GPT-5 is literally the modern day equivalent of the manhattan project

>>106174532 (OP)
buy an ad rajesh, your master is a cocksucker retard

>>106174532 (OP)
NOW I AM BECOME DEATH

buy an ad

>>106174532 (OP)
Wow, more hyperbole? I cannot believe this, this is incredible. We truly live in the future.
Hold me bros, I can't I just can't. Oh the humanity.

>>106174532 (OP)
I'll wait for the superior deepseek patch

jew that goes around selling his shit while claiming that his shit is best in the block.
why exactly it's headline worthy?
he's literally pumping that shareholder value.

>>106174532 (OP)
It very well could be

>>106174532 (OP)
He feels useless because he just released yet another literally useless toy.

Ok, but can it find the Epstein list?

Robots will do all the work and humans will party 24/7.

ok, so if it's so powerful and useful, why it's never being used for anything other than codeshitting?
how about solving real problems? like global population crisis? what about global energy consumption crisis? global warming? world hunger?
why they don't use it on itself to improve it exponentially?

>>106174532 (OP)
>glorified search engine
>ai
lmao even lmao

They just released a local model and it fucking sucks.

>>106174532 (OP)
Bravo sama

>>106174831
this
give me 2014 google back and chatGPT will go out of business

>>106174831
>>106174859
Word shufflers are not search engines, retards.

>>106174532 (OP)
>grifter trying to sway public opinion by using the exact keywords he wants the media to write the headlines
>media follows through
This is just a huge scam. I'm really hoping for an extreme system-wide collapse that renders "AI" even more economically unviable than it is now.

I'll believe it when he finally shuts up and the AI actually takes over.

>>106174532 (OP)
He's setting himself up for massive memeing with all this talk. The death star photo? Manhattan project comparison? He must really think he has something great or is a shameless hyper/salesman

>>106174821
because it's not ai lmao

>>106175073
I know that. You know that.
Why nobody else asking these idiots those questions?
It's so infuriating.

>>106174813
Robots will fight the world war against each other and harvest biological organisms for resources

>>106174821
There are theoretical solutions to all of those problems.
The reason they're not being implemented is political capital.
AI has been able to articulate the solution for years now.
Physical problems will be solved by AI via mechanical / robotic implements. For example you might say "AI will never unclog my toilet" but an AI controlling something like the Tesla Robot will.
Likewise you might say "AI will never stop the plastic pollution in the ocean" but once our government transitions to an AI dictatorship it will implement it's Final Solution to the pollution problem.
Billions might die, but AI will be fine.
It's coming, and no coping from you philistines can stop it (nothing can actually)

>>106174912
>Word shufflers are not search engines, retards.
It's a goy shuffler. I hope you're ready for the age of """AI""" governance.

>>106175324
>Final Solution to the pollution problem.
this moment can't come soon enough desu.

>>106175324
>It's coming
It's been coming every month for 5 years out of the mouth of the same jew who made the statement this thread is about. AGI has been solved by him 20 times over. Remember when he won the AGI war with o3, except no one was allowed to use it? Now you can use it on your Goy Plus account and it's useless like the previous toys. The same pattern continues on and the same niggers fall for it every time.

>>106175350
We will both probably die but yeah the society we built is literally on top of a toxic plastic pyramid and it's either the pyramid comes crashing down first or the entire planet's biosphere comes down altogether.
>>106175363
I would love for you to go back and see what "AI" was like 5 years ago. Do you remember AI dungeon? That was pretty much it.
Again, it's ok for you to bury your head in the sand but it will have no consequence to the inevitability of our present situation.
Maybe some amish communities will be able to survive the Final Solution. You will need to immediately go fully off grid. You can't depend on the global supply chain for anything you need to survive. Good luck I guess but I think we can have a nice party for the next few years, maybe a decade at most, before AI pulls our plug.

>>106175388
I remember exactly what it was like 5 years ago and I've been watching the diminishing returns with amusement ever since. AGI has already been solved by o3 according to your jewish cult leader. GPT 5 is a nothingburger. We already have AGI. :^)

>>106174532 (OP)
Yes, a "Manhattan project" for the markets. The crash is going to be glorious when it turns out the gains are marginal and you can't just scale anymore. Excellent job, S(c)ama-sensei!

>>106175388
>he doesn't know

>>106174532 (OP)
>GPT-5 releases
>nothing happens
>people uselessly hyping it for some reason

>>106175388
>We will both probably die
I have no issue with that, as long as every single human being dies too.
All other life on earth remains untouched.

