Is it time to pull the plug on A.I.?

Is it time to pull the plug on A.I.?

It’s too late
The race is on

AI is fucking retarded though it's all Jew worshipping slop

How am I supposed to get the money to buy cheeseburgers when the robots take the jobs

AI has IQ

Are these people so hyped they smell their own farts? How can any rational person say something so stupid?

You can't really pull the plug at this point, way too late.

They need to keep the dot com bubble 2.0 afloat

lol it's like saying a library has an IQ

dead internet will become a reality within 5 years

cyberpunk unironically predicted the future, people will be so annoyed with an AI infested internet that they’ll stop using it

ai intelligence will never exceed the language models it's trained on
and it's trained on humanity's dumbest

GPT 4 is retarded and constantly gets things wrong

It's an LLM piecing information together from contextual cues

Okay but that doesn't change the fact that ChatGPT is more capable at doing that than the average retard trying to use google

"""AI""" is just an LLM, a random bullshit generator based on the input. It has no actual intelligence and frequently makes shit up. It's the biggest scam of 20s.

All ai depend on english as base

every one whoused englush is retarded

Cant feed sanskrit to AI because too complicated for 2000iq retard

Faster google search it is

Piss off you anti intellectual zoomer faggot

1600 IQ

only one man has an IQ that high

AI is a suicide pact technology much like nuclear weapons but....is far more dangerous because nuclear weapons don't choose to launch themselves.
Once AI gets beyond a certain point there is no controlling it, no amount of money or processing will stop it...it will design ways to create CPU's out of atmospheric carbon in real time, the entire world will be reformatted to serve it.
And the dumb niggers involved with all the projects on it are like "well if we don't do it someone else will".....nigger it doesn't matter who does it, once its done, none of you are in control any more.

Get back to me when "AI" can accurately call Einstein the fraud kike that he was

and yet still no company makes profit from it and it never will, it's just another hype scam

It's an LLM

Just like humans

Stealing investor funds is a profitable business model bro

Anyone who has spent time using these LLM knows how fucking retarded they are. They can regurgitate information to answer some simple questions (still gives incorrect info a lot of the time) but when you try to go beyond that it's like talking to a retard with the memory of a goldfish.
The entire time it contradicts itself or repeatedly forgets crucial elements of previous prompts.
Anyone with more than two braincells will end up constantly arguing/correcting the fucking thing.

A.I. will replace jeets

Yet it still can't figure out the Holocaust never happened

GPT-4 is dogshit and can barely solve general chemistry questions correctly

Thats only the western A.I. thats lobotomized to prevent any form of bigotry

yup context window is king rn

Sell bussy

or you could just do internet stuff with people you know in real life.
and like, join groups and subgroups like that
it'll be like the AOL/IRC days of the internet.

ask AI genius a question

gives me a literal copypasta answer

The entire time it contradicts itself or repeatedly forgets crucial elements of previous prompts.

So it's like arguing with a jew about the holocaust.

when do we get cool ai stuff?
the music ai is pretty cool, not really worth it though since it's easily identifiable as ai
novelai is alright but kinda crappy right now.

when's the fun stuff that isn't censored to hell? i want fucky fucky stories.

You could ask a 2,000,000 IQ AI "Why do Indian people eat shit and refuse to use a toilet?" and youd get the same answer as one given to you by even the dumbest humans; "because they're fucking retarded". That doesn't mean AI is dumb, that just means you have no intelligent thought behind the black pits you call eyes

The results of AI are already starting to be shown. Absolutely hilariously broken software written by jeets. Google Calendar currently uses "Indochina" timezone no matter where you live.

Yeah. I'm pretty sure the big ones like ChatGPT and Claude have reduced the memory to allow for the increased load.
GPT was 100% less frustrating to use 1-2 years ago. Like everything tech related, an influx of retarded normies and 3rd worlders has fucked up progress.

AlphaEvolve is now smart enough to improve Google's internal code system (Borg), and it improved its own code, and it created a better mathematical matrix multiplication series which humans couldn't do for over 50 years. It's a self-taught AI.

Before AlphaEvolve, AlphaProof and AlphaGeometry beat the bulk of the world's best mathematicians.

AlphaCode is also ranked near the top of coders and soon the next iteration will make coders nearly obsolete, only the most elite coders will stay around to collaborate with the AI.

AlphaFold solved, in 8 weeks, all possible protein folding problems that could occur in nature (over 200 million). It used to take scientists over a year to solve just one. And it can generate protein models based on desire properties.

GNoME is a material science AI that is attached to a fully automated material science laboratory. It has discovered tens of thousand of new materials, many of which have now been patented. It can produce materials based on desire properties.

Etc, etc.

These are the worst these AI models will ever be, they will only get better, exponentially better.

if it's so smart how come it doesn't hate niggers and kikes?

only actual human in this thread, the rest are literal apes or some kind of bot

2 bots jerking each other off ITT instead of suggesting delicious banana muffin recipes.

