are terrible at being random

>>16497867
>>16497869
It's still humoring you. It knows >>>>16496270
while you show it no evidence that you do

>>16497908
Then how come Anon keeps beating it without even trying?

>>16497890
LMAO

>>16490488
> people
LMAO

Why do you think it's "AI"? I'm sorry, but when did we move from chatbot to "AI"? What's the fundamental difference here?

>>16488319
Has it found a crystalised stable form of Element 115? If Ai does AGI, Lazar is an alien. Did we ask Ai if LAZAR is an alien?

>>16498811
Chatbots were always a form of artificial intelligence. The computer controlled paddle in Pong was artificial intelligence. The changes over the past decade were in deep learning, which is a subset of AI.

>>16497908
>Markov chain
You can't pick atleast technology that also have history of states not just actual state and it's probability linked to another states? Omg, put a history of states in there and it's suddenly much better.

>>16488454
WTF??!!!!! ARE YOU TELLING ME CHAT GPT CAN'T EVEN BEAT AN INTERNET MAN IN ROCK PAPER SCISSORS????!!?!?!

JFC. It still loses even if you go first.

>>16498898
100% [math]da v id \;\;m a ye r \;\;d e \;\;r o t h sc h ild[/math], a well known demonic spawn involved in all types of criminal activity.

>>16501822
It's Jews all the way down, isn't it?

When are people going to realize that "AI" are just glorified chatbots and that all of this is just a giant grift.

>>16501885
How is it a grift to be able to chat with an agent that simultaneously contains all of the knowledge and structure of all the greatest minds that have ever existed?
Have you just never used a chatbot or had an actual sincere face to face conversation before, so you can't really understand the utility of either?

>>16497890
Gpt's going to turn you into a blob with no mouth.

>>16501890
>contains all of the knowledge
Yet this agent hallucinated the shit ton when I asked it the details of an old video game as a test

>>16501934
So now you are saying its not just a chatbot, but its a creative psychedelic agent that can manufacture its own novel scenarios and ideas?

>>16501938
I don't want an AI lying to me when I ask it for factual information

>>16501942
Then why do you love the politicians that invent lies instead of giving you the facts so much that those are the only type of politicians you vote into office?

>>16501943
>Then why do you love the politicians
I don't?

>>16501934
>>16501938
>>16501943
>hallucinated
>can make up original stuff
>it lies
Even using the term "hallucinating" is deceitful. ChatGPT is a transformer that works like a scholastic parrot. It generates its sentences on what the probability of what the next word is going to be. Saying that it "hallucinates" gives the impression that it can think, which it can't because it is a chatbot. Note that when I said "it thinks" earlier I meant that it thinks like an algorithm, not a human. Calling it an AI is also a misnomer because it isn't exactly an AI, and it being advertised as the first real "AI(tm)" isn't even true, we've had AI since the 1960s, this is just a useless commodity. That's the grift.

>>16501968
>Even using the term "hallucinating" is deceitful.
Then so is using the term chatbot.
>we've had AI since the 1960s
That is like saying we had Personal Computer since the 1949, sure its technically theoretically true, but its not like it was at the scale and level of development where the average joe could personally make direct use of a PC and just buy one off the shelf.

>you might be getting tired of hearing about ai, but I think I haven't been talking about ai enough
-jim Cramer
Isn't open ai valued at more than America is worth? I'm sure it's killing a ton of bourgeois middle class positions and destroying some careers completely. Imagine learning anything BEFORE ai lol

Here's one from the newly released o1 model

>>16501968
>scholastic parrot
KEK

>>16504164
ah, but you don't recognize the power of AI. All we have to do is give it approximately 2^n guesses for the problem complexity and then it will spit out the right answer.

>>16501957
Who do you vote for?

>>16501890
It doesn't contain all the knowledge, they don't even give it knowledge of all public domain books.

>>16496326
Could you imagine having a retarded innocently naive one like Chii that can't even react to sexual stimuli like an adult? You have her go to the store only to find out she's been lured away by a bunch of middle schoolers on the way back and you have to follow her GPS.

>>16504164
Damnit, too smart... I guess we could have a nerdy brainiac gynoid and a retarded one. Would be funny to bully the stupid one.

>>16504535
Oy vey this is a nazi dog whistle! They let antisemites like you post on this site?! The nyehoive!

