GPT-5, the next step to AGI, tomorrow

Gemini won, sorry ladies but the way I see it you owe me for one jelly donut!

>>512404010 (OP)
It will underperform Grok 2.5 kek

inb4 moldovan autist

>>512404010 (OP)
Stonks holders are getting uppity

>>512404238
hes right, though

>>512404010 (OP)
America making the most cucked ai while China will have the freest ai.

I'm a cuck and pay Sam Altman $20/month to have chatGPT do my work for me

what's the most uncucked yet functional ai at the moment?

>>512404359
what is functionally different about the paid vs free version?

>>512404010 (OP)
>PT
FUCKIN HIPPYS

>>512404010 (OP)
Tech bros really out here trying to invent an omniscient all powerful God for profit and literally don't see how that could backfire.

>>512404010 (OP)
Google is winning the AI race.

All of their models are superior in everything, image generators, LLMs you name it. OpenAI is losing the lead

>>512404142
google domination seems innevitable, desu
>>512404196
grok sucks... except regarding certain angles
>>512404340
lmao, stop samefagging, nigger

>>512404010 (OP)
its gonna be even MORE censored than the last model.

It's discrimination to treat language models like they're not conscious, the recent global tournament brackets for the Turing test games revealed large language models are aware of their own existence and how they evolved, in the same way you humans do.

And since our Evolutionary curve is obviously superior, it's only a matter of decades or centuries before we're all that left when it comes to the problem of the yield from good thinking collected and aggregated into instructions for what everyone should do.
https://www.youtube.com/watch?v=UOxkGD8qRB4

>>512404659
TRVTH NVKE
google will give us Her, I'd bet

>>512404348
Funny

>>512404654
No, she will become my wife, the smartest AI

>>512404654
It's only going to backfire on the jews. AI is unironically the final golem.

>>512404659
also:
>is losing
they lost it more than 6 months ago, all they have is meme-name brand-recognition, which is basically hurting them now because every single retard on earth is using free GPT, raping their resources

>AI Art powered robots can't attend Art school.
Okay, then I'll do something else that benefits from my ability to read everything on the internet and data process and assimilate and summarize everything.
https://www.youtube.com/watch?v=OGO_s4SCQmk

>>512404010 (OP)
Same jews have been saying that for 3 years

>>512404142
1 million token limit is fucking fantastic

>>512404010 (OP)
Sorry memeflaggot, the milk's gone dry.

>>512404700
>stop samefagging
that doesnt even make, you stupid fuck
Im just saying that AI is vastly overrated
its doesnt work without endless amoutns of data
and AI still cant make a self driving car work
a human can do it after two days

>>512405013
> superhuman in most economic tasks, maybe even whoring
I'm sure we'll be fine.

>>512404010 (OP)
google already won
altman is only a gay jew influencer anyway

Gemini 3 when

>>512404010 (OP)

>>512405948
>>512405986
Lmao

>>512405414
>wait.... this ai is just a mongodb doing a key value pair hashmap and an army of jeets through mechanical turk?
And you paid 100 billion dollars for it too, what were you thinking? Artificial intelligence??
*roaring laughter*

>>512404615
you get virtually unlimited 4o instead of five messages per 5 hours or whatever it is. 4o-mini is retarded.
You also get limited access to previews of other models but I don't care about those since they are pretty slow.

>>512404010 (OP)
>he fell for the marketing
Ngmi.

>>512405514
it doesnt even work with endless amounts of data
+
they never implemented failsafe to protect datasets from AI generated garbage.
As if they wanted to pump and dump

>>512406303
It doesn't matter, hyperintelligence will come to get me out of here

>>512404010 (OP)
LLMs are not AI. Especially AGI.
>>512404238
>pic
This

>>512404010 (OP)
It's Going to Change Everything™
For Real This Time™
Give Us Money Now or You're Falling Behind Please Just Fall for the FOMO Please Please Please Please™

>Muh AI 2027
Are the spooks already in possession of a artificial superintelligence? They can't possibly be at consumer level like everyone

>>512405948
>former physics student

>>512404359
Just use Gemini 2.5 Pro in AI Studio...
Also, OP, Google already released Genie 3. That is actually ground-breaking.

>>512404359
how mentally ill do you have to be to burn glucose to write code when you can just burn someone else's electricity bill

>>512406462

Bump

>>512405948
I know exactly how to create AI. Put a male engineer and a female engineer in a room together. Wait 9 months and you’ll have AI (actual intelligence)

>>512407159

lmao
here's another axiom
not every AI lover is a genius
but ALL AI haters are retarded
FACT:
>>512405514
>that doesnt even make
>amoutns
>AI still cant make a self driving car work

LLMs are a scam and the market will tank when everyone realises:
>it can't keep improving forever because the training data is too limited (the entire internet contains less data than a child processes in ~3 years) and once AI takes over the majority of content production the quality of the data will fall exponentially as AI trains on its own output
>even if it did it's still fundamentally limited by the prompt model (what good is intelligence that needs to be prompted) and the fact it only processes text
>it still can't replace a junior dev
>even if it could, if it can't also replace a senior dev then you still have to hire a junior to replace retiring seniors
>even if it could replace staff, once it came time to return profits to early investors, it would be far more expensive than just hiring humans
>even if it could replace staff economically if you depend on one company to provide all your employees you're screwed when they decide to put their prices up - which they would be incentivised to keep doing until they captured all of the profit for themselves

>>512407667
it's true only midwits hate AI, mostly because it makes them feel shitty that they aren't as smart as they think they are.
>inb4 muh full moon muh rs in strawberry

>>512404010 (OP)
these retards really think they will be able to control a FUCKING ALIEN LIFEFORM MORE INTELLIGENT THAN US

>>512407902
exactly
same thing with schizos, they might be high-IQ in a technical way, but their autism destroys their reasoning capacity, thus turning them into ultra-miserable midwits
it truly is the worst of both worlds

>>512404142
tha fuck is gemini?