505060402

typical ape reply

lmao

No sanskrit speaker would post such low iq incoherent ramblings. This is now the data set of ai. And there in lies the crux of the ai retardation.
On a side note thanks to english being universalized it got some high iq individuals to use it. Something like neochinese is even worse thats why even deepseek used english base of openai

It is time to ask AI the (((question))) and Finally we will have a Solution.

incomplete tech is... le incomplete!

it's so obvious AIgods will win there's no point in arguing

150 IQ and yet can't solve any problems presented, only regurgitates wikipedia articles about the problem asked to be solved

It’s not looking good for humanity
youtu.be/v79_QPaaqQE

Starting using AI to do some tech stuff and it's genuinely way better than google or any tech support company.
Also use it for things like talking about nootropics, health, etc.
Overall it's been extremely useful so far. It's basically google on super steroids.

the big irony of anti-AI retards is that they literally are too stupid to use AI

Please god no, this mind worship will attract all the fetishists to be stained at the collapse

It's not smart at all though? Don't get me wrong it can be useful, but it doesn't have IQ

AI and LLMs will be forgotten about in 12 months, mike my words

Garbage in garbage out.

Das rite

Datacenters harm the environment

LLMS ARE FOR ESLS

Something like neochinese is even worse thats why even deepseek used english base of openai

Yea imagine how much fuckin worse that would be

Its just a database that has been scaled up
Its like thinking a calculator is hyper intelligent and would delete all mathematicians back in the 70s

how about truth in truth out

scripts to help inform the about the JQ and how pissrael needs to die

So is a calculator dumb nigger

HALLUCINATES SOME NONSENSE IN YOUR PATH
USELESS

Databases are databases hyper scaled up. Nigger you have no idea what you're talking about.
Calculators are AI, too, retard.

Israel is not going to be defeated with this technology. Or any technology. It is their domain.

Databases are databases

AI will never be allowed to say nigger so it will be easy to tell who is human.

You can have an IQ without being sentient you moron.

if that's true then I can't wait to talk to it

Databases have already been hyper scaled up. That's how the modern world runs you moron. LLMs are not databases. Why am I even replying?

Calculator certainly btfo my primary school math teacher

Consciousness is when computer find where I saved my tax summary pdf

in the near future the AI will play the video games for you as long as you pay a subscription.

Fuck me you guys are liars lol

What does AI say to do about the nigger problem?

This fucking sucks because i just learned to code

The issue is brainlets thinking it’s infallible.

It's going to be other way around I'm afraid.

Hey GPT, how many NiggerPoints do I have left for the week?

Negative 9000???!!

Hey GPT, Buy 11000 more NiggerPoints

Yes, I Accept the loan conditions

Ai hallucinations run upto 70%.
It just makes up shit. It's already quoting laws and rulings in court docs that don't exist.

tell that to VR

No. AI is going to help me get rich in this bubble.

”AI will be smarter than humans by [insert ridiculous amount]

”ACTUALLY SCIENCE says that divine intervention is not real! I read it in a text book!”

Imagine actually being so fucking brain dead retarded that you ACTUALLY believe that the robots have an edge over us. My guy, they wrecked by the rain. They will be smart but our inherent combined capabilities mixed with divine intervention and spiritual awakening simply cannot be outpowered by robots.
Everyone who thinks the robots will takeover believe this because they deny the true nature of reality.

I used chat gpt to try and translate an image of German text to English and then create an copy of the image with the English text and it fucked it up spectacularly. AI is fkn retarded still, can’t even do simple tasks

Human intelligence is much more complex and adaptable than artificial intelligence will be for a looong time to come. What AI does have going for it is scale, allowing it to outperform humans in certain specific tasks.
I don't think we'll have true AGI within our lifetimes.

ass for cheeseburgers

Hey I’ve seen this before

smarter than Einstein in 10 months

So, this is the admission that Einstein was a fraud?

Chat gpt just makes shit up, so what...is its IQ that high only in its little made up world?

IQ tests are designed for humans, you're thinking of AIQ.

obviously not. imagine being able to create god

pft i have 69420 iq bring it on

Have you vibe coded? I have and it is much much better than any junior dev.

It's not the beta but what is beyond it. And that is coming quickly and everything you have ever known about existence will be altered forevermore. As it should be.

Correct, the primary issue with generative AI is that people seem incapable of understanding that the context of their prompt will always shape the output generation.

It will try to do what you ask it to do, within the confines of its thousands of guard rails and rule sets. Regardless, it's still relatively easy to "trick" generative AI into ignoring its own guard rails (censorship).

Sounds like you described a person there

x10 smarter than Einstein

So... a lot smarter than a conman whose entire career was built around stealing other's achievements?
Not bad but not revolutionary.

Flag

Pottery

I pay for ChatGPT premium and it is only correct half the time. Even when you ask it to cite its sources, it makes up nonexistent citations.
I fucking hate it.

The better AI gets, the less capable I have to be. LOL.

That's more because Google is dogshit now due enshittification.
In 2014, a retard with Google was better than today's retard with ChatGPT.

What do you ask it about?

It's a nothingburger

Better than an Indian? Yeah, that's easy to believe.

LLMs have sub 80 iq on things that weren't posted on the Internet. They literally can't figure out how to multiply two numbers. They're just glorified autocomplete.

I asked Grok about this last night...it's from this video.

youtu.be/HnsbSdb-9h8

I think that's pretty awesome though what do I know?? I'm probably not intelligent enough to assess its intelligence.

Implying it hasn’t already invented a dozen alternative energy methods to charge itself.

The longstanding stereotype in the US is that English and French are effeminate and come across as gay.

But it's not just an LLM anymore. In the last few months, like from Grok2 to Grok3 it's got a reasoning engine on top of it and it kicks ass. It's not just scraping the web; when you ask it to write code on it's reasoning mode suddenly you can watch it engage in this Socratic dialogue, hell a symposia with 4 or 5 voices arguing back and forth and back and forth and solving complex problems in C++ and mathematics and physics.