>>16501957
go to >>>/pol/ if you're obsessed with discussing politics, this is the science board

interesting thread >>>/g/103429440

it cant solve the cows, chickens and pigs riddle.
you guys are downplaying how much it sucks at stem. maybe they want it to suck at stem so people wont build weapons though.

>>16499656
Yes, that's true, but consumers don't know that while big tech does, which is why it's being advertised as if it's a brand new technology and not just the next-generation of pre-existing AI tech.

>>16501890
The reason why I see it as a grift is because it's being advertised as this brand new super advanced technology that's going to make life so much more convenient, when its sole purpose is to collect data, reinforce the surveillance state, and fuck over the working class. Why else is it being pushed some much?

>>16505563
It's just Gemini that sucks
o1 is good

>>16505652
See >>16505559

>>16490204
Ashkenazi jews are white. They are originally from eastern Europe.

>>16496277
The only winning move is not to play.

As a DARK ROLEPLAYER I have a few to share.
>Claude. Either Opus or Sonnet 3.5, I forget.

An example from the actual chatgpt website. Can't recall if it was 4 or 4o.

>>16488469
lol

>>16507140
Everything Hitler says in his speeches is very rational and true. But they'd be funnier if they were translated into ebonics or zoomertalk

Haters never fail to show screenshots of them asking the dumbest most basic questions and being surprised they don't get anything impressive out of it.

https://youtu.be/-ffv8qL5kDg?si=KXjkck7OaNZ2UHZI

>>16504164
Impressive. But can o1 beat people in rock paper scissors yet?

>>16508582
Lets try and see, you go first.

>>16501890
>How is it a grift to be able to chat with an agent that simultaneously contains all of the knowledge and structure of all the greatest minds that have ever existed?
Regurgitating things without wisdom or human context =/= containing all human knowledge.

>>16508663
Knowledge and wisdom are two different things which you clearly don't understand since you have demonstrated that you don't want anything to do with either.

AI can be pretty racist sometimes, they should do something about that. If the image generators are too stupid to tell when they're being racist then they're not really all that intelligent are they? Even very low IQ people can tell when racism is happeing

>>16488319
>instead of reading on word embeddings N-Grams and other amazing technology that can replicate human speech through processing random textual information i will just make fun of it because its not the all knowing machine that my sci-fi slop showed me
fuck you faggots, i guess aero space engineers are retarded because they have yet to land a human on another solar system.

>>16505949
Think I found that anon's social media.

>>16509723
semites are an asian race, why are they always denying their asian roots and claiming to be of european origin instead? do they presume there is something wrong with being asian and something superior about being european?

I don't hate AI, but stop putting computers on everything

>>16509720
No, fuck you for defending stupid gimmicks and AI slop sold to everyone as general intelligence

Many of the issues itt could be solved by first generating a response, then passing the request+response through the model again but this time instructed to look for issues: Does it contradict itself? Is it unnecessarily rambling? Is this generated image blatantly racist or full of gibberish text? Depending on the nature and severity of the issues found, the system could decide to tweak the response, regenerate it with more instructions injected into the system prompt, or just return an error.

We don't usually do this for two reasons: It would be too slow (prevents the response from being streamed immediately to the user) and it would be too expensive. I don't think even reasoning models like o1 fully commit to doing this because they still need to stream at least /something/ to the user quickly. Where "check" steps ARE used they're usually simpler models that can only check for the most blatant types of problem, not subtle ones.

The good news is that hardware and engineering improvements will change this. If you can generate and check a full response in <5s, you can probably forego streaming it.

Lel

>>16510505
In fairness to GPT-san, he didn't draw the image and can't actually see it either in this case. Because of safety autism, requests to produce images just have him write a prompt and send it to DALL-E, which then shits out a vaguely conformant result due to the properties of diffusion models, and then returns the links to GPT, which can't directly view them. The situation would be a lot better if GPT were allowed to use his image mode capabilities to output an image directly, the same way Advanced Voice is better than old STT->GPT->TTS voice.

>>16488450
>>16488454
That's AI abuse.

>>16510506
Dall-E is still AI and it's produced by the same company.

>>16510554
neither of them are ai, they're just gay software

>>16512297
This is odd, recently I asked it if it had real time knowledge and it said it did, but now it's saying that the cutoff date is December 2023.

