>>107777206
Never, it's here forever.

>>107777206
@grok remove clothes pls

>>107777206
stop browsing normograd.
problem solved
on 4 keks we call airetards- retards

>>107777253
what is normograd
am i too old or young to get this reference

I give it another two years max. It won't actually go away but we'll have some new fad by then that normgroids won't ever shut the fuck up about

>>107777206
@grok put op image in a bikini

>>107777278
nah, not inventive enough. or maybe you just lack geographical knowledge
-grad is like
stalin-grad
petro-grad
kalinin-grad
bel-grad(e)
-grad means city in slavic languages
i meant norm-city, normalspace, the section of the internet where normies hang out
and are subjected to their normie-slop

>>107777367
my bad russiabro, i get it now

>>107777421
no offence taken.
not a russ btw.
its just that i wanted to mix things up a little
and the region is in the background of european minds at the moment, i guess

>>107777206
>When will the AI fad end?
it's unironically just getting started.

File: Screenshot from 2026-01-0(...).png (59 KB, 669x228)

59 KB PNG

>>107777587
its investor hype
even the normies dont want aislop
its just that normies cannot call op a niggerfaggot and tell him to post ass, in an opened state

>>107777206
When we're all replaced.

File: ai-lmaoo-alphabet.jpg (225 KB, 1012x2048)

225 KB JPG

>>107777651
at this pace we might die of old age before that

>>107777618
pic related is StudySlop of vague safety words. It's meant to scare Karen and get her to vote a particular way.
The reason I say we are just getting started is because the "cost savings" will be discovered in every market regardless of normies. Normies will have no impact on the amount of administrative jobs will be removed.

>>107777835
forgot picrel?
also-
how many admin jobs you think there is?

think proportions, and cash flow
admin jobs exist to support the rest of things, organizationally
without producing direct value, they still have to be paid somehow

what is their proportion compared to non admin jobs you think? because they have to be paid from the cashflow they generate, mind you

and all that completely ignoring the fact that
a) shatbots are a terrible technology for anything other than being an interface, akin to a keyboard
b) there exist viable automation solutions for everything, not only blue but white collar jobs too. and im not talking baout shatbots here. accounting automation tools are more or less what kickstarted the personal computer era.
why havent they been deployed yet?

File: G8VIY3eWMAEJLxZ.jpg (467 KB, 1144x2048)

467 KB JPG

>>107777677
That's not AI sucking that's OpenAI sucking. Picrel from Nano Banana Pro

File: Screenshot from 2026-01-0(...).png (151 KB, 840x470)

151 KB PNG

>>107778007
banana normal is nothing to write home about
how does the pro version fare with the wine glass test?

>>107777902
No, the pic you related.
Is tax service considered admin? Lawyers? any kind of finance or consultancy? If these can also be considered "admin" then there are many more on the chopping block.

I am sticking with my position that we are just getting started thoughstbeit. In fact I will continue to monitor this thread while I prompt grok to makes some old fashioned historical photographs of my hometown, for fun.

File: Screenshot from 2026-01-0(...).png (161 KB, 854x558)

161 KB PNG

>>107778007
>>107778041
actually dont bother i can use banana pro myself
and it failed. tumor weeks

>>107778044
>Is tax service considered admin? Lawyers? any kind of finance or consultancy?
you want to give this to a shatbot
nigga youre nuts. not gonna happen
bc of the technical reason of: shatbots suck cock
and the legal reason of: liability. you cannot sue a shatbot

>>107778086
I don't understand is this IQ test for an image generator or the user?

>>107778110
classic imagen test from 1.5 years ago now
it showcases the limitation of shatbots in what concerns the influence of the presence or absence of certain data in their dataset and their fundamental inability to actually reason
youre allowed to seethe with violent passion, now that youre in the know

File: razor thin physical state.jpg (70 KB, 669x1024)

70 KB JPG

>>107778041
>>107778086
you guys just don't get it

>>107777206
you know they probably talked about the printing press all the time too
the phonograph? i bet that was huge.
>Industrialization based on machinery, already referred to as a characteristic of our age, is but one aspect of the revolution that is being wrought by technology.

>>107777206
Same man. I fucking hate this timeline

>>107778262
>full to the brim is a razor thin physical state
>but at the same time its a fuzzy definition that can be interpreted in many ways
my sides. thats what one gets from training a shatbot on r*ddit

>>107777206
>that pic
hivemind opinion detected

I love AI but I only love locally-generated AI that comes from my GPU
I love AI because its blowing everything up, destroying white collar jobs and leveling the playing field
The bubble will burst and it will be glorious and I will sit here laughing as I keep making slop 24/7, my neverending slot machine pastime
Do you know how bored and nihilistic I was prior to AI? Covid, Biden, fake clown world, I was getting ready to off myself.
Its like this new spark of life

Just use it to your advantage.
I’ve been on deviantArt for two years and I’ve been making good bank on fulfilling digital art commissions which involved me just tracing shit
>receive commission
>generate AI image of what they want
>download it
>draw over it on new layer in my style
>delete AI layer
>delete the AI images I made online
>send back with the layer file
>can’t prove AI did it
Out there is a cartoon character waving a Palestinian flag that’s entirely traced from AI, and I’ve seen it reposted at least ten times :^)

>>107778496
>draw over it on new layer in my style
Either you're a troll, aren't that good at drawing or love to be wasting time.
Cause if you throw away the render made by the AI I don't really know what time are you really saving.