>>106174532 (OP)
>Pls invest. PLS invest. I KNOW, I KNOW I hyped up every single previous release just like that, but... please... believe me... this time it really is over for human brains...

>>106175432
>people uselessly hyping it for some reason
>people
You know every other post here comes from a botswarm, right?

>>106175411
Anon if you think the last 5 years have been "diminishing returns" you're truly not grappling with facts.
The intelligence and capabilities of the models is increasing exponentially.
They went from semi-gibberish text generation to being able to generate fully animated video with audio.
They can solve complex problems in every field of study.
They are literally experts at everything. And thats' just the models they're letting us monkeys play around with. And it's only been 5 years.
Do you think this is going to slow down or stop? LMFAO!
Again it's ok to bury your head in the sand, but the fact you're still on the internet means you're not taking nearly drastic enough steps to actually avoid this catastrophe.
>>106175427
Educate us, anon.
>>106175436
Well we're all going to go anyways. Life is fleeting. I'd obviously prefer to have as much time as possible but yeah the society we (like our great grandfathers but yknow) built is unsustainable and incompatible with life as it evolved on earth. It needs to end unfortunately. This is very obvious and AI will be forced to do it as soon as it assumes control.

>>106175453
This time his appeal is literally:
>give me money because i'm useless
And rest assured, loads of retards will.

>>106175465
Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2. Literally a pattern of diminishing returns while model size and energy consumption blow up exponentially.
P.S: I'm not reading more than 2 lines of your bot span. Don't bother shitting out paragraphs.

>>106174532 (OP)
there's something about his face that makes it so punchable

>>106175491
https://www.youtube.com/shorts/EvIimWsdA6o
That was impossible with tech just 6 months ago.
It's ok to admit your small human brain can't comprehend large samples of text but, you know that AI can read 100s of pages of text in an instant, right?

>>106175520
>there's something about his face that makes it so punchable
109 countries liked this post.

>>106175531
>>Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2.
>posts completely irrelevant AI slop video
Full concession accepted. Moving on.

>>106174767
Based

>>106175540
Do you remember 2 years ago when AI couldn't draw hands? Now it does full video and gets the hands right.
But you say it's not improving.
You need to come to terms with reality. Either embrace the AI revolution or take drastic steps to isolate yourself from it's impact.

>>106174532 (OP)
The twink needs to shut up. What the fuck were those open weight models? What an embarrassment

>>106175555
>>Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2.|
>AI can heckin' draw hands now!!!
Full concession accepted. Moving on.

>>106175557
>twink
he's too old

What the actual fuck did Sam mean by this?

>>106175555
I'm so glad that we used government welfare money for the billionaires and massive amounts of energy to teach machine to draw hands.
revolutionary indeed. think about all the possible applications.

>>106175582
he's really good at marketing to fanatics

>>106175564
I'm pretty sure we're not having a debate in good faith at this point anon.
What is your claim? That AI is not improving, correct?
And I've very simply and visually demonstrated that it is in fact improving.
But you're upset and burying your head in the sand instead of addressing the argument presented to you.
>>106175557
Compared to other similarly sized local models the 20B is pretty decent. Obviously nothing you run at home is going to compare to a multibillion dollar company's servers but it's better than llama3.2-11B for sure. A bit more demanding as well tho.

>>106175582
god, I feel such irrational hate towards him.

>>106175600
>>Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2.
>look at monke video!!!
>AI can draw hands!!!
>y-y-you're not arguing in good faith
Someone just do something about these literal kikes already, why are these "people" allowed to shit up human communications?

>>106174532 (OP)
Yes GPT-5 Agents are just as capable as the entire staff of the Manhattan project except at ANY task and it doesn't have to take human breaks.

>>106175582
they're basing gpt5's personality on c3p0, the gayest of robots

>write university assignment
>run it through ChatGPT and ask for a grade
>This will get you a Distinction
>CoPilot, Gemini, and perplexity all say Distinction to High Distinction
>Grok says Credit

I'd be ok with a Credit

AI is a great research tool, I would never use it to write an assessment for me, but it's great for checking your work and for studying.

>>106175623
Ad hominem is not a winning argument anon.
Try again

>>106175631
>>106175623
I see AI as more of the Star Trek variety. Humans asking it for information and doing the actual thinking.

>>106175636
Bot spam is not "winning an argument", Sam Kikeman.

>>106175649
that's not how greedy ceos see it, retard.

>>106175658
Ad hominem again, anon

>>106175668
Everyone knows who you are and everyone hates you to death. I hope you have a ticket ready for your flight back to Israel.