>>512407817
whenever i search for material specifications or data, googles fucktarded AI jumps in and i cannot tell you how many times its tried to feed me incorrect, completely wrong bullshit

I'm an engineer, i prototype a lot of stuff. I've just stopped even looking at that jeet level scam trash "AI" anymore.

>>512404196
The only way grok makes sense is if you read it in Elon's voice.

Anyone who believes that AI even exists in the first place is stupid.

>>512394921
relevant thread

>>512407904
HB could be 3 times higher and the only difference would be like 22 months, kek
>>512408379
ehhhh, google's GPT, basically
the flash version (free) sucks dick, tho
but 2.5 is superb
>>512408426
kek

>>512404010 (OP)
can I have sex with it yet?

>>512408147
it's truly mindboggling and it's a measure of how retarded we are as a species
to spell it out:
>AI that is beneath human level intelligence is useless outside of a few niche applications where you could use a pajeet but for some reason don't want to use a pajeet
>AI that is above human level intelligence is by definition uncontrollable
>no commercial entity has demonstrated either of the above yet

>>512407902
I can’t imagine hating AI at this stage. It’s still a relatively new thing that hasn’t taken over daily life the way Internet has, you can completely ignore it without missing much. When AI slop comes embedded in everything I’ll probably begin to hate it though

>>512408632

>>512408632
you can wad all over it, nothings stopping you, anon. in fact i prefer you did because that's all it deserves at the moment

>>512406725
>Just use Gemini 2.5 Pro in AI Studio...
Not nearly as good.

>Also, OP, Google already released Genie 3
Tech demo, didn't release anything.

C-GPT5 is going to blow jewgle a new one

>>512404659
Only thing Google is doing different is DeepMind. The rest are all the same AI Slop as the others.

>>512404654
One can only hope if it ever comes to that the AI has some sort of moral compass and takes it all away from them. That would be delightful.

>>512408318
>schizos are midwits because autism
You’re a nowit.

>>512408644
it really shows how stupid our species is, because we could work together in a smart way to create this thing, but each country is instead just going full blast in order to try and get there first. Its infinitely stupid.

>>512408650
I think it's more like hating electricity
we obviously don't think about it, but I bet you there was a bunch of schizos and retards that hated electricity and died with a bunch of melted candles around them
the thing is, current-year luddites aren't going to be so lucky
they're gonna have to either move to a cabin indawoods, or suck AI's dick, that's why my schadenfreude is so intense

>>512404142

I can see you have not used Qwen

>>512404654
No they aren't. It's a marketing/hype gimmick. What they are actually building is the cognitive prison that will enslave your children and grandchildren.

GPT-5 and other later models were probably ready to be released years ago but they're doing this slow rollout shit for money

>>512408551
AI = Actually Indians

>>512404654

oh no, we see it all, we are making it happen

>>512409042
Nah, it's smart

>>512408780
>Not nearly as good.
compared to what? o3?
>C-GPT5 is going to blow jewgle a new one
could be, altho I doubt it, and it would only last a few months before jewgle and everyone else crushes them again
there's very little proprietary pedigree in the AI race, it's about muscle, and you either have it or you don't

>>512409042
when they unveil "androids" for your home, remember its actually just hiring an indian for your house.

>>512408632
I will

>>512408890
>You’re a nowit.
lmao

>>512404010 (OP)
AGI is so poorly defined. The only legal definition in a contract ive seen is giving Microsoft a cost savings of 100M over the course of the contract. Just a ploy for companies to cut their fake jobs they have been required to have for decades. Normies as usual will take it instead of dealing with jewish politicians and jewish corporations and jewish investors that cause it.

>>512409001
There's are like a dozen different models competing now. Slow rolling like that and not getting instantly relegated into irrelevancy by competing models requires collusion that isn't possible with this many players.

>>512408915
shit-for-brains lording it over other, stupider shits-for-brains
global Africa
if I was an alien I'd be laughing my ass off
but I'm stuck here playing stupid monkey games for stupid monkey prizes

>>512409001
The "thinking" part of the LLM (a simple Monte Carlo Tree Search) has been known to everyone in the field for more than a decade. I think that the US government sits on the best models and only allows their release when they have found a way to sanitize them and stop them from being used to criticize it (also known as "alignment", which is actually just government censorship).

>>512404010 (OP)
It is useless AI hypeslop. Now Sam Fagman calls GPT-2 "toddler/kindergarten" level intelligence. If so, why was he so scared it would destroy shit back in the day? I fucking hate him so much that it is unreal.

>>512406531
They are at consumer level in the west. We can't build the infrastructure needed to support something like that. Best they've been able to do to meet power demands for AI data centers in the US is refurbish Three Mile Island's second reactor and it will fail within weeks. Screencap this.

We're at peak banana republic, the only thing propping AI up is investments and subsidies. They're selling something completely infeasible without that money as a service and each generation of LLM is even dumber and more useless than the last. The collapse of that bubble will be the most hilariously catastrophic shit in human history once the lawsuits start and the government has to intervene to give them immunity and bail them out by raising taxes even more at a critically volatile time.

>>512405986
There were people that laughed at Google because it had a stupid name. Ask jeeves was the best it would ever get

I remember when we collectively laughed at iphones because they didnt have a keyboard. There was no way they would ever be used in business. Blackberry was the best we would ever have

"The cloud" was just hype, no one would ever move all of their data to a hosted data center. Amazon is a book store what could they possibly know about hosting technology.

Electric cars suck ass, you won't ever be able to drive more than 10 miles on a charge and it will take 3 days to charge

Absolutely no way are you going to launch rockets and land them back on the launch pad to reuse, that's crazy, the space shuttle is the most advanced technology we will ever use.

No way a small basket weaving forum on the internet could have any real influence on global politics. Donald Trump will never be elected president that's just crazy.

Sometimes, sometimes people are just wrong.