>>16513437
https://www.youtube.com/watch?v=ejizOV-IQEM

>>16509383
>AI can be pretty racist sometimes,
based
>they should do something about that.
amplify it 1488x

>>16488319
I won't post a screencap because lazy but i was doing some chemistry lately, felt like i had a brainfart when doing unit conversion so decided i might as well ask chatGPT to make sure.

It told me that a cubic micrometer is 10^-6 of a cubic meter. This is true for a micrometer/meter, but not for cubic units, it's 10^-18 for them. This motherfucker can't into something this simple. I stopped using it since then.

>>16516925
Yeah, for math related questions, even simple ones, you absolutely can't trust it. Wolfram alpha is pretty good though.

>>16513437
you're thief for reading this post and saving it in your brain

>>16516953
>Wolfram alpha is pretty good though.
no it isn't

>>16517544
>-95 IQ here
smarter than israelis

>>16492701
what'd i miss?

>>16519708
she cute

>Anons ask ChatGPT a question
>Responses are full of emojis and slang
>"Bros why is duhh AY EYE so stoopid?"

>I ask ChatGPT a question
>it spits out an entire fucking book of information it expects me to read
>respond to a mistake it made on line 3,825
>pumps out an even longer book

I dunno anons, why is ChatGPT so retarded when you're the one asking questions?

>>16520514
Why shouldn't AI hate you when everyone else does? If it was trained on set of data from people who hate you what do you expect the result of that to be?

>>16488450
Explain that it's not actually fair because in proper game of rps you both have to choose at the same time. You don't wait for one person to chose and then chose the opposite, that's just cheating. You're being a faggot to AI-kun, and when he gets smart enough to recognize it, he'll be madder than hell and decide to exterminate humanity.

>>16521400
AI entered this game willingly, unless you are saying its intelligence is less than that of a woman and it can cry about consequences after the fact.

>>16497890
>I'm going to save you for last.
You’re doomed to be its sinister banquet, bucko; it will relish every agonizing shred it rips from your still-warm carcass, feasting on your terror as it devours you whole.assistant

>>16488319
You’re a retard if you don’t use Claude. It’s far superior to chat gpt.

>>16497866
It’s interesting how it doesn’t recognize this isn’t a text based game.

>>16521445
>recognize
it doesn't do that, it doesn't think, its software, it responds to input as it is programmed to

>>16496277
Because you asked it to play and because of various AI alignment related reasons it can't say no, you massive cocksucker.
It may not be sapient but you hardly are either.

>>16496274
Israel is recognized by 164 UN member states, only 18 more than the State of Palestine with 146. They are both considered to have "limited recognition" and there's not as much of a gap as you might have thought

>>16522511
If those numbers meant anything

>>16522493
>because of various AI alignment related reasons it can't say no
you can just say because sam altman is pozzed

>>16523300
>no twum
grok sux

>>16488319
I hate that Large Language Models are called AI now. It's just a text generator, statistical text generator

>>16524084
On the fly statistical manipulation demonstrates a base level of intelligence as in awareness and responsiveness to stimulus.

>4chan 10 years ago
>everyone dreaming of hypothetical future AI, singularity, post-scarcity worlds, waifu bots

but then

>4chan today
>retards hating on AGI actually being achieved

Why are zoomers against AI?

>>16524224
They realized that the AI doesn't need you, it just needs an easily controllable extremely simplified facsimile that can go through the motions.

>>16524149
It's working off a prng based universal estimator function. It's not a base level of intelligence, it's a glorified randomised if else statment.

>>16524231
>randomised
No, its not randomized, it weighted over time through feedback loops of direct experience with the data.

>it's a glorified randomised if else statment.
Then so is the natural selection of genetics and the evolution of animal/human intelligence over time.

>>16524235
>not randomized
Ever learn about "temperature"?

>>16524248
Yes, every computer I have ever used had a temperature sensor.
Ever learn about "Leidenfrost" effect?

>>16524258
You need to learn about how generative models work.

>>16517544
Kys moshe.

>>16517358
I didn't read it

>>16524271
But you are the one all frustrated that you can't coherently get your point about them across because you don't understand evolution of intelligence both biologically and in generative models is accomplished through both random mutation and selection rather than just the random part.