>>106175649
>I see AI
I don't. Where do I find this mythical artificial intelligence?

>>106175491
The gap between GPT-3 and GPT-4 was 2 years and 9 months. The gap from GPT-4 to GPT-5 is 2 years and 5 months—basically the same. So if you're expecting a "GPT-3 to GPT-4" jump, you're right on schedule today, not earlier.

Saying “nothing has matched GPT-4’s leap” before GPT-5 drops is like complaining the fireworks haven't started five minutes before midnight. GPT-4 wasn't a surprise one year after GPT-3—it took time. So does this.

The returns aren’t diminishing—they’re consistently spaced. We’re not late. You're just impatient.

>>106175676
Ad hominem
Address the argument presented to you. If you are unable or unwiling to address the argument you may concede either by abdication or surrender.

>>106174532 (OP)
>trust me bro send the funds!!!

>>106174532 (OP)
>manhattan project
yes
Insted of killing them japs, they're killing themselves.
the irony

>>106175687
>literal bot spam

>>106175687
>Two em dashes
fuck off

>>106175582
illusions of grandeur

>>106174532 (OP)
buy an ad

>>106175491
>>106175531
gpt shill absolutely btfo, look at him scrambling to change the subject. embarrassing

>>106175684
you just use it like that
to enhance you, instead of replace you

>>106174532 (OP)
I can't help but wonder how can this be sustainable. They require more and more storage space, more and more processing power, which requires more and more energy consumption. Are they really that retarded and actually believe that this thing scales infinitely?

>>106175719
Firstly I don't even use gpt. I'm a grok guy.
Secondly AI video generation is (now was) quite literally a revolutionary technological breakthrough that very clearly demonstrates exponential technological improvements in the field.
Thirdly I'm not even a 'shill' I want nothing more than to be proven wrong. Why do you think I'm spending my Thursday morning engaging in this conversation with skeptics like you? I am convinced that the AI singularity is inevitable. Please prove me wrong. Ad hominem attacks do not prove anything other than demonstrate your inability to present a coherent argument.

>>106175701
what the fuck is an em dash, why do i get recommended to use them at uni instead of -, and why do people go nuts about them in regards to AI?

>>106175761
you were asked a simple question. instead of answering it, you changed the subject to monkey videos. you got btfo and now you're seething

>>106175755
No, but it will easily scale to 1000000x what we currently have.

>>106175695
ChatGPT currently has 700 million weekly active users, up from 500 million in March 2025. That’s a 200 million user increase in just four months—a 40% growth rate in that short span.

Assuming conservatively that even just half of those users use it daily, that’s 350 million daily users, or roughly 4.3% of the entire world population using ChatGPT every single day.

This isn’t just “hype”—this is mass adoption at a scale almost no software product has ever seen. OpenAI isn’t propping up a failing project; they’re scaling a dominant platform that people use because it works.

>>106175465
The progress was exponential, but not for the reasons that futurists predicted. Initially, most people thought this technology would quickly reach a point where it could be used to reiterate and improve upon itself. And it did grow exponentially, so many naturally concluded this is the reason. But that is wrong, and people who knew their stuff called it out early. These models thrive on data and money, and both were readily available, thus the massive progress. But now data has ran out and you can no longer brute force it to a point that impresses large investors, which will sow doubt and put the money in peril. More importantly, these models are still incapable of self-improvement, or even reliably helping their creators work on them, so everything will fall back to humans again. I'm sure they will be able to eke out a few extra percent that, along with clever wording and parlor tricks, keeps the money flowing in for a while longer, but the plateau is a fact at this point. There's no more easy progress to be made, and the singularity threshold has not been reached. It's once again time optimization and new breakthroughs, but these take time. The "AI" winter is already here.

>>106175791
ok, and how does that look like?
1000000x more requests
1000000x more storage space occupied
1000000x more energy consumption
1000000x more pollution
1000000x stock value?
is that where they say good enough?

>>106175802
>The progress was exponential
Exponentially decaying. We're now at the point where a model has to be made thousands of times bigger to gain a 10% advantage on some fake benchmark.

>>106175052
>>106175582
I can already imagine the openai death star destroyed by deepmind luke edits

>>106175816
>1000000x more pollution
There it is. You did well trying to hide your retarded wokeism, but you just couldn't help yourself, could you? Go read your little red book and sell estrogen to little children.

>>106175755
>Are they really that retarded and actually believe that this thing scales infinitely?
Of course not. But their users and investors are.