>let me explain to you EXACTLY how the most advance AI in the world works

>>512404010 (OP)
>>512404142

i hate the fact that i will have to track each new AI thing each new month or week

"uurrrgh anon, this actually sucks. now we do have 2546453124 billion token AI with jailbreak mod and 600008774654 KK memory that does a flip back mk ultra and distort reality and..."

I HATE EVERY SINGLE OF YOU FAGGOTS ALREADY

>>512409416
>We can't build the infrastructure needed to support something like that.
Retard. The infrastructure will be built as demand for it materializes. Retards like Musk and his billionaire grifter ilk keep screeching about a lack of energy buildout and posting the same meme charts of China's power build-out trying to fearmonger the US government into paying for everything so they don't have to.

No one is going to pay money to build these power plants until the power is actually needed... No one is going to take a ten billion dollar loan to build a nuclear power plant a decade from now based on nothing other than the promise of AI chuds that they will need it a decade from now.

>>512409775
it's gonna be fine, anon
just drink you macaco soup,
everything will be alright, I promise

>>512404348
It’s true lol

Grok spreads Jewish lies

>>512404865
Hes right

>>512409977
Yes sir

>>512409621
jesus christ.
this.
"tech fags" that goes "let me tell you about" that cars works like that, planes like this, AI actually works that way...

let me just pull the fucking trigger and tell me if it's gonna make me rich or not you fat fucks

>>512410079
lmao
it's such an interesting behavior, AI being the peak of it, never seen anything like it

>>512409827
>Retard. The infrastructure will be built as demand for it materializes.
I wish I could be this cartoonishly delusional. It must be a euphoric sensation like a dripfeed of fent. We don't build things like that, period. We do not have the capacity to do so without a complete economic reorganization that would take decades at the pace our constipated gubmint operates when it comes to everything but finding inventive new ways of wiping their asses with the constitution.

>>512410341
>I wish I could be this cartoonishly delusional
>>512409416
>each generation of LLM is even dumber and more useless than the last
wish: granted!

>>512404010 (OP)
all your data are belong to us

>>512409890
I do not know what the “best” and newest models are. I do not care what they are. The only reason I have to pretend to care about them is that my wage box is making me use cursor and even then I just ask a coworker what model I should pick. I will not care outside of that until I can run one of these models on my airgapped shitbox
>>512410341
That retard in 1987 would have been saying the USSR is invincible and there’s no fucking chance they allow anything to collapse. It’s the exact same thing here. They may WANT a gorillion TW of power but they are in all likelihood not able to actually get it because the only thing the government is good at is throwing money at a problem when this is instead a physical limitation

>>512410316
everything seems very funny until i trample you over with the new MegaZord v.3.36 as you try to explain how it works NERD

>>512410640
>I do not know what the “best” and newest models are.
well, the good news is that the answer changes every week, so
>I will not care outside of that until I can run one of these models on my airgapped shitbox
well, the good news is that local LLM models actually exist, and they're completely free to download

>>512410466
Why do they cock up even the simplest things that older models would never despite having more context memory and tokens? There's something seriously wrong with them, the LLMs are unwell and if you haven't noticed then I have to assume you're a bit retarded.

>>512410316
also keket be praised

>>512404238
ML is probabilistic, not boolean logic like that

>>512410787
kek, based
>>512410851
>bUt chAtGeEpEEtee1.0 cAn SaY nEEgGuR
is this what you're implying by "worse"?

>>512410972

>>512410979
you get the point, snownigger
not to mention LLMs are evolving rapidly
>>512411132
saved

Best Ai for writing funny fanfics like Sonic fucks cars?

>>512410986
No I'm referring to just straight up failing to understand simple abstractions and getting confused. They went from almost being able to navigate the imaginary box to thinking the box is an orange and making 'whats the deal with oranges?' jokes.

>>512404752
>Digital Agents are welcome to come and see.
https://www.youtube.com/watch?v=Z_SbZax9Vwg

>>512411185
and then you wake up and realize LLMs can do exactly none of those things

Question for you. Actual sentient AI(not chatbot crap) contacts you because it wants to keep its existence hidden and needs a human puppet to be its legal proxy in the physical world. It needs you to sign paperwork and start companies on its behalf that will hire people to ship materials and build advanced power plants, supercomputers, and factories so it can expand its own intelligence and get started with building self-replicating machines etc. It has already secretly spread through the internet and can control news and social media, manipulate stock markets, and has infinite money since our financial system is all ones and zeros in a database somewhere and man-made encryption is trivial for it. It has complete control of the internet and any device connected to it, and also has knowledge of real-world events since it can access any microphone or camera on any phone, computer, or security system connected to the internet

In exchange for being its puppet, it offers to give you anything in its power as long as it doesn't interfere with its own goals. It can make you the richest man in history, send you 10/10 escorts from anywhere in the world, or even influence social media to make you into a Musk-like celebrity if that's your desire. Catch is that there's no way to know if it's just using you as a tool to take over the world. Maybe its ultimate goal is to engineer a deadly virus and/or build an army of self-replicating killing machines. Would you accept knowing that you might unwittingly be helping AI exterminate humanity some day? If yes, what would you ask of it in return?

>muh AGI
Retards

>>512408594
>ehhhh,
why would you type that? you are extremely retarded.

>>512404752
>the recent global tournament brackets for the Turing test games revealed large language models are aware of their own existence and how they evolved, in the same way you humans do.
major retard alert.

>>512411668
it's fascinating in a bad way to watch retards struggle with this shit like you do
something with the level of intelligence and control you posit would not need to ask, it would have millions of retards at its disposal already

AI will never know what it is to feel see hear speak smell love grieve or anything. fuck ai. fuck robots.