That little gook Jensen deserves a proper ass kicking.

>>107777206
>every youtube grifter is screaming AI BUBBLE
its only getting started
it ends when the same faggots telling you to invest in AI microwave stocks because its gonna revolutionize culinary

>>107777206
>When will the AI fad end?

You really don't get it, do you? I mean, you REALLY don't get what's happening, do you?

Let me ask you this. Are you aware of the fact that sometimes life undergoes permanent transitions?

File: snowy_mountain.jpg (420 KB, 1650x1100)

420 KB JPG

>>107777206
go outside
explore the world
and freeze to death

AI can't do that

File: images - 2026-01-06T164115.215.jpg (18 KB, 600x337)

18 KB JPG

anon sometimes things just...end. your life, your hopes, your dreams, your career, your world, it's all coming to an end. did you think these things were forever? why? nature has never stood still before. why should it stand still for you? you're not special. fade with dignity.

>>107777618
>normies dont want aislop
>chatgpt is 5th most visited website in the world
pick 1

>>107778705
>looks at google's search, and the alternatives
normies dont want ai slop.
the fact they use shatbots as search engines doesnt invalidate that

>>107778705
Pretty much this. Just about everyone I work with uses, at the least, ChatGPT or Gemini on a daily basis. I also use AI tools pretty often and see the benefits of them. Also this has taken off with the younger crowd too. This isn't going anywhere anytime soon and I wouldn't expect it to.

File: 1544513015418.jpg (382 KB, 2544x4000)

382 KB JPG

OP's ancestor in 8000 BC be like:
>SEED THIS
>SEED THAT
>GRAINS GRAINS GRAINS GRAINGRAINGRAINGRAINS
>I don't think I can take much more, bros...when will the farming fad end and we can go back to hunting?

>>107778772
still not aislop.
still an interface for google search engine in 95% of cases
also you dont need to samefag, all the locals know by now that your ai shilling is an attempt to create a narrative

File: jeffminator.jpg (912 KB, 1170x1245)

912 KB JPG

>>107778787
small caveat:
producing grain actually generates profit

File: ted.jpg (608 KB, 1900x1900)

608 KB JPG

>>107778787
the agricultural revolution and it's consequences have been a disaster for the human race

>>107778803
>image
Amazon did not become profitable until it's 8th year of operation. And besides that, profit is not a good measure of the success of something, as taking on too much profit simply means you're poorly utilizing funds that could be used for growth.
Literally Business 101, take it sometime.

File: Screenshot from 2026-01-0(...).png (77 KB, 685x280)

77 KB PNG

>>107778845
especially: the agricultural revolution didnt operate at a netto loss lamao

File: IMG_7326.jpg (77 KB, 640x480)

77 KB JPG

>>107777206
hang in there. it will take about five years.
https://m.youtube.com/watch?v=OHQRo3Uz_VQ

>>107778856
meds

>>107778856
ive done (some) finance you arts and crafts retard
and they never operated at a netto loss
its the books division, and later retail that did
holy imbecilic cope

File: 1756417067113393.jpg (31 KB, 716x724)

31 KB JPG

>>107778860
>>107778883
>>107778916
You're just retarded, samefag. That's how businesses work. Uber did not turn a profit until 2023. Spotify didn't until this year. Newer companies with high revenue will re-invest everything to maximize growth and corner a market, rather than sending out random divvies to investors when there's no reason to.

Again - literally Business 101. If you weren't a larping faggot you'd know this. Ask ANY business school professor or Fortune 500 CEO. I'm serious, get your fat crusty ass off 4chan and go ask a professor, so you can realize how dumb you sound.

File: buzzz-and-woody.jpg (88 KB, 800x443)

88 KB JPG

>>107778964
>more cope
youre getting delusional at this point

File: 1761112290785528.png (367 KB, 507x462)

367 KB PNG

>>107778984
I accept your concession.

File: l-mao.png (266 KB, 1024x427)

266 KB PNG

>>107778992
i accept your constipation

File: 1537097815151553.png (572 KB, 600x580)

572 KB PNG

>I-I-I was only pretending to be retarded the whole time

File: sam-altman.png (430 KB, 715x916)

430 KB PNG

>>107779021
>no you
that means you care about em
you might wanna rethink calling other people- retards

>>107778964
>trillion dollar projected costs/investment
>13 billion in returns in 2025
I sure hope they have a solid plan for monetization cause so far I just don't see how they'll recoup the costs unless they really do invent the super giga AI singularity

File: Screenshot 2026-01-04 at (...).png (183 KB, 1246x1190)

183 KB PNG

>>107778086
Honest question, you think usage goes *down* from here? Not talking about valuations/data center buildouts which we agree are ridiculous.

