Thread 106356145

What happens when the ai bubble bursts?

>>106356145 (OP)
We start another bubble about, like, sentient snacks or something.

>>106356145 (OP)
I don't think it will. People are disappointed that GPT5 was less of a jump than 3 to 4 but it's still impressive technology. I can easily see AI receptionists mass replacing human receptionists in the next few years. They only need to take calls and add appointments to the calendar, all within what AI can do right now.

The real goal of AI is replacing white collar work. A decent AI that can do spreadsheets can replace trillions of dollars worth of office drone salaries

>>106356157
>it's still impressive
So you keep saying, but the only thing that's impressive is the energy and water consumption of this glorified autocomplete.

I love Chinese technology and the Chinese people(well, the cool ones at least).

AI bubble will continue until advertisers stop spending money on Alphabet and Meta.

damn i forgot about that video. the glass of wine out of nowhere was perfect.

>>106356157
>it's still impressive technology
no, no it's not

>>106356157
>A decent AI that can do spreadsheets can replace trillions of dollars worth of office drone salaries
http://archive.today/T7DsU
>Researchers said “95 per cent of organisations are getting zero return” from their investments in generative AI, the technology that has sent US stocks soaring to record highs in recent months.
>“The story is spooking people,” said one trader close to a multibillion-dollar US tech fund.
>“Just 5 per cent of integrated AI pilots are extracting millions in value, while the vast majority remain stuck with no measurable [profit and loss] impact,” the MIT report said.

Fugazi.

>>106356145 (OP)
The whole economy collapses.

>>106356145 (OP)
I imagine a lot of rape.

>>106356202
Pajeet

>>106356145 (OP)
>What happens when the ai bubble bursts?
They will move to quantum computing at that point there are processors worth selling.

>>106356179
I think you're expecting too much when LLMs have only been around for 3 years. It's still very new. Computers were around in the 80s but mass computerisation of businesses wasn't really kicking into gear until the mid to late 90s

>>106356145 (OP)
I fear it's not a bubble, it's another industrial revolution.
The real bubble is the tech skills bubble.
Now that AI is a thing we just don't need tech workers like we used to.
There's going to need to be a lot of reskilling

>>106356163
It makes a lot of twitter art troons really mad, it's worth the investment for that alone

>>106356163
It's impressive that it can use natural language at all.

>>106356313
>>LLMs have only been around for 3 years.
>LLMs came out in 2022
Fuck off retard.

>>106356145 (OP)
Basically nothing, a few of the smaller AI companies will fail but OAI, Google, MS, etc are to big to fail and will get multi-billion bailouts to secure their position in the market.

>>106356145 (OP)
bail outs

>>106356318
>It makes a lot of twitter art troons really mad, it's worth the investment for that alone
This, all the terminally online freaks get mad about.

>>106356327
>a few of the smaller AI companies will fail
Shame.

>>106356157
>it's still impressive technology
Ah yes, let's keep thinking a giant web scraper that gives you the wrong answers with fake confidence is technological genius. In a best case scenario it will have nuclear powered datacenters that work at the capacity of a jeet call center worker, complete with the childish gaslighting and groveling.

>>106356324
>ChatGPT is a generative artificial intelligence chatbot developed by OpenAI and released on November 30, 2022
?

>>106356313
>I think you're expecting too much when LLMs have only been around for 3 years.
I'm not expecting shit, dickhead, I was literally quoting your post.
>A decent AI that can do spreadsheets can replace trillions of dollars worth of office drone salaries

No mention of in the future, or eventually or anything. And it's still wrong.

https://finance.yahoo.com/news/firing-700-humans-ai-klarna-173029838.html?guccounter=1

>After Firing 700 Humans For AI, Klarna Now Wants Them Back

The tech is not there, it doesn't seem to be getting there, either.

>>106356316
>it's another industrial revolution.
This is propaganda, there are no parallels with an industrial revolution.

>>106356318
Look, those are gonna be mad about something regardless, you can't base an entire imaginary industry around it.

>>106356449
>The tech is not there, it doesn't seem to be getting there, either.
People said that after Will Smith eating Spaghetti. The tech has come a long way in a short time. I can't see anything human receptionists do that AI can't eventually do. Likewise some data entry desk jockey. A few companies jumping the gun before it was ready doesn't mean much, LLMs will eventually be able to replace most white collar work, that's simply a fact unless it hits some massive wall in the next decade.

>>106356323
Not when it's copying it from the internet. It's a really fucking dumb regurgitation machine.

>>106356484
It's not copying from the internet though. It's creating entirely new text in a contextually appropriate way to respond to queries. You can't google search what an LLM says and see anywhere that it's drawn that text from. It's training data informs its responses yes but that's like saying a human is a regurgitation machine because you rely on data you've absorbed over your lifespan to inform your own thinking.

>>106356480
>it's simply a fact that this will happen unless it doesn't
The wisdom of tech bros is truly unparalleled.