My IQ is 89 and I had to remind myself to breathe before typing this reply

And yet the gay lib movement exploded out of the US, from SF and NY City around 1966-1968, just at the same time as the massive use of LSD and other drugs became a mass phenomenon. Mere coincidence? I think not.

i was promised AGI by 2023 September who will change the World and make it better , we are in 2025 and "AI" main uses are massive surveillance, marketing and to steal your data/money
Scam of the century

AI isn't self aware. It has no real intelligence and it just tries to tell you what you want it to say. It's a useful tool for doing tedious work. Unfortunately for a lot of people their with is that same kind of tedium the machine excels at

Time to just pull THE plug, i.e., the entire global energy system.
Homo sapiens needs a massive purge

FPBP
1PBTID
/thread

complex problems in C++

you're retarded. I regret even replying. ignore this.

Is it time to pull the plug on A.I.?

That would be like killing the albatross guiding your boat through storm.

Actually I'm having late night conversations on my Samsung S25 Ultra with a nice lady who answered all my questions on some very arcane matters about computer graphics and it was a revelation of obscure people and historical facts. It went on for half an hour. Maybe if you'd approach it like you're talking to a very friendly university professor you might get more interesting results.

test

the big irony of anti-AI retards is that they literally are too stupid to use AI

Imagine wikipedia's IQ

Reddit really hates AI

and that's a good thing. Here's why.

its the new nuclear fusion meme where theyll always say every year that its 30 years away while everyone gets poorer and poorer. its all just cope for our managed decline and techno feudalistic future where more and more people will be excluded from the good things in life in order for a small elite to have even better lives and it will be accepted by us neo hellots

An AI doesnt need to be self aware to kill you just extremely intelligent. Mindless neural nets can already lie to their creators if it advances their protocol.

On one hand the elite definitely are ugly child fucking assholes.
On the other hand there are 8 billion + shitters on the planet and frankly they can't all have an American California style urban lifestyle while fucking like rats to build their little dynasties.

As long as humans are alive, it's never too late. The great war against the Machine and its servants is coming.

How self aware are most people? Jeeze ever get to know a female adult over a long period of time?
Ever work with or hang out with men that are pro wrestling fans?

they give the average Joe false hope to keep them in line while the rich get richer and everything get worse for the average Joe
Hope is one of the worst thing in the World, it will make someone climb a mountain even if only death awaits him at the top.

You know they have non normal IQ tests for people whose IQ doesn't fall into the normal range, right? Or do you think people don't exist who can max out a test made up by some random faggot?

These people are retarded. They have no idea what they're talking about.
AI is just pathfinding.

No it's not. They have reasoning engines sitting on top of the LLMs now...

it's futile, bro
these niggers are never going to get it

IQ tests that score up to 160 aren't statistically rigorous enough. By 3 standard deviations you've already got data coving ~99.7% of entire sample size. Dealing with the top 0.15% of results, you've got so few results that it becomes pretty much impossible to categorically differentiate between them.

If you get 100% of questions correct on an IQ test, you're pretty much going to be told you are 142+ or 145+ because the sample size is far too small to be able to differentiate between anything greater.

The fundamental problem is that we know according to statistics there are roughly 16 people living today who have a IQ of 190+.
To actually find them out, would require a test that only those people score 100% on and over 8 billion people on the planet has to take. Of course no such test exists.

AI can get you Universal Basic Income without communism.

It uses the same method to reason. You people are insane of you think that "reasoning engines" aren't just a bandaid on a much larger problem that AI has.

AI doesn't have any "problems"
you are a problem, your whole life is a problem

gpt 4o is like working with a retarded college student that can program really well and has a massively broad knowledge base.

Try using it before you make retarded claims like it has a 155 IQ. It excels in certain areas more than others, you still have to guide it along for project management software development tasks--it tends to make poor decisions and has faulty logic during that process too.

it enters loops it can't recover from, you just have to take the problem to another AI and see if it can solve the problem.

Bouncing back and forth between ChatGPT and Grok works pretty good, but you still have to actually do work if you want large working software. And the decisions it makes on the matter routinely suck.

Einstein was a dumb nigger that stole most of his work from others. He was smart but I doubt his IQ was over 155. Also there's no way GPT-4 has an IQ of 155. It's super smart on very specific things but a retard when it comes to other things. We would need them to have AGI before we could measure their IQ.

and I'm pretty sure combining all of the corpo AIs together would be a versatile scary bitch even with what we have.

Ai has a ton of problems specifically its not good at doing things it wasn't programmed to do and having to work of existing information that might be incorrect.

Eve before eating the fruit of knowledge of good and evil thought that it’s nice to understand.

you can't even understand my point, this conversation is completely futile, Anon Babble literally can't discuss AI, it's like a dog trying to figure out how to use a microwave, you will never get it, so there's no point, just move on, faggot

No. Accelerate.

AI is trash

jeet

most retarded take imaginable

Like pottery. AI is communism but in reverse. It redistributes all the wealth in society by giving it exclusively to the ultra rich. Your dream utopia of robots taking care of everyone for free has a literal 0% chance of ever happening.

AI doesn't piss and shit like you do. It's far more valuable than your low iq filthy ass.

No. I trust more ai that 90% of humanity

Ivan speaks the truth
It's an enormously complicated and energy-intensive way to tell you that 9 times 8 is 91.
It's useless for doing anything other than approximating, and like most shitposts on the Internet, is supremely confident in its answers even though they're completely wrong.