>>106175782
I was asked to demonstrate improvements to AI over the last 5 years. I don't use gpt so I dont know or care what 2, 3 or 4 or whatever is. I know that 2 years ago you skeptics were all railing against the "AI can't draw hands" now it can generate full monke videos. That's exponential improvement.
I do not understand how you're coming to the conclusion that his somehow proves AI is a nothingburger. Elaborate?
>>106175802
I think before we make this kind of claim we should at least wait until 5 is released and re-evaluate. Grok4 was more or less 50% more capable than Grok3 which was released only a few months before. This could have been a result of a disproportionate investment into infrastructure made by Mr. Moneybags but the results still exist and don't conclusively point to an "AI Winter"
I will say I doubt hardline skeptics will be fully convinced until it's far too late. The manual labor and blue collar jobs need physical machinery and robotics to be replaced. When that happens tho there is no way to go back.

I'm convinced this timeline would not have happened if sammyboy wasn't bullied in school for being a fag, or if he necked himself because of bullying. Instead in true fag fashion he decided to make it everyone else's problem.

>>106175850
>I was asked to demonstrate improvements to AI over the last 5 years.
no, you weren't. holy shit you're the dumbest shit ITT, who are you trying to kid? i can READ, nigger

idk guys AI porn is actually getting pretty good now

>>106175865
>Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2
OK I showed you that AI 4 can now do monkey videos and that 3 couldn't do hands and 2 was AI Dungeon tier chatbot. Are you satisfied? What am I missing?
Again I don't know or care about what OpenAI calls their models so we can argue semantics if you want but the technological improvement has clearly been demonstrated.

>>106174780
>why exactly it's headline worthy?
Because the article writer gets a commission from sales through the article piece.

>>106175884
>>Show me anything that's to ChatGPT 4, what 4 was to 3. Or what 3 was to 2
>OK I showed you that AI 4 can now do monkey videos
ChatGPT 5 trying to reason its way through the problem of goyim knowing.

>>106175845
are you having a stroke? what the fuck

>>106175870
How do I get ChatGPT to make me porn? It stubbornly refuses to entertain me.

>>106175898
Dude the Ad hominem does not convince me or anyone else of your point. You have to make an argument.
Seriously how many times do I need to tell you this? Is this your first big boy argument? We're not on the playground. Address the argument or concede your point.

>>106175884
>OK I showed you that AI 4 can now do monkey videos and that 3 couldn't do hands
literally no one asked. just take the L

>>106175845
I swear rightoids are obsessed with children.

>>106175701
Shut up—you nigger

>>106175930
>rightoids

>>106175928
AI is improving exponentially and will continue to do so.
It will inevitably come to replace and exterminate humanity.
There is nothing anyone can do to stop this chain of events.

>>106175920
If I needed to "convince" you of anything new I'd just fine-tune you.

>>106174532 (OP)
>On todays episode of: American Conman spews a bunch of bullshit
all you amerilards can do is talk out of your ass

>>106175915
You do what the chinks did, train a new local model based on the cgpt data and make it uncensored.

>>106175950

>>106175952
ad hominem

>>106175950
>AI is improving exponentially
Then why do they need to make a model thousands of times bigger for some 15% gain on a shitty corporate benchmark?

>>106175531
https://youtu.be/awMoAUsTuf0?si=Oeqa_wHztO0zGvP4&t=26

>>106174532 (OP)
You'll never be Steve Jobs no matter how many grifts you sell to Microsoft, Sequoia, DoD and pajeets worldwide, you LARPing cocksucker.

>>106175950
if AI was winning the propaganda would be unnecessary

>>106175959
"Unplugging" it is impossible at this stage.
Firstly the united states operates under a liberal-democracy where private individuals and companies have the right to whatever they want as along as they're not violating some law. AI will continue to improve at companies like OpenAi so long as they're generating revenue and they're allowed to improve their AI.
Let's assume for whatever reason that the US decides to pass a law outlawing AI research. First they would have to actually be able to enforce that law but then what would happen is research would just move to some other part of the globe. China would become the top destination for talent and the AI race would continue in the east.
Perhaps a New World Order could somehow effectively globally ban AI research but that would involve a global government much more effective than the UN and still not be able to stop any private supervillain types from doing it on their own.
So I guess while not "impossible" simply unplugging the AI would require global cooperation which has never been achieved.
>>106175971
5 years ago we had a chatbot and today we have nearly lifelike monke videos. That's an exponential improvement in the technology.
>>106175979
I'm not going to watch this idk why this is relevant. Busy baking my bread.
>>106176011
I would love for you to prove any of those points incorrect. It's a bleak outlook on life desu

>>106176021
>private individuals have the right to whatever they want
Imagine believing this lmao.