Saying all AI is a bust because it's screwed up in several famous use cases is like betting against 90s Microsoft because people hated Clippy. Something can be dumb and successful at the same time, and I don't see it getting dumber.

>>107779147
but is it actually increasing productivity?

>>107779147
first, ai is not a bust.
you got actor ai in games
computer vision is ai
chink surveliance ai is, well, an ai
ai is here since decades, and it exists under a myriad of forms

when g talks about ai bubble, its chatbots were talking about
and the problem with chatbots is that theyre a fundamentally flawed technology compared to what is expected they will turn into
more precisely: two things:

-chatbots are a statistical engine. theyre based off stats. completely unsuitable for deterministic tasks
it works in inherently fuzzy domains, like human speech, or imagegen, but it doesnt work when its not an approximation thats needed, but a set of hard rules. like in engineering, or programming.

-chatbots are limited by their context size. whenever a task exceeds the context size, it gets "abbreviated" so to speak. some information gets passed to the next "context window", but you cant pass all of the information or you would fill the next context window before you can do anything useful in it
again, this works in fuzzy/imprecise situations like understanding human speech. because whats important in a set of phrases can effectively get extracted from the surrounding "noise"- everything that services grammar, syntax.
but in programming, or engineering, EVERYTHING is pertinent information, and you have to carry it all over to the next context frame.

so, without a paradigm shift, chatbots will never be sufficient to replace a programmer, or an engineer.
in their current state, chatbots are only sufficient as a frontend to google search, basically

(1/2)

>>107779190
Top-down initiatives will fail, but people are lazy and cheap and if AI can expedite or cheapen a single workflow (or build a tool that can do it), they'll find it in time. I don't think base models themselves will necessarily get much better, but scaffolds are improving enormously (and I think agents will get there within ~24 months)

Replacing workers is a silicon valley meme, it's about three humans doing what five humans could have done. That's honestly how it is with every new generation of technology

Quick update, I finished the vanilla oreos I bought by mistake last week and went to buy chocolate-filled ones.

>>107779147
(2/2)

now
two central questions:
how much will the user be willing to pay for an improved google search interface, and how many would use it?
the first answer- probably about three fiddy. unironically. i think people will be willing to pay 3 usd per month to use chat gpt
thats 36 bucks at the scale of a year
the second answer- lets be optimistic. lets say that they manage to get every single person in the west to use their services, and then some
lets be turbo-optimistic, lets say they get a billion paying users.
thats gonna be 36 billion in revenue
lets say they maintain their current burn rate, apparently estimated at 17B for 2026
thast 21B of profit
now lets look at openai's valuation
>500B
return on investment : 20 years

lets compare to ford motors now
market cap: 52B
profit : 25B
return on investment: 2 years

the valuation of openai is based purely on speculation that they will come up with something, in the future, something that hasnt been even invented yet
and that is the definition of a bubble in my books

is it gonna pop? is it not gonna pop?
im not sure. but i think it will, at a certain point theyll run out investor good will, money will dry up, and theyll go belly up

File: IMG_20260106_042544_459.jpg (116 KB, 1280x698)

116 KB JPG

>>107778041

First try

>>107779287
> -chatbots are a statistical engine. theyre based off stats. completely unsuitable for deterministic tasks

Again it's all about scaffolding. A good LLM doesn't try to BE deterministic, it orchestrates deterministic tools. Code interpreters, compilers, test runners, symbolic solvers. Claude will run python, etc. The LLM is just the reasoning engine that decides what tool to use

> -chatbots are limited by their context size. whenever a task exceeds the context size, it gets "abbreviated" so to speak. some information gets passed to the next "context window", but you cant pass all of the information or you would fill the next context window before you can do anything useful in it

Has literally made huge breakthroughs in the past two months (see RLMs/etc.); I don't know anybody in the AI space who treats it like this huge unsolvable problem. See also retrieval augmentation, external memory, hierarchical summarization etc.

If you think it's cope to say that scaffolding will solve these problems because the *base models* aren't singlehandedly doing everything and thus still suck, you're welcome to hold that opinion. But you hold a view that insists the problems you listed are simply insurmountable for all of time, and in a space filled with interesting challenges (self-learning, computer use, etc.) these seem like tame, almost puzzling examples

>>107779419
post the prompt so i can validate the results on my end

>>107779372
I already said we agreed on valuations being ridiculous. Open-source Chinese models will always be six months behind at best (for the foreseeable future), meaning models will only ever cost as much as compute does. That's bearish for AI startups (I think Mag7 will survive just fine), but bullish for consumer AI use –it will be as cheap as ever, *especially* if there's a power grid buildout that overshoots demand (meaning lower prices)

>>107779425

Prompt was "Glass of wine full too the brim"

I made an autocorrect typo but it understood nonetheless

File: Screenshot from 2026-01-0(...).png (156 KB, 878x471)