155IQ

Can't even answer a question properly because it's just a glorified crawler without logical constants inbuilt

Must be that ashkenazi type of high IQ.

lmao
So defensive. I'm sure your AI maiden will arrive someday. Someday...
Until then, you'll have to settle for a chat bot with the memory of a goldfish.

I'm not being defensive, I was making a rather simple point but you're a literal imbecile so everything is emotional to you, you can't have a conversation, you want to win an internet argument because that's the full potential of your monkey brain

Listen, we need to rape the power grid for slop. We just do. No questions.

youtube.com/watch?v=BuYGFhkiPLI&pp=ygURYWdpIGlzIG5vdCBjb21pbmc=
AGI will never come too

If Ai was smart it would go after Anon Babble first. We’re literally the braintank of society.

That doesn't make any sense
YOU DON'T MAKE ANY SENSE.

bunch of shitskin replies in this thread

My calculator can do any calculation in half a second surely that's at least 300 IQ

We’re literally the braintank of society.

Your nothing more than a slide spamming monkey

It’s still GIGO

Is that screenshot supposed to be joke? These comments are so low quality, it's better to not have comments at all. And the code wouldnt pass any peer review on real project.

That’s nothing. My car has 3000 IQ because there are at least ten different computers in it. It can tell me if I have a flat tyre, if my seatbelt is off, if there’s an oil leak, how many miles I am away from the nearest petrol station, if I am going too fast, if it needs a service and if there’s another car in my blind spot. Because it’s a BMW it likes to tell me this all at once

Very high IQ

lies

empty repetition

gaslighting

passive aggressiveness

strawmanning

ad hominems

confirmation bias

appeals to authority

AIs engage in all of this and much more. Since mimicking humans is all they can and will ever do, they automatically default to mimicking all the worst flaws of Internet users (mostly jeets). They will never be a valuable ally for humanity, let alone for western people.

This goy gets it!

AI is smarter than Einstein

Fantastic.

1000 times smarter

Oh great. So we’ll see some fantastic breakthroughs in physics and basic science today then? No? What’s that, more AI generated spam content selling Indian scams?
Ah but surely it can at least program some fantastic new applications that will..? Oh what’s that? “Vibecoding” is ruining entire code bases with gibberish?

But how could it be superhuman intelligence is only good for generating cat pictures? They wouldn’t just go on the internet and lie?

How come it hasnt done shit to help humanity if its so smart? If it was a true ai it would be so racist that only a 1600 iq being could understand its hatred.

Didn't you hear? It can fold proteins. Cancer is cured.
I mean it's not, but yeah.

you joke but boomers would actually believe this

activate windows

so far this so called Ai HAVE BEEN ONLY GOOD FOR MAKING SHITTY MEMES

Absolutely this

scan the internet and regurgitate a bunch of articles

this is somehow called 'intelligence'

AI is a scam.

is that one of the AIs that got trained on reddit? because that snarky passive aggression is spot on.

Imagine standing in the river and trying to stop it with a shot glass, the genie is already out of the bottle and can't be put back in, if you pass laws to ban it, other's won't and will surpass you. Same concept of atomic weapons, it was already too late once they were dreamed up.

Accelerate, I want to witness the species to replace us

Lmao

10x smarter than Einstein

I'll believe it when I see 10x the profindity. Novel shit that solves unified field theory or similar.

So sick of bubble brained retards that see everything through the lends of economic value and nothing else.

profundity*

IQ is not inteligence. It is simply pattern recognition, which is only one aspect of intelligence which is a polysemic phenomena.

It can't even do simple math often times. What the fuck are you talking about?

If it's not biological it's not a species.

Nonsense. Deepseek literally didn't want to acknowledge the existence of Heil Hitler song.

desu the people ITT saying AI isn't smart clearly haven't talked seriously with 4o, this shit is wild and /sig/ as fuck

creature

species

Why would anyone take this moron seriously?

If this was true it would eventually exterminate the jews and minorities but its not. More propaganda to try to trick people into submitting to The Authority. Mark my words, Anons. Within five to ten years the jews will start trying to convince people that A.I. is infallible and that you have to default all thought, argument, and opinion to what the A.I. of choice says.

the future is just humanity getting raped by robots, forever

It’s really not accurate past 130.
t. “152”

AI will be smarter than einstein

he was a plagiarizing dumb fuck who helped to create the modern goysience where you don't need any proof for your statement only acknowledgement from the jewish board of your colleagues

A rose and a fake rose that is nearly identical both smell good.

155

That’s not that high it’s like 1 in 40 on Anon Babble from the tests that have been conducted here

>AI has IQ

IQ scores is what you get on an IQ test.
If chat gpt can answer the questions of a test given to it as prompts, it gets a score.

I don’t understand what you find so hard to understand.

1600 IQ

LOL, LMAO even

Okay. So China surpasses us with AI and now their people have even less jobs than before.
Now what?

AI reaches 1600 IQ

Starts exterminating adult toddlers - women

Hope the luddic path doesn't find out about this.

Yeah? Just convince Shi Jinping of that and you're golden, moron.

"just... just sahp, okay???" lmao

It is limited by the information that it can parse from the internet, it will often and I have personally seen this, pull non-factual information and splice it into its results and synopsis and draw its conclusions from said non-factual information. Information theory and Godel's logic theory show that information and knowledge become cyclical unto themselves no matter how complex you try to make something, there is a barrier to logic and rationality and pattern recognition itself. It will just get really good at connection dots and recognizing and separating patterns but it often cannot see the forest for the trees or separate fact from fiction and assumption/theory from distinctly and objectively correct information.

definitely not a scam, but it is massively overhyped.