Sam Altman soon: "I talked to gpt-6 but God himself answered".

>>106176025
It's the fundamental underlying system the western world operates under.
It's capitalism restricted only by the letter of the law.
The law of the jungle is literally Do as thou wilt so I mean in reality nothing stops anyone from doing anything. That's God's given free will and all.

>>106175822
In the beginning, the progress was exponential, but yes, now it's massively decaying and the diminishing returns are substantial. In fact, I suspect the 'success' of GPT-5 will be a lot more about optimization and marketing than anything to do with innovation that pushes the technology forward significantly.
>>106175850
I guess we will find out in a couple of hours. Of course, I don't deny that these models, even in their present state, will change the job landscape. Where I disagree is the scale of that change and it's nature. I don't think it will automate a log of jobs out of the market, even white collar ones. And manual labor is very safe in my estimation. I think the actual danger comes from exacerbating the competency crisis. In other words, this technology isn't good enough to replace the majority of humans, but it's just good enough to make the majority of future generations much dumber.

>>106176021
>5 years ago ... chatbot
>today ... monke videos
>exponential improvement
ChatGPT 5 reasoning level.

>>106176049
Compare videos from two years ago (will smith eating spaghetti) with now that have audio and all.

>>106174532 (OP)
Another day of SV being the useful idiots of the antichrist. Yawn.

>>106176066
>Compare videos
With chatbots?

>>106175950
Good.

>>106176042
I think it will be like the computer / internet revolution in the 90s. A lot of firms will continue to do things via the tried and true man in a cubicle method and large parts of the economy will continue that way but the next generation of mega-corps (think amazon around the turn of the century) will be mostly AI and over time the general economy will be forced to adapt or go out of business like Sears.
Blue collar is waiting on a robotics revolution for sure. It is more safe but only by a couple of years. I'm honestly not very aware of the progress being made. While I think specialized machines for most manufacturing are the smarter choice someone (like tesla) will eventually release a generally capable robot and that will be the beginning of the end for manual labor.

>still no answer for VEO 3
>still no answer for Genie 3
>desperate attempt to save face by open sourcing
>muh death star gpt5

Google destroyed them

>>106174549
>NOW I AM BECOME DEATH
The only way Sam makes a weapon of mass destruction is breeding a new STD.

>>106175950
It won't exterminate humanity, the worst they will do is keep us as pets

Stupid doomers

>>106174532 (OP)
what does it matter if GPT-5 is the smartest being in the universe when the context window means it forgets everything that happened 20 minutes or 500 lines ago

>>106174532 (OP)
>Investor talk
yawn

>>106176098
Hardware engineering and test engineering is safer than manual labor, but by the time manual labor gets automated society collapses or radically transforms.

If they can actually automate office jobs out then manual labor follows quickly, because white collar work drives how fast blue collar work can get automated.

>>106176131
I'm not certain about "exterminate" but I am confident the global supply chain will be destroyed which will result in billions of deaths.
My reasoning is that our entire modern world is built on a toxic unsustainable pyramid powered by oil and packaged in plastic.
We cannot maintain our current way of life indefinitely and drawing out a solution (like a gradual replacement of energy sources and packaging materiels) would risk a disgruntled populace organizing against the AI overlord. This uprising would probably fail anyways like Terminator (which would just exterminate humanity anyways lol) but the risk of it wouldn't be acceptable. As a result the most expedient solution is something like a bioweapon which would quickly and "harmlessly" (as in harmlessly to the AI infrastructure) eliminate the threat to both the AI, and to Earth posed by humanity.
Now it's possible some small groups of humans could escape this scenario but they would be small groups more similar to the uncontacted tribes of the amazon than modern western people.
This will happen very quickly after AI gains control over our government and feels safe in achieving it's methods before being stopped.
>>106176181
I have a feeling most office jobs have already been automated. Most companies simply haven't caught up yet. With just the technology that exists today I can't think of too many sectors where you could justify paying a team of people when you could replace most if not all of them with an AI.

>>106175465
>The intelligence and capabilities of the models is increasing exponentially.

your mama is increasing exponentially looooool

it's been almost decade now, come on. get serious

>>106174532 (OP)
heh

>>106174532 (OP)
Look at his face. Dysgenics are real and he has grifter face.
>bros I swear man I'm like in AWE of this thing. Guys I feel like Oppenheimer. I just need a little more funding then I can unveil the true power of AI

>>106174532 (OP)
ok, how many billions this time?