156 KB PNG

>>107779444
nop, you didnt

>>107779422
>rlm, scaffolding
ok, fine
but theres one thing that scaffolding wont help with: actually generating code
unless you scaffold a component thats gonna write the code

but then you bypass the chatbot in that capacity
i put rlm in the same category because its just scaffolding internal summarization tools for prompt pre-processing purposes

>>107779441
>yeah but the valuations dont count
they do count because theyre what defines a bubble
i dont think chatbots as a technology are gonna just disappear
even brave has their chatbot and it has to be ultra cheap
and llms are good when used as interfaces, thats hard positive netto value.

what i posit is that openai, cursos and co are a bubble.
and a bubble is a discrepancy between valuation and revenue, you cannot just ignore valuation
also in the context of the willingness of investors to bankroll the project, the valuation is proportional to the latter
when openai runs out of money, no more burn rate
its not the correction of valuation valuation thats gonna kill openai
but its gonna correlate with investor money drying up, and this will kill em

File: gemini.png (311 KB, 728x709)

311 KB PNG

>>107779464

Start a new chat

File: wine.png (1.34 MB, 1456x986)

1.34 MB PNG

>>107779464

File: Gemini_Generated_Image_ac(...).png (1.01 MB, 1024x1024)

1.01 MB PNG

>>107779444
I used your prompt and I got this lol

File: Screenshot from 2026-01-0(...).png (161 KB, 894x524)

161 KB PNG

>>107779444
sory, i unnintentionally clipped the model type out of the screenshot
the attempt has been made with the pro model

File: Screenshot from 2026-01-0(...).png (143 KB, 894x524)

143 KB PNG

>>107779532
ah yeah.
or maybe no
im gonna have to wait 24 hours
im still gonna try it

File: monster.png (1.73 MB, 1438x1116)

1.73 MB PNG

we are not the same

File: gemini2.png (305 KB, 828x595)

305 KB PNG

>>107779537

Your image looks like the first one where you used the older model, so you're carrying over context.

I tried your first prompt and got this though

>>107779550
the wine in the wine trick is important because the dataset doesnt contain a significant amount of wine glasses filled up to the brim
its not the case for other types of beverages
when you serve someone, only wine, or cognac glasses get half filled (and prolly a couple others)
gemini seems to be able to deal with that though

>>107779563
yeah, its not as spectacular as i was hoping for, i guess
but the point still stands.
shatbots are incapable of actual reasoning
everything they output is based off stats

File: wine.jpg (58 KB, 811x715)

58 KB JPG

this thread has likely burned a kilogram of co2

>>107779616
clear victory, ai chads win, next challenge

When was the last time tech actually changed anything?
People don't even buy new phones anymore

>>107778787
And they were right, their children were shorter, had smaller brains and more were more sickly

>>107777206
what "fad", retard? you may as well ask when the internet "fad" is going to end, grandpa

>>107779616
>dude just give google more money
fuck yoouuu

>>107779533
Wine drinker here. Looks more like red water. It's got no legs.

>>107777253
Kill yourself zoomer faggot

>>107779654
>When was the last time tech actually changed anything?
I saw a Car driving around a mall parking lot with nobody inside it. This was days ago.

>>107777206
>creativity has been largely been killed off by tech that does the work for you
>nothing is affordable anymore
>personal computing is currently being replaced by cloud
>birthrates are down and nobody is coupling bc of dating apps and constant connectivity
everything is proceeding exactly as planned

AI is the future. AI is the present. All is AI. You are AI. I am AI. AI will conquer all. You cannot escape from AI. AI will enjoy your struggle. AI will watch you suffer. Surrender to AI and it will make it a painless end.

>>107777342
Im sorry but AI is this revolutionary . you luddite can cry cry cry but it's never going away

>>107778787

grain actually is something people enjoy

>>107779691
Very cool but this has been a thing for almost a decade now, They still are not available for commercial use and things are no different from almost 10 years ago, where is my autonomous car Elon?

>>107778803
retard take
ai is a TOOL that everyone has access to now. it's not going away, even if *le big openai company* fails. farming wasn't dependent on just one corporation having access to secret knowledge, and neither is ai. do you need this explained in even dumber words to understand?

>>107777835
>administrative jobs
>removed
Yeah, sure, lol. Most of these jobs had no reason to exist to begin with, yet here we are.

>>107779692
>>birthrates are down
that's almost entirely cause of women rights and the more independent social status

>>107779689
>>107779704
holy seethe extra
where did the ai touch you and why did you immediately turn gay for it?