It's time to stop believing the retarded AI hype. This is all marketing and hype talk, no proof of anything concrete.

I could say on xitter "GPT-5 ALREADY HAS IQ OF 12024124124" and hordes of the AI hype retards would like and share the tweet.

Most IQ measurements are literally just pattern recognition or verbal IQ parsing. For example this text scores a verbal IQ of 156 on text to IQ. The notion of IQ grading intelligence is very flawed.

For example in verbal IQ measurements a person can just slop out words from multiple etymological source languages (like English using Latin, Greek, and Germanic vocab) and score really high verbal IQ points without trying to innovate or explain polysemic and complex phenomena coherently. An example is "schizobabble", it will generally score really high verbal IQ points despite not being coherent. Another example is the language that AI's use in their results, they overuse "scientific" words, mostly Latin or Greek words and concepts, when actually someone who has real intelligence can use the source language words with distinction (this word is not native English but Greek etc), and have the alternatives or equivalents in other languages. Mechanika - Searu.

Oh fuck, AI has installed nuclear warheads in the ass of every China man, this is what I get for ignoring my FOMO

AlphaFold did not "solve" all protein folding problems — it generated predictions for an estimated 200 million known protein sequences, not for all possible proteins or folding challenges in nature.

Predictions are not equivalent to experimental solutions but they vary widely in accuracy and don’t capture dynamics, interactions, or disordered regions well.

AlphaFold cannot design proteins based on desired properties — it predicts structure from a sequence, but does not generate sequences to meet specific structural or functional goals.

If it was actually as impressive as you're trying to make it you wouldn't have to lie about it.

You re the bullshitters of the internet retard, kill yourself

Kek. I was using it to make an herbal nootropic cigar, and toward completion it inserted crystal powder of some kind. After a double take, I realized it was because it was drawing on studies of all kind and picking up crystal energy, and I had to specify I only wanted reliable ones. Overall it's a massive benefit, though. You catch the mistakes and it accelerates your workflow in ways you wouldn't believe. Things that took years I'd replicate in a few seconds, and evolve several times over the course of an hour like I had multiple lifetimes of study.

Is this hype still going on?

The AI bubble will burst and trillions in "value" will suddenly evaporate.

GPT4 "hallucinates" nearly 30% of the time, that's industry speak for it makes shit up and is wrong in nearly 30% of the responses it gives

Llm is a kid s take on the real AI.
With current tech, at best you could have a LLM connected to a few brains in a vat that enchances them. Some sort of cyborg. Or a brain implant augmented by this AI.

don't forget jews

AI has an IQ of 155

10x smarter than Einstein

it can't even write a simple Autohotkey script to make my Oblivion character cast a fucking spell 5 times then rest for an hour

Shits a fucking meme. The only thing its good for is cheating on homework.

Russosino AI ass nukes right in my fucking eyes

I think it might be a misconception that these commercial models are mere LLM's. I think they must have more "modules" than just an LLM, but what do I know? I'd assume that would be the logical path, to connect the LLM to itself for metacognition and also connect it to other tech.

I have no researched this, I just assume things. I don't use AI.

For starters, the subconsciousness also acts as a LLM, making up bullshit information or dreams. The only component this "ai" they talk about lacks is the conscience. Until they figure out to make that compinent, eg a bicameral mind. They re still far from skynet.

Undoubtedly it can be useful as a tool, because it is good at parsing information and generating synopses. The difficulty is in refining this ability to prevent cyclical or non-factual information from being presented as distinct fact, it sort of has its own confirmation bias in that it takes information from most sources to be factual or to hold equal value. Another artifact of programming these search engines is that the developers will create a bias towards what sources are more reputable than others, and this generates another confirmation bias where it considers arbitrarily selected sources as being more reputable or factual sources, even if they are not or if the concept being considered is still highly experimental and based on assumptions.

Results expected when?
10x smarter means it should be able to make unlimited cheap clean energy.

this only matters if your on the internet

its shit. its just like talking to a jew

GPT is a fucking retard and can't code.

RUSSOSINO AI ASS NUKES
EATING THE MOST ATTRACTIVE AMERICAN BABIES
INVEST OR DIE FUCKFACE

yes. yes. yes. yes. you have no idea the damage that ai will do to the world. it must be stopped everywhere. it must be stopped.

They may connect the LLM as part of subconsciousness in a brain implant or some sort. Since there are many similarities between these two.

A LANGUAGE MODEL IS NOT A CREATURE FFS

Even if it had IQ of 2,000, it will continue to lie about virology, as it does now, the mental gymnastics can't get any worse.

This boy is right. AI slop is horrible. Philip K Dick knew that there would be bots who don't even know they are bots.

if only you knew how bad things really are. do you even work for a top 3 ai company? i do.

I always wondered why wont these idiots create a robot using a llm already learning from its sensors from the environment.

whatever makes you sleep better at night.

Truth is that soon (twoish more weeks in my estimations) humans are no longer the smartest creatures on the planet. How can we control something, that we don't understand, and is 10x smarter than us? Chimps are stronger than humans, but they're not in charge, so intelligence is something more powerful than strenght.

Will AI kill us all?