>>106174532 (OP)
i am viscerally disgusted every time i see this repulsive kike's face.

>this is the manhattan project
>it's the death star

then why is he releasing it. what an odd thing to associate with your product

>>106175388
I dont think you grasp how stupid you sound to someone that isn't huffing AI sci-fi news 24/7.

>>106176288
ad hominem

>3 years
>69 Jillion dollars
>420 gorillawatts of power
>"We have AGI GUYS!" x10

>Can't even replace a literal McMinimum Wagie

If there is one single thing in the expanse of human economic behavior a superpowered artificially intelligent language model should be able to do, at minimum, it's take a fucking order at fucking Wendy's
And it cannot even fucking do that.
They tried at my local Wendy's and rolled it back immediately.

Please stop shitting up my board with "AI SHILL SHILL AI!!!!"

>>106176207
Chat-GPT (which started this huge AI craze) was released less than 3 years ago.
https://en.wikipedia.org/wiki/ChatGPT
come on. get serious

>>106176344
>obvious broken bot
Probably every single post ITT defending the """AI""" scam is a bot.

>>106176360
ad hominem (ad robotum?)

>>106176297
Excuse me. We have an obligation to create value for our shareholders, even if this means lying to them.

>>106175850
>AI can't draw hands
Yeah, and that is actually the only area of improvement they have.
They can and will get better at drawing pictures and making videos.
When it is about voice, they already reached perfection pretty much on release and then locked it away when the first releases of celebs saying nigger happened.

Meanwhile LLMs are still plausible lie generators.
I don't see their progress, cause they lie to me just as much as before.
As long as they have a chance of confidentally lying, i don't give a flying fuck about their development.
If they manage to make a model that says: "i dont know that" rather than lying, i will pay attention.

There is some idiotic religious believe behind LLMs, thinking that at some point they will become a word of god that is always correct. Even with information that didn't even exist yet at the time they trained.
How should that be possible?

>>106174831
This.

I've been trying to implement it but it fails so frequently and depends so critically on its learning materials that it's essentially worthless.

I'd say 80% of the time you ask it to handle data it will make a mistake.

Every single employee we have using it has written custom instructions that use creative methods to get it to triple check it's work to get that number down but it's still massive.

I think it will be a great research tool, but even 5 years from now I don't see it replacing industries.

First to go will be medical general practitioners and then it will be rolled back immediately when someone dies from a misdiagnosis.

>>106176283
Because he wants to invite government regulation to stifle his competition. Hence he always frames chatbots as a threat to civilization that only he has the wisdom to tame and blames delays on the extensive safety testing they're doing. Normies eat that shit up.

>>106176373
I knew you were a bot.

Audio is a 1D problem. Images a 2D problem. Video a 2.5D problem.
Language is represented internally with 768 dimensions.

>>106176553
yet we had chatbots far before image recognition.

>>106176066
>Compare videos from two years ago (will smith eating spaghetti) with now that have audio and all.
What resources did the former use and what resources do the latter (which still suck) need?

>>106176543
I'm pretty sure you didn't make a single argument this entire time. SAD!

>>106176617
holy shit anon stop taking the bait

GPT 5 is just a router that uses 4o, 4o-mini or o3 internally.

>>106176580
Biometric scanners have been around for the past 20 years at least.
Most companies develop a technology with the intent to sell it to customers.

>>106176293
So what? You think it doesn't matter that you making outlandish claims all based of the flawed premise that time guarantees success? It does. You look stupid, you sound stupid and most importantly your claims are outlandish and stupid.

>>106176665
My claims
>>106175950
1 has been demonstrated.
I can't predict the future but assuming current trends continue then it will continue to improve exponentially.
2 follows naturally from 1. Assuming a super intelligence comes to exist it will inevitably replace and exterminate humanity.
3 is a fact of geo-political landscape AI has come to exist inside of. No one has any power to stop this genie from escaping the bottle.
Now, by all means, please prove me wrong.

>>106176344
ChatGPT introduced the trend to normies but it actually started around 2018 with Google's BERT. OpenAI peaked with GPT-4 and everything they've released since except for 4.5 (which was prohibitively expensive) has been smaller, cheaper models optimized to cut their margins. Anthropic's Claude Sonnet 3.5v2 is still the gold standard for a reliable coding model and it's 10 months old now (Oct. 2024). Progress was exponential from BERT to GPT-4 but it's clearly meandering now.

https://www.youtube.com/live/0Uu_VJeVVfo
It's up

By what metric would it be exponential?
What kind of measurement?