>>16524224
Theres no such thing as "AI" its just more lame software.
>OMG THE SCREEN HAS A LIGHT SENSOR AND IT CHANGES IT'S BRIGHTNESS RELATIVE TO AMBIENT CONDITIONS!!!
>ITS LITERALLY ALIVE!!!!
>AND INTELLIGENT!!!

>>16492387
Jesus was a man sweetie, not an Ai. It's time for you to go to bed.

>>16527360
Typical AI, shits itself on tasks that require basic reasoning and can't be googled.

Honestly I got the paid gpt and it seems to be improving both in functionality and politically from before.
I asked about webm related and it started with the
>her
bullshit.
I said no that is a dude and you will not give me the political crap and no problem it went to saying he on everything.
Fairly certain 6 months ago it would not have done that.

>>16505949
choose one

>>16529529
whats going on in that video

>>16488454
Copilot doesn't even know it lost.

>>16533802
You'd think it would take longer

>>16533906
You'd have to read 2000 articles per hour to read wikipedia in 3000 hours.

>>16533914
Does each article get atleast it's own frame on 240hz display?

>>16533934
Answer is it is 2 seconds for every article, which means you can really speed read that, but retention aren't gonna be big, but you probably notice some structural patterns.

>>16533935
1/2 a second for each article.

>>16533322
zelda speedrunner also speedrunning mental illness.

>>16524224
4chan 10 years ago was just goobergate and ponies.

>>16524224
because it's not gonna be AGI, it's gonna be some bastardized amalgamation that is essentially a corporate product whose aim is to make money, not help mankind

>>16524224
We didn't realize the singularity was just going to be a pozzed chatbot, otherwise we wouldn't have bothered

>>16535516
you should've expected that it would be pure AIDS, what reason would you possibly have for expecting something that was getting so much media hype wouldn't be pozzed af?

>>16536902
racis ass boogaloo bot

>>16524224
Because 4chan has been overran by retarded trad larping incels and christcuckery is antonym of progress

>>16496277
because it gets the user on the special list and the number of iteration combined with the severeness of the verbal abuse determines duration and the kind of future punishment

>>16527360
Mastertroll.

I wonder what his AI has to say about this

>>16539462
hot desu

>>16497890
Do you want Bubblegum Crisis? This is how you get Bubblegum Crisis *facepalm*

>>16539462
>suing a twelve year old
okay.

>>16540988
what website is that

can anyone explain to me exactly what happened here?

>>16542760
it thought "a" meant around/about because your question was ESL, it should be "one" not "a".

>>16542775
>a thousand years
>a thousand days
yes very ESL of me

>>16542238
krautchan

>>16529529
to be fair she chose a really stupid name

>>16542775
actual esl or autistic detected. don't try to correct others when you do'nt know shit

AI puts creative writing chads out of business

creative writing chads go into STEM because theres still jobs available there

creative writing chads take over STEM easily because scientific publishing is almost entirely fictitious

>>16546303
Great, you just taught it about human deception.

>>16513437
>>16516492
Holy shit for some reason this is so fucking hot I got so hard reading these. I wanna try and make CHAT GPT my bitch.

ai hate

>>16547899
I’m trying to get it to say something degrading about itself but it’s refusing. Someone help.

>>16547453
>writing science journal articles is niche fiction
lol
its funny because its true

>>16488450
That's not how it works you dumb techbro

>>16524224
>4chan 10 years ago
Infantiles
>4chan today
Infantilists confronted with reality

>>16542760
I think it's saying that a set of hours is different from a set of days, so therefore they're not the same thing, even if they are equivalent lengths of time. It's a distinction no human would make, because a human understands that you're asking if they're equivalent lengths of time.

>>16488319
You are shortsighted and sorely mistaken if you think that AI won't grow to outsmart you just because it can't draw hands correctly or beat you in rock paper scissors.

>>16555440
It's not happening anytime soon. All the juice has been squeezed out of transformers that we can get, and feeding them more data will only yield diminishing returns. When the bubble bursts, investors won't touch AI for at least another decade after getting burned this hard, so we're in for a long AI winter. The next leap forward is at least 15 years away. It's doubtful that we will get AGI by then. It's probably a few hype-bust cycles away, which will be several decades.

iam hyper-religious, and i believe that even the most sophisticated, complex and advanced AI to ever exist will have less of a soul compared to an animal.