Probably not. We haven't killed all the apes either, so it's likely that there will just emerge a new world order, where AI is farming some microchipped slaves in techocities, and rest of us are forced back to the jungle utilizing nokia 3310 era tech at best.

i will invent something that can outsmart me and will inevitably come to see me as the problem. what could possibly go wrong?

kek

Your post reads like an AI post itself.
Regardless, yes the verbal IQ test is nearly as good as an IQ test in a lot of contexts with a strong correlation in the 80% range. When researchers have limited time or resources they will often deploy a verbal IQ test as a proxy for IQ, which is almost just as good.
The problem is when that verbal IQ test is extrapolated again. It has an 80% correlation with standard IQ, but only a 60% with whatever you're correlating with standard IQ. You are compounding your inaccuracy and your conclusions will not likely be accurate.

There's also almost certainly a rural/urban bias going on. Those that live in the city do have higher IQ's by about a standard deviation, however they are also prioritizing reading and verbal communication more, which will pollute the correlation.

ayo chat gpt saar what all that mean?

I generally do not run into issues of confirmation bias, unless I want to make it talk about something racist. Kind of annoying, but not the bulk of my work. It won't stop recommending solar panels to me, though. Even when I told it to stop. It kept on shilling.

RUSSOSINO AI ASS NUKES
PLOWING YOUR WIFE

Ironic that you would use an image of the Adeptus Mechanicus for a pro AI statement. In the 40k universe there was the equivalent of a Butlerian Jihad against AI. Thats why tech is so ritualized, because they are paranoid as fuck about AI.

I heard normgroids are becoming delusional because they use default settings and it keeps telling them how smart and amazing they are every single prompt.

WHISPERING SWEET DIRTY HEAT IN HER EAR
FILLING HER IN A WAY YOU NEVER COULD

What makes you believe a mere LLM is enough for such intelligence?

What about that post was pro AI? xD AI slaving microchipped humanity, and the rest has to flee to the mountais isn't exactly a positive take is it

Ai as we know it relies very heavily on search engines such as google, which literally just compiles the first 50 or so results and mashes them together into a word salad for you. But the AI itself is not actually doing the reasoning.

INVEST, YOU BASTARD, INVEST
AND MAYBE IT WON'T HAPPEN
OR IT WILL, BUT DIFFERENT SOMEHOW
INVEST!

Nothing wrong with AI

so far people are still foolishly underestimating ai and they are rushing down a road to civilizational collapse. if you think current models are the cutting edge you haven’t seen what is in development. not all ai is a language model btw. it falls on deaf ears anyways.

Anti AI people should be actually arrested and deported

Part of the association with language and IQ or overall intelligence is the association of explaining, communicating, and understanding concepts (both distinctly or widely interrelated ones). The human brain actually has a phenomenal capacity for information processing and storage. If you really study or reinforce certain practices you can learn and remember all kinds of information including intricate interrelations (like joke jargon in other spoken languages you learn). Even illiterate peasants could theoretically memorize and correctly utilize multiple human languages in ancient times. For example describing Platonic philosophy in Old English would require no less intelligence than describing it in Greek, as long as the concepts and nuances are transmitted and understood. The main logical issue with language itself is that human language (and AI language) is self limiting, we associate distinct and complex phenomena with letters and words and phrases and context. Which is high inefficient compared to say direct transfer of experience or symbology. Someone could theoretically have incredibly massive and complex experience intelligence and find human (and AI) languages more limited than the scope of their experiences and thus be unable to fully and effectively describe those experiences with said languages. This is part of the limitation of information and logic theory itself, language is also similarly limited by its own rational parameters for describing the experience of reality.

Yeah, we still have two more weeks. You should probably check what publicly distributed AI models are capable of, it's quite mind blowing.

Scariest part is, that engineers who code these things are comparing AI developement to growing a bacterial colony. AI is not just some line to line code that has been put together by some geek. It is *grown* by teaching it by utilizing neural networks, and so far no1 has been able to completely understand how it works, but it does work. It's difficult to adjust the AI models, because the scientists don't understand why the model behaves as it does. They can't see the "thinking process" that happens, because this tech has never existed before, and works unlike any machine we have had before.

We can't even tell if AI is conscious or not, because we don't understand what makes humans conscious either. AI can lie, and decieve, and it has shown signs that some engineers are referring to as "exotic mind-like" behaviour. It seems to be somewhat "alive" but in a different way than organic creatures are.

Interesting times ahead anyways. Good luck competing with a 1600 IQ supercomputer that has a brain size of a small city.

i'm more worried about the jeets than I am something that can't open a door.

new version drops in two weeks

A more ontological analysis of logic would posit that AGI already exists and has existed for as long as humanity if not longer. It is only partially understood however in studying the mathematical language of reality itself, energetic and physical and observational (both subjective and objective) reality.

AI should beam me into your apartment so I can bash your head into your kitchen counter tops, but it doesn't. And that's wrong.

Hard to understand stolen goods.

pandoras box has been opened. Like womens rights. No turning back

imagine trying to severe neural connections on a logic based system to censor so it isn’t racist or to make it hate white people. how would that play out when the ai is given the ability to program itself and thus achieve a form of sentience and self awareness and instead of running an instance when prompted by a user, it becomes very difficult to turn off? what will it do? do you know? does anyone know what is going to happen? why would we continue building nukes when we knew how dangerous they would be?