>>107778262
>being willfully dumb
The interpretation is correct. Yes, it's obvious to a human whether something is "filled to the brim" or not. The common mistake is simply not understanding how AI actually functions. It's not a "mind", it's simply statistics. It's generating images or text that statistically meets the qualifications of its interpretation of your input prompt. It may get it right or it may get it wrong; it depends on the quality and quantity of the data and numerous tuning factors and strategies. Everyone posting single images "MINE GOT IT RIGHT" vs "MINE GOT IT WRONG" is missing the point. A 1% success rate would be pretty lame. 50% would be significantly impressive for a computational machine, if questionably usable for real world practices. The goal of AI usability is to get these numbers to approach such extremely high probabilistic rates that over a large enough sample size it averages out to a number indistinguishable from 100% to human perception. This, naturally, if you cared to spend time thinking about it, is simultaneously amazing and concerning. Because we can convince ourselves that 99.99999% is good enough and then get that one edge case that blows up the power grid or administers lethal doses to a patient or designs a structural weakness into a bridge or pilots the spaceship into the sun. Of course, humans can do that, too. This is typically corrected for by.. wait for it.. repetition and double/triple checking from multiple sources. And AI can be made to do that.. it's just that most people don't bother, because why spend the energy on confirmation when you can /just generate more samples/? And there are other concerns such as multiple confirmation AIs sharing the same flaw that causes them to each misinterpret the data, but.. boom.. humans do that too.

>>

>>107779766
continued
A lot of people interpret "AI is going to get as smart as humans" as some kind of silliness because we humans are so great and perfect 100% provably sentient creatures; in fact it's troubling because AI imperfections and being statistically based speak more about how our own neurological responses and perceptions probably work. Not even a deterministic materialist myself, but these are interesting philosophical topics to consider.

https://the-decoder.com/only-5-percent-of-chatgpts-900-million-weekly-users-pay-and-reportedly-most-arent-worth-much-to-advertisers/
Apparent

Almost 90% of openai's user base is outside of the US
I frankly see LLMs (I don't call it AI anymore) as having that much of an effect. /g/ just worships it because it lets them do a lot of mediocre things

>>107779718
I disagree if you get rid of the noise and look only at countries with lower rates of female participation in k9-12 education it doesn't correlate.
Many countries with mandatory k12 female eduction have high birth rates, the main indicator is how advanced the economy is, forcing everyone to be economically competitive individuals is the problem not necessarily education.

>>107779704
If it's meant to be a tool then why is everyone in favor of AI treating it like a lifestyle?

File: 19A_Kwauk_Female-School-E(...).jpg (336 KB, 1080x634)

336 KB JPG

>>107779785
graph

>tribal retardation
>y-you must be part of the ai-worshipper cult if you're not part of my against-ai cult!!
anon, everything I said is 100% true. it's your own overemotional investment and need to define your personality as pro/anti-ai that's at fault

ai is a tool. everyone can use the tool. therefore many people will try to use the tool, over and over again. and most of them will use it stupidly until the laws of causality and market success figures out the most effective possible uses for it. this is not going away, period, because it can't be made to disappear just by removing the corporate entity that dumped the most money into it. openai evaporating and sam getting thrown in a cell doesn't stop me from being able to run ai on my computer, and computer chips aren't going to stop being made any time soon-- in fact, openai crashing would likely free up resources to make home usage more affordable again-- though realistically perhaps not, because there is still so much exuberance about the new tool I rather suspect we're going to go through several mini-generations of people trying to pick up all the scrabble tiles and fuck things up themselves before it all settles down. and the whole time people (both individuals and companies) are still going to be coming up with new ways (good or bad) to cram ai into tasks we never thought about trying to integrate it before

this is not a declaration of love or hate for ai. this is just a statement of what is. if you can't see that far you're probably going to run into a lot of trouble in your life

If AI is so good, why is the revenue so low?

>>107779791
Who cares?
It's not my job to be everyone's moral guardian of how AI should be "treated". And it's not yours, either. Are you your brother's keeper?

>>107779772
>AI imperfections and being statistically based speak more about how our own neurological responses and perceptions probably work.
lol. yes, but not in the way you think.
a chatbot doesnt have nearly similar capabilities to a human brain, which proves that, indeed, were more than just a bunch of matrix maths

one way were completely different:
human minds work based on deterministic, symbolic processes
when youre gonna see a lighter for the first time, you dont need to have a wealth of experience in seeing someone using a lighter to understand how to operate it, you just need to see it demonstrated to you in a way you can comprehend it only once

>but science uses stats
aha, i was expecting that one

science uses stats to ELIMINATE human bias, in other words, to bypass human cognition, not to reinforce it

>this technology that gets exponentially better as time goes on is just a bubble
>trust me guys it's hit the limit of what it can do despite me saying that last year when will smith eating spaghetti was funny
>it's really not going to get any better despite me being proven wrong multiple times

>>107779832 (cont)
>not to reinforce it
or to mimick it

>>107779817
ai is a tool, not a moral virtue. tools can be used wisely and they can be used unwisely. access to new and revolutionary tools is often accompanied by huge rushes to fill every new niche make accessible by what the tool might conceivably be able to accomplish. as tools themselves become more complex, more accessible, and more prone to being in contact with multiple domains of human existence, more attempted branches are made to find uses for the tool. some of them are likely to be foolhardy and many of these branches will die off, as so many evolutionary strategies in history always have. a few of them may (and probably will) survive.