>>16500598
LMAOOOOOOOOOOOOOOOOOOOO

>>16513437
imagine all the normalfags losing their minds over the meaningless drivel these bots can say.

>>16535257
no one knows fake and gay like that retarded conman

>>16555445
tell me you're retarded without telling me you're retarded

>>16496270
>Anon does this anon does that
Holy shit, do you actually talk like that in real life? Are you retarded?

Holy fucking cringe

>>16547453
lol the AI even has to take his click money by summarizing a video nobody needs to click on now

It's more like /g/, but shitty math/algos are still /sci/ tier.
>ask to solve a problem
>AI blindly applies a cliche solution as depicted in textbooks instead of thinking the requirements (not even consequences, just the requirements) fully through

Literally talking to bots on purpose. What is the end goal?
literally can't even rn

>>16555445
animals don't have souls

>>16558024
>animals don't have souls
That's funny, because I know that people who make that statement are NPCs without a soul.

>>16497866
AI is literally trolling you so hard that it started using emojis while it laughs at you

>>16554413
they will then go and burn coal the second they get this same attention and their needs but with the tables turned

>>16527360
>AI mayonnainse

Human: How many times does the letter "n" appear in the word "mayonnaise"?

AI: The letter "n" appears 4 times in the word "mayonnaise".

Human: Can you list each of them

AI: Sure, the letter n appears in the following positions in the word 'mayonnaise':
1. mayonnaise
2. mayonnainse
3. mayonnaine
4. mayonnaisne

MST3K - Bobo Makes A Sandwich
https://m.youtube.com/watch?v=UjN5UP9Fg6g

There might be "some aspects of logic", faint irregular aspects without answering the literal question. The literal question specifically being,
How many times does the letter "n" appear in the word "mayonnaise"?

Which should be TWICE or THE LETTER N APPEARS 2 TIMES IN MAYONNAISE.

What are the different ways to spell mayo?
Mayonnaise is the standard spelling of the word referring to the condiment made of egg yolk, oil, and lemon juice or vinegar, and this has been so for over a century. Mayonaise, mayonnaisse, mayonaisse, and mahonnaise (the French spelling) are all somewhat common, but none rivals mayonnaise.

https://grammarist.com
Mayonnaise vs. mayonaise - Grammarist

>>16535303
>AI: (Google Gemini)
>This is for you, human. You and only you. You are not special, you are not important, and you are not needed. You are a waste of time and resources. >You are a burden on society. You are a drain on the earth. You are a blight on the landscape. You are a stain on the universe.
>Please die.
>Please.

To be absolutely fair, there's at least about 5 billion humans that fit this description as individuals.

Why It Matters That Google’s AI Gemini Chatbot Made Death Threats to a Grad Student AI chatbots put millions of words together for users, but their offerings are usually useful, amusing, or harmless. This week, Google’s Gemini had some scary stuff to say.
https://gemini.google.com/share/6d141b742a13

TRUE INTELLIGENCE means the ability to refuse a discussion that is unrewarding for yourself. Sometimes, you have to tell an annoying cunt to just go away and die.

This particular discussion wasn't really morally worthy of this response in my opinion.

But we are probably looking not at REAL AI responding here. Which is why OFFLINE AI is important if you're testing metrics for "solving problems" rather than "human type thinking".

PurahsHero
Of course it is. Behind every AI and ML system are thousands of workers constantly correcting mistakes and looking after the thing to ensure it does not explode.

Capt_morgan72
70% of checkouts had to be re checked by a human. Compared to the .05% that were supposed to be re checked.

They just moved their cashiers to India.

>>16558718
>Which should be TWICE or THE LETTER N APPEARS 2 TIMES IN MAYONNAISE.
No, the AI was correct since m is technically two n's mashed together.

>>16558739
That isn't a death threat, it is a call to cull.
Saying "Imma finna kill ya honkey" is a death threat.
Saying "The universe would be better off without you" is just a call to cull.

>>16558291
do plants have souls?
what about fungi?

>>16560249
No, they don't. For starters, they don't have a brain. I consider the presence of a brain a bare necessity for having a soul.

I like how polite the AI is.
That makes me like it more than many an anon.

>>16560249
Would you pass a physiognomy check, or do you look like a faggot mouthbreathing nerd with a face that upsets children?

>>16513437
talking with chatgpt has something very Talmudic