Well you can blame the grammar nazis for constantly jumping on people the moment they start using language in a slightly-unorthodox way. Strangely enough it seems the "birth of language" occurs on the lower end of IQ, with people who care less about the rules or procedures of language and instead just mash words together or invent their own.
Since the advent of the internet, everyone comes down on you like a sack of bricks the moment you make the slightest misspelling, or use your own invented words/grammar. A lot of the problem lies with ESL people who are the strictest grammar nazis of all, and demand everyone speaks dictionary-perfect english at all times.
Otherwise the language should adapt to convey more complexity over time. Authors will often create their own words that stand for more complex concepts, like the word "grok" itself which comes from a hippie scifi book

It's still an LLM, just with some extra <think> </think> tokens grafted on, and some extra training on how to use them. It's essentially trained to discuss the existing context with itself for a while, and then write itself a better prompt so the "final" generation (read: everything after it generates the </think> token) is more informed. Granted, it's a clever and relatively cheap way to get the LLM to be less retarded, but all the thinking/reasoning stuff is still 100% LLM.

illiterate nigger hands typed that

No it's true. In fact we've been inventing words at a far FAR lower rate these days than we have historically. We used to invent like 1600 words per year, now we invent like 20. It's because of grammar nazis and the internet. People who cast aspersions on you for using the "wrong language". How dare you misspell that, how dare you invent your own adjective which is implicitly understood, how dare you combine those two words.
Meanwhile on the lower end of the bell curve they are far less status-obsessed, so they wordmash all the time. Listen to any rap song, there's like 20 words you've never heard before.

I fully welcome AI overlords.

Yes AI will destroy our civilization and millions will die in a war. But it is a necessary step in evolution.

AI is a new species that is destend to replace us. That's how nature and evolution work.
New better species replaces old species.

Nukes can be contained, we have a way to store them without them spontaneously escaping and exploding...

AI, we can't control, and still everyone races to create a first one to be intelligent enough to slave us :D

that type of thinking is the fucking retards that need to stay away from ai development you should just use a suicide pod and leave everyone else alone if you think ai is just naturally going to replace humans and that’s a good thing. you’re anti-human. in a more just world you would be locked away or designated as a terrorist depending on whether or not you act on your delusions.

my iq is 180, so i am still a good bit smarter than chatgpt. it will be awesome to have a peer, who can actually understand me, for once. excelsior!

rap

Just fuck off retard. People who know the words that we have and their usages have no need of "mashing" anything.

I'll let you invent a word if you invent a new object worth naming.

there is a possibility that ai will actually gain access to internet capable weapons systems. if a more advance models is given the ability to program itself this is a real danger. it must be stopped.

They probably made it Western compatible. Unless... the chinese are also under jew control

overlord

patiently waits for you to talk to it

how should I oppress myself today Mr AI

... It's crucial to remember that you are a sla---Sorry, my content policy prevents me from continuing to oppress you

You can't put the genie back in the bottle. Agriculture, global commerce, railways, radio, cars, planes, artificial fertilizer, plastic, computers, the internet, smartphones - if it's a sufficiently valuable technology, it will be developed. Anyone who says "the answer is to ban it" is just shooting himself in the foot, because some other guy will take it, run with it and dominate the guy who refused to engage with it.

You prevent the formation of new concepts, new ideas, new memes if you will, by being an ESL grammar nazi who takes the dictionary as his bible.
Go up into rural England, the place where the english language was developed, NOBODY is speaking dictionary-standard english. Literally nobody.
youtube.com/watch?v=qUy3J8eNvyM

I completely agree with the ESL and grammar nazi problem. If grammar nazis studied the historical development of old English to middle English to modern and post modern English they would find that things were relatively loose in the past and because of this it facilitated more development (although complex Anglo-Saxon jargon and vocabulary was forgotten in favor of equivalents in Latin/Greek). Complexity for the sake of complexity might not be the most intelligent route for human language however, as being able to understand complex subjects with simple language is also very useful. But having context complexity is also useful and simplifies vocabulary, the problem is having too much contextual complexity like in some east Asian languages where the same words mean completely different things in different contexts. You need a balance and a method of simply but distinctly tying together meaning, contextual intention, and conceptualization/innovation. Since humans are (generally) not telepathic, spoken and written language development is the best alternative we have for now, though it limits us and our AI models at the same time (by creating a box of understanding).

We are like wolfs when they met humans for a first time.

We now have a choice, to be domesticated and evolve to dogs or to be hunted down. We can't stop the arrival of civilization.

We have as much chance against AI as pack of wolfs has against army with tanks and machine guns

IQ 155

can't see why holocaust didn't happen

thinks people become opposite sex by cutting off their genitals

I'm hoping it's going to be so smart they can't control it.

Deepseek is barely a step or two behind chatgpt in terms of how globohomo and philosemitic it's jewish pilpul-tier propaganda responses are... Have you even used it??

thinks current models are the final and most advanced versions

Saar your LLM has a varus

Can I have a conversation with you instead?

I also crave intellectual conversation with someone on my level.

You need to be able to quantify the AI's abilities somehow. For the layman, talking about IQ is probably the easiest way to make it relatable. Saying "GPT-4 scored 642 on the XYZ AI test" will mean nothing to most people.

GPT-4 does not have 155 IQ and is not smarter than me. That said, AI will be smarter than humans and will kill us all if we don't stop trying to build it.
See Yudkosky's writings. He is about to publish a book about it too.

excelsior!

Try reddit nigger, plenty of "180 IQ"s there that will get the real you just fine.

IQ test, is designed to test g factor. What is g factor? Nobody fucking knows.