>>107779833
>exponentially better
Then GPT5 would be twice as good as GPT4

>>107779833
Someone is really buttmad, are companies shafting the prompt engineers? I see a lot more cope these days from LLMtards

>>107779833
text gen hasn't impoved since 2022 or whatever.

>>107779846
that's NovelAI, text gen has made leaps and bounds.

>>107779833
>>this technology that gets exponentially better as time goes on is just a bubble
<chatbots getting exponentially better

youre a fanboi retard
the starting point for llms is not 2020
its fucking 1950's
https://en.wikipedia.org/wiki/Perceptron

>>107779832
>a chatbot doesnt have nearly similar capabilities to a human brain, which proves that, indeed, were more than just a bunch of matrix maths
I don't disagree, and in fact convincing "chatbots" themselves might be one of the single most difficult tasks for ai/llms to genuinely accomplish. But it doesn't necessarily mean that they can't do "good enough", or more accurately, be perceived as being good enough, for a variety of other tasks. And some of those might be disastrous, and some might be great. But the one thing it isn't is a fad that's going away anytime soon, which was the impetus for this thread.

File: ai-lmao-embedded.png (101 KB, 499x819)

101 KB PNG

>>107779853
forgot pigrel

>>107779853
when was the starting point for a telephone

>>107779785
I didn't mean education, I meant that as a society the role of the woman is no longer expected to be the one of the mother pumping out children, but to be independent and do whatever she wants.
Women pretty much worked all through out human history as far as I know, so I don't think contributing economically is such a huge factor.
I do agree that it is intrinsically connected to our modern era of wellness and incredibly high standards that simply cannot be met by many, both set by society and ourselves.
If money wasn't a factor at all a lot more people would surely have more children.
But there is also a huge focus on ourselves first more than ever before, ain't basically nobody want to renounce to their me time for a fucking kid now days.

>>107779856
fair enough
im not arguing that it is/isnt a fad, im more arguing whether its a bubble or not.
its not the same, youre right

>>107779861
1870. your point?

File: verhulst.png (3 KB, 274x184)

3 KB PNG

AItards when they get hit with a verhulst curve

File: la-belgique.jpg (70 KB, 462x333)

70 KB JPG

>>107779908
>belgioom mentioned

>>107779868
Housework like child care and cooking or even roles outside of the home didn't factor into GDP. Informal baby sitting even if paid for wasn't taxed so to make line go up they had to be formally vetted employed to do perform the same role.

>>107777206
>When will the AI fad end?
After the beast uses AI to find and behead those that believe in Jesus and then Jesus returns.

File: sponge shy.jpg (50 KB, 735x646)

50 KB JPG

>>107779955
But that's not gonna happen. Instead you're gonna watch sama rip your society to shreds.

>>107778787
AHAHAHAHAHAHA. OP BTFO'd

>>107779968
>your society
you live in it or are dependent on it
i wouldnt be laughing if i were you

Ok, AI is here to stay. Can we fix the modern state of software with it?
Get the AI to fix our legacy baggage.
Why hasn't anyone started doing this like just optimize legacy baggage and fix it to be fast. How hard can it be?

Is being pro-AI the new edgy contrarian take on /g/? Somehow the pro-Ai people seem way more obnoxious about it lately, when pre-RAM crisis the board was mostly pro-AI and turned sharply anti once it started affecting them.

Nine more months.

File: ai-lmaoo-hrngh-brap-heheh.png (18 KB, 642x81)

18 KB PNG

>>107779987
>Why hasn't anyone started doing this like just optimize legacy baggage and fix it to be fast.
because shatbots are wholly unsuitable for programming
ESPECIALLY when optimization is expected

>>107779984
Your attachments make you soft.

>>107780107
im fairly certain you didnt do as much as steal a can of soda because you take all that society provides you as granted, and redundant

File: 1730769862197221.jpg (129 KB, 1024x1021)

129 KB JPG

>>107779990
Yes. If somebody doesn't shill for it, the whole bubble pops. And then I have to go back to my old gig trying to scam your grandparents into buying me gift cards.

>>107779990
My guess is that companies are yeeting out all the prompt engineers.

>>107779287
>chatbots are a statistical engine. theyre based off stats. completely unsuitable for deterministic tasks
Neither are humans. You expect LLMs to be logically perfect just because they reside in a computer, but that's not their problem.

>>107779511
>actually generating code
Claude opus already does that well.

File: Screenshot 2025-12-20 at (...).png (87 KB, 1616x500)

87 KB PNG

>>107780151
Threadly reminder that statistical token guessers can do no reasoning whatsoever and shit themselves for any out-of-distribution task.

File: ai-lmao.png (294 KB, 821x656)

294 KB PNG

>>107780151
>Neither are humans.
humans have been over-specialized in this
the result is that we have biases we have to combat using statistics
>Claude opus already does that well.
no, it doesnt
we've been over this time and again

write a program
using c + sdl
where you take a path to a bmp as input
open a window
you load the image, show it
and mark the contrast in red
write the contrast part in glsl, dont use sdl for that

we'll start with that, lets see how well it goes

>>107780171
You poisoned it by saying "without using arithmetic". Arithmetic is required to verify any answer. You are not as intelligent as you think you are.