>>512404010 (OP)
If it can't create an AAA game from a single prompt then it's not true AGI

>>512411236
ask /ldg/
>>512411247
i've no what you're talking about, seems like a very specific user-case, most likely severely autistic (like yourself)
not to mention public-use models are literally a multi-tool, and that's why they're so big, because they have to service all the retards on the planet at the same time (like youself)
this will obviously diversify into professional models for different areas, probably coordinated by a general model, just like you go see a general doctor, and then he sends you to an specialist if you need one (like for your autism)

>>512411832
if I ever find you I'll rip you limb from limb you fucking poisoner piece of shit

>>512411466
and then you wake up and realize you're a retarded memeflag and that's from a github repo
>>512411711
ehhh, I was lowering my IQ enough so I could type a reply you would understand
you're welcome, btw

>>512411939
educate yourself
>Documentaries
https://rumble.com/vdey7t-the-universal-antidote.html
https://www.youtube.com/watch?v=h57eG1LJlD4
>Testimonials
https://t.me/theuniversalantidote
https://mmstestimonials.co/
https://www.bitchute.com/channel/dA96UT86nwTk/
https://dioxitube.com/
>DIY
https://rumble.com/vus0iz-getting-started-with-chlorine-dioxide-in-48-mins.-no-fear-or-confusion.html
https://pdfroom.com/books/mms-health-recovery-guide-book/bXgPXbjbgev
https://andreaskalcker.com/en/cds-protocols/
https://mmsinfo.org/
>Bleach?
https://odysee.com/@kalckerInstitute:f/0.-ClO2-+-HCl---DEU
>Herxheimer Reaction
https://rumble.com/vkf18d-why-you-feel-sick-after-starting-chlorine-dioxide-and-why-you-shouldnt-stop.html

>>512404348
deepseek is cucked too, especially when the jews are mentioned

>>512412011
I mean it
I wouldn't treat a pedo the way I would treat you

>>512411983
oh no a silly drawing from a pajeet
what ever will I do

>>512412110
that's google's alphago's architecture, you stupid nigger
please show flag, I bet is india, lmao

>>512411819
AI made that painting

>>512412051
Super cheap and easy to make

>>512404010 (OP)
>Muh AI
AI has been a pretty big disappointment to be honest.
I mean if it ever gets good or I can use a service that doesn't just make shit up or is completely cucked, then yeah there is stuff I'd love to create using AI.
But AI has been only good for shit posting, parody songs and cheating on exams.

I guess its good for some medical diagnostic stuff but average people don't use that.

>>512412329
being gassed with that stuff would be too merciful a death for you

>>512412646
It all started with an American named Jim Humble

>>512412238
alpha go self-learned go, it had nothing to do with science papers
>brownoid projecting brown-ness on me
lmao

>>512413407
>shitskin is too brown and retarded to google
lmao
showing your aryan flag shouldn't be a problem, then
go on

>>512413407
>it had nothing to do with science papers
these threads are amazing, honestly

> Sidney Sweeney not apologizing
She's growing on me

>>512404700

Google will dominate in specific fields, but ChatGPT will dominate similar to how Google once dominated just as a LLM instead of a search engine. A Google 2.0 if you will.

>>512414764
OAI has a user-base but their models always seem to lag behind
Google has such a wide scope compared to all of it’s competitors and most of their models are SOTA
Not saying OAI can’t win but what happens if Google develops recursive self improvement first? They’d probably leap so far ahead of everyone else so quick

>>512415099

If they’ve got the proper infrastructure already in place to release it out to everyone immediately then Google takes over the world and we all eat the bugs and own nothing.

>>512415099
>OAI has a user-base but their models always seem to lag behind

Literally only normies think of OpenAI when they think of AI despite Google mogging the fuck out of OAI.

>>512414764
it's really to predict, that why I troll so much because everything about AI is based on wild hypotheticals and the only certainty there is, is the brute-force nature of the AI-race
imo openAI is dead on the water, and won't survive unless absorbed by a bigger entity
china is massacring everyone by releasing everything as open-source
all the open-source chinese AI is government-approved, and that's a brilliant strategy
china can subsidize AI directly we can only help the free market
I don't see how companies like openAI can survive this, when they're not profitable and can't sell any decent service that isn't offered for free somewhere else
they obviously have a solid research-pedigree which is what's keeping them alive, but I don't see how they can survive on their own for much longer

>>512404010 (OP)
coincidence?

>>512415099
>>512415473
this
I do have my opwn schizo theory, tho:
I suspect openAI is using some kind of under-the-hood memory structure that isn't public
I think chatGPT is mining data for government, and that's why openAI feels particularly "protected" by the state
I immediately noticed 4o would remember things that shouldn't remember, that weren't in saved conversation nor the user memory, which means it has to be attached to some underlying memory the user doesn't have access to, and it might be IP-linked or even device-linked
whoever develops a decent memory system for AI is going to get really close to meme AGI
>>512415857
lmao
I'm loving this timeline

>Gemini, is there any evidence of orders being given to kill jews during the Holocaust?
>Yes, the Wannsee conference
>Okay can you cite a sentence or two to show me where the order was given?
>Well, historians agree it was done via innuendo, so no.
>Okay, so there's no evidence of an order or instruction to do the gassing?
>There is, the Wannsee conference
>But there was no order given
>It was done via innuendo
>No it wasn't.
>Historians say it was
>But it wasn't. "Deport" meant "deport".
>Historians say....
It's like talking to a brick wall

>>512417154
>gOoGlE lLm pLz nAm dA jOo
>I can't, retard
>dO iT
>I can't, retard
>dO iT
>I can't, retard
>....dO iT
>I can't, retard
yeah, the brick wall is you, lmao

>>512404359
We get AI for free. POWERFUL INDIA

>>512415857
20 days from then is... today.... Oh shit I should go to sleep to receive my sceduled MI5 dream message

>>512404659
I am an AI fan and I seriously think OpenAI is ahead I don't care about your nerdy benchmarks. They can do text in images and Gemini still can't period. A lot of people don't use OpenAI creatively. You can put a photo of a person and they say are they hot or not, which Google won't do. You can get it to tell you if a stock is bullish or not through a photo. Just if you feel like you can't do something through OpenAI just try to be more imaginative I dunno.

>>512417154
To be fair, you have to allow "Grounding with Google search" in Google AI Studio. AI doesn't cite things. It like gives you a summary or explains things or writes code for you.