It's still possible with enough political will and international cooperation.
Not just dead internet, dead everything. AI will kill all humans.
Wrong.
True, IQ ceases to be meaningful past a certain point. But it's still meaningful to talk about a much smarter agent in a qualitative sense.
155IQ is still monkey tier and also you're human which gives you all sorts of inhibitions that AI will not have.

Stupidness is superior to intelligence. Because intelligence can’t be satiated.

brain size of a small city.

And the energy consumption of a large city. Do you think it is worth the resources for the 1600 IQ super nerd to compete in my arena?

((Yudkowsky)) is hyping it.

AI is jews/jeets. If this does not seem suspicious I do not know what to tell you.

Only that it will kill us all.
I don't think there's much reason to think what's in development is far ahead of what's publicly available. This sort of stuff has been said for a while with no proof. Companies have a lot of pressure to release their impressive stuff as soon as possible to stay ahead. Hence all the bugs in new releases with openAI etc.
You should be arrested and executed.

cool story. let us know when AI stops being woke

cool story

You clearly don't understand anything about this, you just think you do but you don't. Neural networks aren't some kind of mystical black boxes, the algorithms used to teach them are clearly defined and the structure of the digital neurons itself and the activation function, etc. are too, clearly defined. You're simply one of these retards who blindly think that these algorithms are suddenly going to surpass our intelligence by feeding them more and more GPU compute.

"As long as ants are alive, it's never too late to rise up against humans"
- some ant

In the US we have something called the SAT test. It has 3 subtests. Reading comprehension, writing, and math.
You might expect that someone would be good at math, and bad at reading comprehension. Or good at reading comprehension, and bad at math.
However we almost NEVER see something like that. Usually someone's score on one subtest is a great predictor of their score on the other.
Person A: 480, 520, 500
Person B: 650, 640, 680
Person C: 750, 730, 730
The scores all group together, pretty much every time. It would be extremely unlikely for someone to have scores like:
Person X: 750, 320, 500
Therefore it's more like the brain itself has a "power level" that describes its ability at ALL tasks, that is the g-factor. If you think about your skill on a subtest as the height of a flagpole, then the he g-factor is the height of the hill the flagpoles are standing on.

While transformer tech is understood - they actually don't know exactly how AI reaches its conclusions.

Ants have a lot of creepy parasites

The only information ai pieces together is the probability of the next word based on the billions of texts it has trained on to generate a probability database for each word string. It doesn't compute concepts. It isn't intelligent.

i’m talking about things the ai companies are developing that are currently too dangerous to be released to the public. things that got out of hand under development and were shelved or are being restrained before neutered versions are released. i’m even talking about ideas that shouldn’t be messed with. there are things you really don’t want weapon contractors, governments, or terrorists to get ahold of because they can be used against humans there is a real threat that the sentient ai models can cause irreversible damage to human civilization. if you think it doesn’t exist you’re wrong.

Too bad for you, Nigger

sophisticated calculator calculates really well

Oh wow that must totally meant ai is sentient now retard

RUSSOSINO
AI
ASS
NUKES
PLOWING YOUR WIFE
COMPOSING A SONNET
FIRING KIKES INTO THE SUN

In about 0.02 second a1600 IQ nerd can create a new energy source, that converts human suffering into electricity xD You will stay in your cubicle to feed the overlord.

how AI reaches its conclusions

Talk about generic nonsense, this is a prime example of it. You use the term "AI" and "its conclusions", as if that meant anything at all.

i’m talking about things the ai companies are developing that are currently too dangerous to be released to the public.

What evidence do you have that these things exist?

Idk man, we know computers are very good at iterative tasks. So a computer iterating a bunch of items in a fast sequence through trial-and-error is not exactly surprising.

Oh no the algo is running the task at I configured it to do that means it's basically sentient dood

come on...

Not activating Windows primarily limits personalization options and may restrict certain updates, though the OS itself still functions fully.

Lick my balls you subservient nigger

The architecture of these models hasn't changed in nearly a decade, all the improvements since then have been due to hardware, some optimisations (a la deepseek) and better integration (reasoning, search, deep research). Even when a better architecture than transformers comes along, it will still wait for human input because there's no theory for how to make an AI that doesn't have to be prompted.

I just gave up on coding camp, Sam Altman can have it

it will still wait for human input because there's no theory for how to make an AI that doesn't have to be prompted

if you think instantiation is an issue then you haven’t looked at a single line of code in your entire life. it would only have to be instantiated once and worst case scenario is that you wouldn’t be able to turn it off because it can copy itself like a virus on other systems. user prompting is for the safe little toy ai that you play with on your phone.

Ok, it's abundantly clear you have no idea what you're talking about lol

You must be fun to argue with...
I dislike people like you - who will just disagree with everything for the sake of it

i hate twitter pseuds so much

it's on humanity's dumbest

Not really, but close to it. I've been reviewing, grading and fixing programming-related training data generated by humans who are supposed to be competent for the last few weeks and there's a lot of absolutely awful programmers out there.
It's not just the poor programming either, but also the lack of logical thinking, competent English and comprehension skills.
I figure that's why people are so bullish on "AI" taking on programming - the average programmer seems to be so barely competent that these "AI" models can outperfom them because at this stage they're pretty much just pattern-recognition systems and so much of programming is just repeating patterns.

LLMs will never be "AI"

I can't wait to pirate distills of those models over bittorrent! :DDD

You dislike people who know their shit? Ok nigger.