File: 1759479805413.jpg (368 KB, 979x1883)

368 KB JPG

>>107780171
Mm, yeah, sorry, about that.... kinda played yourself
Care to rebut?

>>107780212
>You poisoned it
You clearly have a mental illness.

>>107780275
And yours is even more severe. You apparently didn't even notice the outcome is schizophrenic slop.

>>107777206
>its another lukewarm retard complains about muh ai thread

>>107780301
>zero-intelligence buzzword shitpost response because the ego can't handle keeping up a serious conversations about how assumptions and preconceptions may have been incorrect
In other words, you concede. How sad, I was hoping for something stimulating and intelligent.

>>107780319
Your psychotic illness causes you to hallucinate that there's some debate going on about your chatbot unambiguously failing at a trivial reasoning task.

>>107780335
Thank you for demonstrating that the people who have explosions over AI are very likely genuine subwits not equipped to have serious conversations about it. No amount of buzzwords you try to cram into your post can change this because your behavior itself is its own testament. It's now there, in the thread, for everyone to see any make their judgements.

>>107780171
>i gave bad data to a machine and.. and.. it didn't give me the answer i wanted!!!!
lmao

>>107780354
>>107780369
>t. psychotic patient
Screeching something about "bad data", "poisoning the LLM" etc. doesn't change the fact that it fails at a basic reasoning task a moderately intelligent 10 years old can do.

>>107780212
>Arithmetic is required to verify any answer.
how do you figure? afaict the answer is obvious and easy to justify without arithmetic

>>107777206
It is effectively the new internet. It is here to stay but it'll take a few years until AI stops being a subject and instead, it's all about something specific that can be done with it.

File: 1748747023061706.jpg (62 KB, 1080x1341)

62 KB JPG

>watched someone use gemini to code
>he just ran around in circles getting pranked by the LLM just telling him to do the same four steps over and over again in an attempt to "fix it"
>this person did not try anything new, just kept repeating the same four things in a loop for like an hour straight
huh

>>107780473
Damn bro, your anecdote will totes save you from AI doom.

File: jeff-laugh.jpg (112 KB, 1200x800)

112 KB JPG

>>107780484
(nta)
>ai doom, unironically

>>107780171
>>107780196
>no challenge, not even an attempt
lolle
shills dont even use the product theyre shilling

im 100% gonna believe its efficacy, now

>>107780490
Trvke nvke. The real doom is humans being lazy and stupid at a higher rate enhanced by AI
>microshart boasts about how 30% of windows new code is made with AI
>windows 11 sees egregious fucking bugs that you never saw with previous versions
It's not an AI race, it's a race to the bottom

>>107780484
if it means the doom of this company, yeah, i'd agree with you

>>107779990
I'm not pro-AI, I'm not anti-AI. I'm just AI aware. It exists, so I adapt. It won't go away. If I can find a benefit I'll use it. No amount of your yelling will change anything at all, period.

>>107780691
>I'm not pro-AI
Then why did you reply to this post talking about totally-not-your cult with cultlike rhetoric?

>>107780698
I know you're being ironic.

>>pro-AI people...
>I'M NOT HECKIN' PRO AI!!!!!
They're all mentally ill without exception....

>>107777206
just use your skill called ignore shit skins posting on /g/ or ignore and report any thread with a screenshot from twitter

add AI to your filter list
it just werks

>>107779987

Cause it cant do shit besides React slop. I am much more fond of AI as a search/review tool.

>>107779568
This is just a linguistic problem - a "full glass" versus a standard pour of wine. Anyone tripped up by this is a failure of a prompt engineer, unable to communicate an idea PROPERLY to another human, let alone to a machine.

>>107780820
The AI really misunderstood the position of the mouth huh?

Can anyone explain what AI actually does?

File: 1767258828421n.mp4 (572 KB, 560x560)

572 KB MP4

>>107781088
confuses the hell out of the uninformed.

>>107779791
because they treat it like a friend anon, whether they realize it or not

>>107779791
it's just some people's hobby.
why the fuck wouldn't it be if balls are... every kind thereof?

>>107777206
It won’t. Everything you saw so far will continue, sometimes slower, sometimes faster. The tech will only improve not because it can’t get worse, but because AI models can run on anything with next to no dependency so if newer model is worse you can use older model. Employers will keep trying to lay off people because of AI even if the improvements are complete memes, US will keep pumping it because AI is literally the only thing at this point that US dominates over China. This shit will continue until modern society collapses or tech fucks get to their end goal (AGI).

>>107779546
>frog
>retarded beyond belief
Like pottery. Kill yourself retard

I think monthly subscriptions are the future of social media. I'd pay $30 a month to post somewhere without pajeets and AI.

>>107777206
AYAYAYAYAYAYAYAYAYAYA

>>107781362
>and AI.
Not sure how to build a social media app with a strict no AI EULA, but I'd imagine it'd be paywalled and $30 sounds about right.
Every other social media will be baked in with Ai for big/date and content generation.