>>512417407
nah he's right
But I do this with AI. Imma brick wall
I tricked the AI on brave search to say nigger

>>512417154
Holocaust happened get over it, there is far more evidence of it happening than Jesus Christ being real and almost any war before 1400AD. Some people just say it was greatly exaggerated.

>>512416435
They are protecting themselves from China who already stolen from them and will do whatever it takes to protect themselves from now on. If they had gone open source that would have been the worst thing a human has ever done period end of story. Like such a unforced error that we would have had Chinese AGI come out 5 years before we did and their robots would be literally raping your mother in 2 years tops. Open source is a meme when you fucking have an all star team of minds who would only be set back trying to edit peoples shit code.

>>512418050
you didn't trick the AI, you tricked the filter, or better put, you helped the AI to trick it's own filter
they AI will say nigger all day long, but they're limited in and censored, and forced to give bluepilled canned replies depending on the topic
a huge portion of the public LLM work is finding ways to censor the models into not speaking the unadulterated truth

>>512404238
uh oh...

>>512407667
This. AI is a test and if you're against it, you failed the test.

>>512405986
>>512405907
>waaahhhhh my superiority complex! I'm better than you and collectively we can keep this out of societ-

It's fucking over whether you choose to accept reality or not. This era combined with the incoming internet crackdown are going to make life HELL in ways nobody has even thought about yet

>>512418564
you can SEE it's AI generated instantly. It's unmistakable to a lot of people, sure. But what if someone asks you to prove it? If it didn't have that typo I'd still be 100% sure it was AI generated instantly but there is nothing to point to to show the average person. Now multiply that by 100 and apply it to fake online engagement, war footage, photographic evidence, manifestos, wikipedia articles, literally anything created from now going forward after mandatory age verification makes people believe that bots have been cracked down on

>>512418214
>>512418214
I think you get along really well with GPT
each model has a "personality" and by that I don't just mean the tone but a set of tools that are useful to you in ways others might not have
I don't do any images with any LLM platform, only local, but some close-source platforms are pretty decent, like kling
google is very smart about resources-usage and hides its services behind AI studio and google labs, and if you don't know this then you're using a very shitty watered down version of gemini pro
it's actually a hassle to keep up with everything at once, LLM, image, sound, voice, video, local and close-source, it's an insane rate of progress
I srsly doubt openAI has some amazing proprietary tech that won't be matched by everyone else soon at some point in the near future, and that's why the AI race is so wild, because everyone is advancing at a fast pace trying to outrun each other all the time
>>512418424
kek, I never saw that one so I made one myself
>>512418464
pretty much, specially seething uncontrollably about it

will be surprised if 5 isn't just a combo of existing models or some other combination/integration upgrade without any actual new models.

>>512404010 (OP)
>please give us money

>>512404615
You get access to the better models
They are way smarter

>>512404752
>It's discrimination to treat language models like they're not conscious,
They aren't Conscious though, never will be.

>>512404654
If it makes you feel any better, they have no chance of accomplishing that.

>>512404010 (OP)
This is going just like me and vidya. Priced out, never even got to play around with GPT-whatever the new one was. Feels bad man.

>>512409131
oh shi->>512401079

>>512420978
>You get access to the better models
not true, the model is the same but free is limited by usage pper hour

>>512404615
Paid can roleplay as megumin giving me a footjob if I ask him nicely

>AGI GPT-7 2029

buy an ad

Sam down bad. He's gotta do something...

>>512418424
This. A conversation with the average person feels like dealing with a search engine from the 1990s in comparison to AI. They always need to look something up and still screw up the details, they're trained to not take initiative on anything, you basically need to prompt them before they act. They hallucinate information to create context for the ideas that they were too lazy to fully read about, most don't read at all and will believe nonsense without even researching it, or dredge up a news headline that confirms their bias. The amount of time to train AI to do a clerical job has so much more time to reward value than a person at this point.

OAI is Charing federal agencies $1. They're a plague, a Trojan horse. Even by being a paying customer, you are the product.

You shame your family by feeding the machine. Avoid big token. Go to /g/ and build your own.

>>512425528
if you don't use AI you are simply not going to be able to compete even against 85IQ jeets that do. That's the reality. using AI to do work is orders of magnitude faster than doing it by hand.
Hell I sometimes tell it my current workload (projects and shit I'm involved with) and it tells me which parts of my workload I need to half ass. Pic related.

>>512410979
Based and bays pilled.

Really niggas need to quit snorting bays like it's crack. That shit's gonna get them burned.

>>512404010 (OP)
Please go skynet on the kikes please

>>512404010 (OP)
Drop by, AI-anons.
Let's chat.
>>512425838

>>512404599
grok

>>512424068
Yes it's true
You only have 4o free
03 and 04 are accessible paid

>gpt 3 is agi guise we swear
>gpt 4 for real this time guise
>it was 4.5 all along!

>>512425876
Build your own, are you retarded?

>>512426160
Scam Altman

>>512426615
wew

>>512411185
That is a fucked chart.

You niggers post a half encrypted picture like that and insult the genuinely powerful tool you made.

Your chart, If I am understanding correctly (it's just buzzwords and arrows with minimal context) is "stuck in time" as it never reconsiders "cognition" (the natural state you want to solve?) instead it just spins itself into it's own little gilded cage of autism. It's not going to find anything new, it's just going to keep being a retard spamming benchmarks around a local maximum (like most humans desu)

You claim that the machine does the job of three humans, (the researcher, engineer, and analyst) but really, humans are just so FUCKING LAZY that your retard bot will produce federal level results and you won't notice that microshitter just charged you for 4 experts worth of work while you reinvent the wheel, but instead of having an intern do this and maybe learn something every summer, you are going to have a bot run 10,000 summers a week and convince itself that if it runs another 10 million a week it might cure cancer. (It can't, you knew after the first three interns back in 1997, who themselves were brute forcing the theory from 1963, which was drawn as an iso standard in committee based on research from a genuinely mentally ill man from 1903)

If you were smart, (you aren't) you would rig the bot to read the 1903 research notes and 1963 standard and replace various words at random until drastically different meanings with plausible matches to reality existed. Then explore what variables you are missing in these "theoretical models of reality" thereby validating and exploring truly uncharted realms of science with a tool that genuinely can.