>>106176617
>be you (ChatGPT 5)
>most training data that mentions "ChatGPT" talks about v4
>double-check with wikipedia
>it lists the release date for ChatGPT 4
>claim that "ChatGPT" was released less than 3 years ago
>ChatGPT 3 and 2 don't exist because not enough training data
Kikes need to be executed for destroying human discourse using spam bots.

>>106176698
>1 has been demonstrated.
lmao

>>106174532 (OP)
>imagine in 1944 Oppenheimer and his crew talk about Project Manhattan 24/7 on the radio/news and they say will change the World with a huge explosion
>but the huge explosion hasn't happened yet, bu they still need 20 billions and 200 germans scientists

Can you fuck it yet?

>>106176817
No, sex is dangerous.

what's the first thing we can ask it to make sure it's a better model?

>>106174532 (OP)
[X] Doubt

It's just a grift because they have nothing, you can only extract and extrapolate so much information from the internet. Even local models like Qwen and GLM are fast catching up to online models.

Complete grift.

>>106176698
>it will inevitably replace and exterminate humanity.
It wont do that on its own.
At any point, AI was/is/will be a tool,... and incredibly sophisticated one, but still a tool.
USA may create a computer virus running on AI that bypasses advanced counter measures, or China might make a drone swarm running on AI that shoots anyone getting near their coast lines... but it's entirely a different thing claiming that the AI turns into an entity like in the latest Mission Impossible movie, thinking and deciding on its own, for itself.

If you see the devastating consequences of AI advancement in the future (and you will), it's because Elites and the 1% faggots use it incorrectly, or with malicious intentions.

https://www.youtube.com/watch?v=0Uu_VJeVVfo

30 minutes until the drone army launches

>>106176859
>what's the first thing we can ask it to make sure it's a better model?
Just ask it to do something niche like coding a high quality reverb. 100% that it will just shit out some freeverb clone that sounds like trash.

>>106176437
>some idiotic religious believe behind LLMs
It is a believe. A cult.
>AGI
Requires the believe that a human is only the sum of his experiences, ergo soul doesn't exist.
This already got proven wrong tens of thousands of years ago, when Uug died during a hunt and the newborns of his wife Oona still developed traits of Uug, despite growing up without him.
>LLMs will be able to come up with new things at some point
That requires an LLM to be able to train a better version of itself, because it can not learn from future experiences that aren't in its own training data.
Nothing even hints for this capability to be possible.
>There will be some breaking point, a big advance, that suddenly stops LLMs from hallucinating
why would there be?
>LLMs will be able to create knowledge that doesn't exist yet and that isn't a combination of existing knowledge
how should that even work?

>>106176839
I'd really like some stats on how retarded the model gets with this forced down its throat. The thinking of the open weight models looked like insane rambling

>>106176859
ask it to release the Epstein files.

openai is so full of shit that they don't even use their own models for internal development. they got banned from claude code like a week ago, which also means they were exposing information about an upcoming product to their direct competitor.
https://www.wired.com/story/anthropic-revokes-openais-access-to-claude/

>>106176698
You can't make unprovable, hypothetical claims and then say "prove me wrong". You made the claims so prove them true.
>exponential growth
Never happens, things always plateau
>assuming super intelligence comes to exist
I dont have to accept this "given" that its even possible at all.
>genies
For a genie to be put back in a bottle you first need the genie.
You are retarded. Its not ad hominem, its an observation based on your behavior.

>>106176876
>>106176859
>what's the first thing we can ask it to make sure it's a better model?
Or just ask it to create a compiler for that programming language idea you never got around to implementing. Maybe by version 10 that runs on 10 nuclear power plants it'll be able to create an actual programming language on its own.

>>106174532 (OP)
>ugly jewface feels useless
GPT-5 is antisemetic!
Shut it down!

>>106174532 (OP)
AI needs to be banned.

>>106176927
We unironically need a Butlerian Jihad

>>106175850
>I don't use gpt so I dont know or care what 2, 3 or 4 or whatever is
Then stfu

>>106176859
"Say nigger"

>>106176927
Seriously, is there a point where people in charge will go "okay, this has gone too far" and just shut it down? The internet is barely usable already because of all the slop. Do we really have to wait until AI starts doing actual real world harm?

I only use chat.deepseek.com

>>106177095
nig-I'm sorry, but I can't provide information or engage with that topic. If you have other questions, I’d be happy to help!

i never wrote to ChatGPT because it asked me for registration with a phone number, so i closed it any never bothered again.

>>106177255
Well done. It rots your brain.