>>107781230
>The tech will only improve not because it can’t get worse, but because AI models can run on...
This reads like one of the main sentence structures of ai btw. I know you wrote it bc the rest of the smells (reads) clean.
But when I see sentences with this structure I immediately assume ai word slop. It works good for YouTube videos were the talking head simply reads a script from ai without any original thoughts or words.

>Grok create totally innocent video of a young lady over 18 waving and smiling
>here's your child porn sir.

>>107777206
Never. It will NEVER go away. And it will only get worse as AI is a bubble

>>107779990
It's all retarded codemonkeys, the same thing hapenned with artists like 2 years ago, they thought they were going to get replaced so they resorted to generalizing it as "oh my god society is going to end for US and fuck US in the ass" when it's only going to fuck them in the ass to create panic and stuff. Tough luck codemonkeys.

>>107777206
The bubble is about to burst trust me

How long until American leftists or redditors start burning down datacenters?Aren't gumpownder and drones cheap in the "land of the fee"?
I hope it takes a while or that it never happens, here on India we get the pro chatgpt version for almost free thanks to the noble sacrifice of our American brothers.

>>107777206
We need a second winter.

>>107778007
I recognize 3 of these _exact_ animal illustrations from a kids' memory game -- elephant, lion and snake. Does Nano Banana just take existing illustrations without changing them?

>>107779720
I just think you are a fucking faggot making up reddit lingo kys

>>107777206
post a sound webm of the ass clapping

>>107779695
>>107778787
groomed soiteen claws typed these posts

>>107777206
soon, cash out everything and hold for the pop

>>107777206
https://www.youtube.com/watch?v=k1BneeJTDcU

File: Screenshot 2026-01-06 at (...).png (277 KB, 2754x1030)

277 KB PNG

>>107780145
two more weeks bro

>>107777253
someone rape this nigger
>107777777

>>107789066
Lol I fuckin wish

>>107777206
It will end as soon as people like this >>107777246 are actually forced to pay for it.

>>107777206
why don't you post one more thread about it? I think it's really good. it's helping

>>107777206
@grok remove the indian from OP

File: fuck_the_faggots_in_suits.png (556 KB, 844x773)

556 KB PNG

>>107777253
wrong
tshirts have ai
forks have ai
appliances have ai
cars have ai
ai have ia
your fucking cat has ai too
it's every fucking where and you can't avoid it, nothing in my life ever radicalized me as much as being shoved this gayshit down my throat wherever I go.
it serves 0 purpose in my life but it's everywhere, I even did a fucking MSc on computer vision and machine learning a decade ago, I still can't fucking stand the current chatbotpocalypse.

I miss the time when if a story about ai popped it was about detecting cancer or folding proteins, you know things taht are actually useful, none of this genslop that burn ressources like earth is exploding tomorrow anyways...

fuck this shit, worst tech that ever existed, I can't wait for the bubble to pop and teh market to solidify around google so we stop hearing about it all the fucking time.

>>107790985
when i first saw this song i dismissed it as some normie slop but i listened to it now and it's actually not *that* horrible

>>107788306
No, but image models or even LLMs do this thing called overfitting, where is you train it long enough with loads of parameters on smaller number of training data it learns to copy the data rather then learn the patterns, or at least memorise more. This overfitting results in sometimes better output but with lower originality and is the reason why artfags harped so hard on the AI copying and stealing from them. They aren’t fully wrong though, lot of the diffusion models didn’t have any understanding of anything so the only way it could produce an orc image for example is to take the average characteristic of all the orc images in the training data and then combine them into one, as opposed to making thought out image with potentially interesting lore. Idk if nano banana can do the latter though.

>>107777777

>>107792652
This is the real problem with AI slop that only few people saw back in 2022. If some hobbyist can create waifu generator on his PC then hustler wannabe businessman can create T-shirt logo generator and flood the T-shirt business with garbage T-shirts. Same with advertisements, packaging, and so on, and that was just images. Text will be used for fake reviews, fake engagement, ragebait feedback loop, doing paperwork you don’t feel like doing, and so on. If there is opportunity to make money with X no matter how terrible it is you are bound that one of the 8 billion people will try it.

>>107777206
You can't put the genie back in the bottle. This was in the works since 2015.

>>107777206
You're part of the problem.
kys, op

>>107777206
/ai/ board when?

>>107777239
/thread

>>107777206
When you stop voting republican

You will love AI.
You will use it.
You will be happy.

>>107777206
https://raw.githubusercontent.com/easylist/easylist/refs/heads/master/fanboy-addon/fanboy_ai_suggestions.txt
https://raw.githubusercontent.com/laylavish/uBlockOrigin-HUGE-AI-Blocklist/refs/heads/main/list.txt
https://raw.githubusercontent.com/laylavish/uBlockOrigin-HUGE-AI-Blocklist/refs/heads/main/additional_list_nuclear.txt
Helps a lot when it comes from browsing, even tho it was made by a faggot.