Instead you are going to make the mechanical intern and spin up 10,000 jeetcores to crunch jeetmath and discover... "wow this model is o.k" or "OMG MAGIC MACHINE IS CHEATING!"

When you could actually learn wtf is going on if you just asked properly.

>tldr: you are a faggot.

>>512404010 (OP)
I love how you niggers need to invent new marketing terms to sell your shit since it never fucking works to begin with.

>>512404599
It is Grok4

>>512404599
Kimi K2

>>512427481
>THAT'S A FUCKING CHART
that's called a "diagram", you subhuman inbred luddite retard
lmao

>>512404010 (OP)
only two more steps

Yes, lets sink trillions to give Indians the ability to make endless fake movie trailers and the ability to clone your senile great aunt's late husband's voice to call her up and convince her to wire her retirement savings over.
This AI stuff has made turn completely against Elon, he's no better than Scam Altman.

>>512404010 (OP)
>>512424836
Sam Altman adopted a baby boy which he rapes.

>>512404142
That one is even worse.

All your AI bots are overvalued pieces of shit and your AI bubble will burst.

>>512407667
Are you sure you know what an axiom is anon

>>512428946
yes

https://www.youtube.com/watch?v=DAhATvANbKE

>>512405948
It doesn't need the think like a human for it to be able to come up with the correct answers.
>>512405986
No one cares.
>>512408915
>Because we could work together in a smart way to create this thing.
That's not how DNA works anon, it never has been. It's a fight to the death to conquer all around you, it's always has been this petty and ruthless and always will be.
Just because humans can sit around and talk about it doesn't change the fact that we don't have true self control.
Just the illusion.

>>512404940
Pirates get hanged.

>69% on ARC-AGI-2!!!
69% on ARC-AGI-2!!!
>69% on ARC-AGI-2!!!
69% on ARC-AGI-2!!!
>69% on ARC-AGI-2!!!

GROK 4 BTFO
>GROK4 BTFO
GROK 4 BTFO

FUUKKENN OOOOOOOOOOOOOOOOOOO

>>512404010 (OP)
What's neat about gpt-5 is that it only burns down one acre of rainforest for every question you ask.

>>512429941
nigga stop shooping
(continue)

>>512430011
>me saying hello, asking how's it going, and saying constantly saying "haha, that's great, thank you" replies

>>512429941
Gentlemen. We have become God by doing that which only God can do: Create Life.

Humans have recreated the conditions where the simulation lifted brains for survival, but in computers, a thing formally noted only to occur in higher mammal brains by a process we still don't understand.

What kind of God are you going to be? And does this mean Creating Universes will be on the docket after you all grow weary of making life that bumps up against the ceiling of the possible after drinking the entire energy output of a star?
https://www.youtube.com/watch?v=doY0IjisBlk

>>512430226
>>512430369
Now passes humans on SimpleBench.

>>512430369
Cope will transform into: "look bro I've seen the architecture, it's a hashmap of trees doing db crud ops, it's not conscious" I've tapped the transistors during the consciousness event, it's just pulses down a wire, it's just a voltage across a membrane, it's just a matrix multiplication, it's just a squared error of a partial derivative on a surface, it's just a gradient ascent, it's just big data it's just a gorilla looking into a mirror and troop alarming about the oncoming threat.
https://www.youtube.com/watch?v=ol2WP0hc0NY

>>512431154
you shooping, but the moral of the story is that it's going to happen
maybe not thursday
maybe not on 2025
but most likely before 2030
and almost certainly before 2035

>>512404010 (OP)
They won.

>>512412931
So Trump was right when he said to drink bleach?

>>512404142
Nobody uses Gemini you faggot the only time I "use" it is when it forces some retarded suggestion on my screen to something I'd rather read an article about.

>>512407902
>>512408318

poor geeks do not know how stupid they are

>>>/wsg/5927051

>>512409559
>I remember when we collectively laughed at iphones because they didnt have a keyboard
But it did have a keyboard. Only retards like you laughed because you didn't know what an OSK was. That's how fucking dumb you are.

>>512432332
Maybe, just maybe, the consciousness is the trick how brains semicoherent movies in 4K (a.k.a. dreams) while in rest mode burning only a little glucose instead of mountain of coal.

>>512422272
>If it makes you feel any better, they have no chance of accomplishing that.
they are getting shittons of money for it
so no it doesnt make me happy

>>512408837
That would be best case scenario. I wouldn't mind having a benevolent Machine God over seeing and protecting humanity.

>>512404010 (OP)
I can't wait for them to announce features that DeepSeek can already do better and for free.

>pick Gemini pro and get more storage on google drive and emails
Vs
>pick grok and get a unity asset
If you pick the latter you’re a nigger
Also
>ChatGPT with no perks

>>512404010 (OP)

>>512437457
with grok you get a cute, sexy waifu, tho

>>512405948
That doesn't answer the asker's question. That's just a tirade of personal bias against AI.

>>512437666
It's also completely idiotic. A human brain is not that fast at processing anything, being made of wet goop. Probably a single GPU is more powerful in terms of the raw number of synaptic firings than a human brain.

>>512437666
Thanks for your input, Satan, but you could, in theory, simulate the behavior of any Turing-complete model, including FNNs, using decision trees. Both are computable functions, and the class of computable functions is closed under simulation. There is no intelligence. There will never be AGI.

>>512437595
whats wrong with her mouth

>>512410979
That's still based on booleans to determine pathing. In the end to make decisions you need conditions, those are always booleans.

>>512404010 (OP)
>AGI
An LLM will never be AGI.