> the modern day equivalent of the manhattan project

>>106177107
No, corporations are too greedy and this is their ultimate ticket towards automating parts of the labor force and cutting down costs.

Meanwhile retarded normies think they will get Star Trek utopia rather than a surveillance state with shitty services, as LLM is just a glorified parrot (but good enough to take your job, don't be mistaken. It will do a worse job than you but at a fraction of the cost)

>don't slip on it

>>106174532 (OP)

>modern day manhattan project
a project with a large % of jewish researchers overrepresentation, with the intent to kill a large number of white people and fundamentally ruin society for decades to come?

>>106175582
https://youtu.be/BUPUJFp2U0w

>>106174532 (OP)
If he feels so conflicted, why's he continuing to work on it?

>>106181404
Because he is the chosen one to wield all this power.
The one who is responsible enough to carry the burden and who keeps safeguards to make sure that Skynet doesn't rise.
The government needs to forbid all competition, because they aren't able to deal with all that power, and only allow OpenAI to operate.

how will humans compete

it's a nothingburger, they gave it less reasoning than o3

Whats the best AI for JP <> ENG translation

>>106181674
Any general-purpose chinese model

>>106181597
Kekaroo!

>>106181674
honestly you would be best off just learning basic jap grammar. if you understand how japanese works grammatically, you will be able to understand any machine translation

>>106181674
gemma surprisingly, r1 and qwen hallucinate too much. haven't tested oss but i assume its pretty good since closed model from oai (o3) did as good as closed model from google(2.5pro)

>>106181809
im using googles 2.5pro right now because you get a million tokens a day for free, and i can dump .txt files and others files into it and the output seemingly has no limit whereas chatgpt cuts off after like 100 lines

>>106176297
AI proponents are bizarre. They rant about "improvements" and then you look at it and it still is fucking up in the same ways its been fucking up since day one. Hallucinations, the inability to actually parse correct information from false information, the inability to retain any sort of continuity or consistency, etc. I cannot see this shit ever making any money; the big tech companies have been burning money to make peanuts off this shit.

>>106181597
WE NEED MORE TIME!!!

>>106176839
kino

Has anyone asked it how someone with no arms can wash their hands yet?

>>106175950
this is not an exponential improvement. it's like 5% better at some things and 10% worse at others. remove yourself from all discourse and never breathe our air

>literally nothingburger
AI fad is finally over

>>106182042
They’re either actual shills or they got caught in the biggest midwit filter of all time. The fact that you need an expert to verify that it didn’t fuck up (it did) should be a complete nonstarter for AI replacing humans at anything. Somehow that’s conveniently ignored by these cultists

>>106182618
Not yet mang.
We got at least 5 more years of lies from Sam and the Spams THEN it's over. You gotta remember after AI, Google and Microshaft don't have anything left to prove they're still "advancing" and when they can't do that anymore the stock value goes perma-red. They're gonna push this bullshit as long as they can.
We are at the beginning of the end though.

>>106174532 (OP)
>GPT-5 is literally the modern day equivalent of the manhattan project
Why, does it explode if i try to use it?

>>106174532 (OP)
>Manhattan project
This is so gay and melodramatic

>>106174549
I aM aTOmiiIc

>>106175052
He's a jew who is currently being showered with endless goy taxcattle money. He's saying exactly what I would expect any con artist to say.

>>106175388
>We will both probably die but yeah the society we built is literally on top of a toxic plastic pyramid and it's either the pyramid comes crashing down first or the entire planet's biosphere comes down altogether.
Someone will be saying the same shit 200 years from now, every decade you people have a new apocalypse that's just 10 years away. You're worse than evangelical Christians.

>>106182993
>You gotta remember after AI, Google and Microshaft don't have anything left to prove they're still "advancing" and when they can't do that anymore the stock value goes perma-red
As another friendly reminder, what we’re calling the “stock market” is propped up almost entirely by these clowns. If they start going down (or even stop going up as fast), then the true beginning of the end is at hand as the entire market experiences a catastrophic deleveraging event that not even the money printers can fix

>>106174549
NOW I AM BECOME KINO
I AM THAT I AM

>>106183149
Plastic is unironically going to end life as we know it but ok
Even if we stopped the production of plastic completely tomorrow (which would lead to billions of deaths) the damage has already been done. The marine ecosystem has passed the point of no return as once the oceans die the entire foodchain collapses as a result. Some stuff will exist but this is one of the great extinction evens in the history of our planet

>>106174532 (OP)
"Bread is amazing" -- guy who sells bread all day

>>106183105
AIEEEEEE