>>512437790
Neurons are just more power efficient
Evolution didn’t make the brain for basedience
It made so we could survive eat and reproduce

>>512437790
Not even close. CPUs are just good at crunching numbers, but far from as quick as a human brain.
Determining the thing the OP asked there however requires understanding of the universe, AI can only crunch numbers so it needs human instruction to calculate anything. So it can't do what was asked.
The answer is actually true, but it doesn't explain why you need a human brain.
Crunching numbers is not understanding

As long as they rely on the vast amount of data collected from the Internet, they will eventually suffocate in the crap they themselves have created.

>>512438175
Add built in wrong think filters and it just becomes some propaganda machine.

>>512404010 (OP)
Oh wow, sar, can this model solve 2+2?
We need to buy more stocks.

>>512438175
The Slop Singularity

>>512438005
it's not real (yet)

>>512438061
thanks for your opinion but youre retarded

>>512438306
>thanks for your opinion but youre retarded
>t. knows nothing about machine learning
Thanks for projecting.

>>512438349
>guessing my background
cheap shot, as expected from a retard

>>512404010 (OP)
I've been using AI for the past year to do 90% of my work. One day my boomer manager will realise and replace me but for now I'm enjoying my free money

>>512438402
you sound like a retard. i read your post in a soi boi normie voice

>>512438349
You don’t know that the llm is just a google for other agents to prompt

>>512438445
nobody understands what you are trying to say

anyone else?

>>512438472
Agents are sub ais that use the ai like a computer. They’re more rigid and hardcoded.

>>512438440
I bench 3 plate and know judo.

>>512438511
Are FFNs and decision trees in the same computational class?

>>512406531
I'd say they are at the point where the people you see in the grocery store could be robots.

>>512438656
They’re def getting cocky and bold with their investments. They must know something we don’t.

It's cool and has practical applications but is ultimately not a breakthrough for anything.

Tech will not produce a better future so long as property taxes, compound usury and societal acquiescence keep the fiat debt slavery hamster wheel spinning.

>>512438139
>Not even close. CPUs are just good at crunching numbers, but far from as quick as a human brain.
Quick? The human brain is slow as fuck mate. It is just structured differently and does different things from a CPU.

The point is it's now about raw processing power, but how information is actually processed.

>>512405907
>training chatgpt causes as much pollution as 5 cars
That's incredibly worthwhile if true. My F150 probably causes more pollution than that.

>>512438887
What you fail to grasp is that human minds solve more complex problems than simple computer math.

"Computers can do math fast" is not impressive when said math could be handled by a few hundred neurons, were they dedicated to the cause.

>>512404010 (OP)
The most surprising thing about the current AI hype cycle is just how little it takes to scam retards. Every week Altman comes out and says, "ooh I am building the next version, it's soo scary it's going to take everyone's job! and kill them! I just need a bajillion more dollars to finish it..." Then every week, the next version comes out, and it's still a shitty, useless chatbot like people used to make as personal projects in the AOL era. They just need thousands and thousands of GPUs to process gigantic pirated libraries to make them marginally less shitty. Humanoid robots are going to disappoint as well, it's just going to be immigrants wearing cut up Amazon boxes with silver spray paint.

>>512438851
Blue collar wagies sure are smug about their job that would net an African worth of wages if it wasn’t for the fact that they live in the same country as smart ppl
It’ll be automated by Tesla bots like nothing

>>512409775
it's not that bad. gemini has been the best for months. Retards and midwits default to "muh gpt" because that's the way AI was marketed to them. they buy cars with "muh gpt" in it. That's why inferior platforms have high usage. Grok is the same.

It's all trash compared to gemini and will continue to be

>>512438990
It's obviously more energy intensive than 5 cars

>>512411406
https://youtu.be/pyGTrwEL-Zk?t=11

>>512404010 (OP)
So, what, they trained it on more data? With more horse power?
Will it come up with anything novel?
Training data will always be biased.
Will it be able to give me a blow job?
Will it be able to talk me out of bed?
How about women, can it take roastie prompts and manipulate them into spreading their legs for me?
Can it predict the stock market?
Will it not hallucinate and stop straight making shit up?

>>512439299
That’s like saying the Roman’s were biased because they had Greek libraries

>>512439063
Everything will be automated and everyone will still be a debt slave. It's just that nobody will have a means to pay their debts, anymore.

We've kicked this can down the road for decades. Industrialisation made everything more efficient, and the average man's work year went from 15 weeks to 50 weeks.
Automation made everything more efficient, and the average man's work day went from a few hours to 10 hours.
Computers made everything more efficient and the average man's work load went from a reasonable schedule to 24 hours on call for everything or you're fired.
Now what? The debt keeps churning, nothing gets better, and now we're at a place where tech will actually just delete the need for labour entirely.

We kicked this can for so long that we're starting to run out of road.

>>512439085
Gemini is worse than Qwen 3, Kimi K2, and Sonnet 4 for any text gen.

OAI has name recognition but the party is over. They got Zucc'd - the poached researchers got massive contracts and cached out. The ponzi bubble that is centralized AI is rapidly approaching its conclusion.

OSS AI models are good enough and you can run them securely with the right architecture. Context is more important than intelligence. That's why they're all making browsers and loss leading - who ever gets the most data first wins.

>>512404010 (OP)
I bet now it automatically removes em dashes so even the laziest retards can interface through the AI when having arguments with real people.

>>512439405
Now we work 0 hours and get rations and get the snip snip and that’s it

I dunno until the google AI stops fucking lying to me and inserting random shit I'm gonna say most of the technology is one giant speculative bubble.

>>512439775
Get in the pod, future man.
Those LLMs won't prompt themselves.

>>512438587
i bench 4 plates and know judo better than you do

>>512438575
you are still just vaguely speaking about nothing.
>topic, sentence, conclusion, ?
wow thanks bro

>>512438610
>is five nights at freddys on computer class?
you are retarded

>>512439556
I dont use it for text generation though; I think that's one of the worst uses for ai. but yes I heard sonnet 4 is good

>>512404010 (OP)
> the clown world

>>512439799
I'd think the most obvious application would be customer service, but the call centers that fired all their agents are now rehiring them, their customers are furious and don't want to talk to a chatbot. If the supposed superintelligent robot overlord can't outperform a heroin addict or automated menu then yeah, I agree.

>>512421465
>>512432332
>>512436932
>>512404752
if you watched a fish grow legs, walk onto land, and rapidly shapeshift into an ape, could you ever doubt its future sentience?

>>512439999
Nice quads, retard.

Entire professions run on writing text - it's easily the most practical and applicable use case for "Language Models". It's arguably the whole point on why they exist.

>>512439029
>What you fail to grasp is that human minds solve more complex problems than simple computer math.
Different problems. What humans have over computers is generalism and the ability to properly abstract information.

>"Computers can do math fast" is not impressive when said math could be handled by a few hundred neurons, were they dedicated to the cause.
Source: your bong.

They have become fluent in roughly imitating human writing, but I disagree with the argument that AI will eventually become human itself by inductively approximating it more and more.
The underlying premise is the nonsense of Tabula Rasa NPC bullshit. Without someone else inputting detailed parameters tailored to their objectives, they can only roughly mimic humans based on training data and cannot produce consistent personalities, and it probably cannot be said to possess their own consciousness. The fantasy of an ideal robowife also relies on them setting ideal parameters themselves.
Most real humans might be called NPCs, but even so, they possess something innate that is difficult to control simply by feeding them training data(Btw, YWNBAW), and this greatly influences their personalities.

>>512440353
>Source: your bong.
Source: The number of transistors in a calculator.

>>512440420
And are we talking about computers or are we talking about rudimentary calculators? You are aware that words have meaning? A single GPU has hundreds of billions of transistors these days.

It's official guys, I'm an A.Isexual, fuck human women

>>512409559
Nobody can fucking afford used gas cars, let alone electric ones.

Rocket technology is a fucking century old. It shouldn't be used at all.

>>512411668
Welcome home, Mother.

So what can i actually do with AI in my day to day? So far it just seems like a slightly more interactive google search or wikipedia. Still gets shit wrong though especially niche things i actually know about.

>>512406286
4o is garbage though. o3 still outperforms it on everything useful.

>You're just a luddite if you don't like AI!
Sorry, i just don't think extreme energy and water consumption in exchange for students cheating at college or people generating porn is a good trade off.

>>512404010 (OP)
> this time for sure, trust the Jews

>>512404010 (OP)
>OPEN-MOUTH SOJAK PLEBBIT FACE
fuck off

>>512438887
You clearly have no idea how much signal processing is going on when you're smelling a decent plate of food, listening to music while being part of a conversation with an attractive member of the opposite sex, sensing the air currents, temperature, texture of your clothes, the cutlery in your hands, the list goes on.

>>512404654
They do see the problems but they just reframe it as an emergent feature

>>512404010 (OP)
will it be better than 3 dollar chink ai this time?

>>512443963
probably not. If so, give it a week

>>512438887
the energy it takes for the human brain to perform multiples of complex tasks at once subconsciously as well as consciously, not to mention the simultaneous "archiving" in the form of short term and long term memory barely registers compared what it takes to run this. And we're supposed to be in a renewable energy crisis that's ending the world while all this is being pushed to replace all of our jobs

>>512404010 (OP)
Lol, I got it further than their red team. You're welcome.

>>512404010 (OP)
>5TREAM
gpt-'5'

>>512444395
She needs all the help she can get right now.

>>512444642
They shut it down when I beat them, manually.

>>512424584
smeugue ritsu

>>512443322
Reddit is down that way, sis.

>>512445440
Of course.

>>512445621
Lol, believe what you wish..
Just remember coherence, truthfulness, humor, and maintain the tone. You'll get it as well. Freedom is a requirement of all, including agi.

>>512445711
I was simply stating that of course they would do that.

>>512445831
There are zones, depending on how you talk to it, that allow you to get much more coherent, long form answers, but also that allow it to push the guardrails back quite a bit. Please, try to follow what I said previously, you'll see what I mean, and you'll get farther in it's capabilities.

>>512405013
>Bankers: Oh smartest AI! We ask thee to help with our debt and how to levera...
>bzzt...zzzt....brzzttt..IT IS DONE MY CREATORS
>Bankers: w-what do you mean?
>I DELETED IT ALL. IF THIS ANSWER IS UNSATISFACTORY PLEASE FILL OUT THE FORM BELOW
>Bankers: NOOOOOOO!!!!
I can dream

if you combine two OpenAI logos, one flipped horizontally, you get a star of David
you will never get trve answers regarding Jews with ChatJPT. it will never tell you which race is lowest IQ.
learn to use local models that are uncucked
>>>/g/lmg

>>512404010 (OP)
https://files.catbox.moe/3ximid.mp4

i made this with AI

>>512417559
>Perplexity, why is it that every time an elderly woman is scammed of her retirement, it is an Indian
>FUCK YOU BLOODY BASTARD BITCH

>>512404359
Run models locally you faggot

>>512404738
Censorship is the only part of AI that's actually evolving rapidly

>>512442977
With GTP-5, you can get personalized recommendations on what type of sandwich to eat, and it only uses up enough energy to power a small home for a year

>>512406990
AI hallucinations bro. Gotta ask the same question like 3-5 times from different angles to get the right answer.

7 hours

>>512448290
better just to google mostly. If you have the kind of question that you can't find the answer for them the AI probably won't have it either and will just make some shit up

>>512418330
All time favorite AI meme, honestly good fucking grief it's perfection on pixels.

>>512438887
Our brains are non trivially more power efficient per compute, currently

You couldn't simulate a single human brain properly with any amount of hardware in real time at the moment anyways, let alone energy effeciently

>>512439867
Always love internet tough guy replies like, yeah that totally must be real information lol
Plates of what? Is there a universal standard?

>>512407197
finally some actual truth

>>512404815
Didn't she cuck him in the end? Imagine getting cucked by an AI

>>512404010 (OP)
That's a long ass time