>>106356504
>tech bros
Technology board retard.

>>106356504
Yeah I pegged you as one of the seething art troons. AI has already replaced your furry art slop commissions so I don't see why you're skeptical over it replacing other low skill low effort jobs as well

>>106356480
>People said that after Will Smith eating Spaghetti.
What's the relevance here? You seem to be throwing random, unrelated events around like they're gonna save your argument.

>>106356484
Before transformers no tech could do it.

>>106356499
>It's not copying from the internet though
Literally all the training data for the bigger LLMs came from the internet directly. Why are you even trying to discuss this?

>>106356484
Yep, and its even more pronounced for people like us that post on the internet a lot. We're the ones making the posts it learns from.
Imitation is a sincere form of flattery, sure, but you don't need a machine to look up an answer you gave someone else and give it to you. That's silly.

>>106356562
No tech could do what? Did you run out of script, Devansh?

>>106356559
>What's the relevance here?
It's AI? And people said it sucked and that it wasn't going to lead to anything. And it got better, a lot better in a few years. AI gets better over time so saying that Will Smith eating spaghetti proved that AI was no good back then is just like saying just because AI isn't quite ready to replace workers enmasse yet means that it's no good now.

>>106356572
It's not "regurgitating" it though. It trains the transformer. That's like someone writing a novel is "regurgitating" it from what they learned reading books.

>>106356619
kek AI can meme

>>106356589
There was no technology that could have a grammatically correct conversation in natural language. LLMs do have the problem that sometimes the information is incorrect, but when was the last time ChatGPT gave you a response that wasn't grammatically correct?
You probably just didn't use chatbots back then but I even worked on them in uni and achieving that wasn't possible back then.

>>106356619
>AI made Michael Bay obsolete
damn

>>106356639
Kind of wild you can actually have a coherent conversation with a machine and people scoff and say that it's not impressive. Talking with something that is not human and getting intelligent contextually aware responses isn't something that has ever been done before, it's a huge achievement

To all the haters of ai who say it’s useless, if it’s useless then why are half the threads on this board about ai or ai generated content? In only a few years it has taken over the internet and it’s only getting worse, I’m not saying that’s a good thing but it shows that tons of people are using it

>>106356675
Because zoomies are braindead consumers who can't think for themselves.

>>106356675
It's useful for mass generating low quality content

>>106356701
Shit AI is going to replace our posting on 4chan

>>106356619
CLANKER
SOVL

>>106356619
this is hilarious ngl

>What happens when the ai bubble bursts?
Same that happened to Internet after dot-com bubble.
https://www.youtube.com/watch?v=63f35w3j3IE

Also, generative AI should be a bridge to a truly AI, not the main thing.

Ideally: they get what they fucking deserve and we see a kind of soft tech reset.
Realistically, they'll keep pushing it until you literally have no choice (we're already kind of there, nothing ever happens)
Please let me be wrong.

>>106356619
The way they fell of the plane.

>>106356629
Fire from matches can't melt steel.

>>106356157
>People are disappointed
This doesn't have anything to do with the bubble. The bubble is bursting because AI doesn't deliver anywhere near expectations. Smart money started selling once it became clear agents were virtually worthless.
>They only need to take calls and add appointments to the calendar, all within what AI can do right now.
Here's the thing. It can't actually do this. It ends up doing something retarded. It can replace specific things receptionists used to do, but it can't act autonomously as a receptionist in a way that outperforms a traditional robo caller or whatever.

>>106356145 (OP)
America collapses and China takes its rightful place as the top of the world.
Mulattos will then complain about how it was the Chinese who lured them into the bubble, pointing at all those Asians working in AI companies.

>>106357113
It's only them who think they are in race, it's kinda like trowing a stick and the dog catching it, but you throw nothing and the dog keeps running at nothing.

AI can replace at least 75% of the workforce. Not sure what the CEOs were expecting. 100%?

>>106357175
>anon don't know ceos are trying to hold ai to keep their jobs

>>106356629
I must be a AI boomer because I unironically enjoyed watching this.

>>106356323
it's crazy how quickly normies forget about technology. AI in its current form has been going for like 10 years and even before that markov chain generators (basically current Glorified Autocomplete but not as good) have been around as long as computers have

the main difference between the old stuff and the new stuff is the new stuff has web scraping so it has a bigger data set to pull from

>>106356619
this actually fucking rules and no human could make something this funny

>>106356145 (OP)
The quantum computing bubble starts.

>>106356145 (OP)
>What happens when the ai bubble bursts?
wtf does that even mean? Do you even know or understand what you are attempting to say?

>>106356157
The problem with GPT5 was that Sam Fagman thought the last time people interacted with his chatbots was back when GPT4 released before all of the o1-o4 models. So compared to GPT4 from 2023 it is a huge jump, but compared to o3 it’s unnoticeable improvement. But the public has been using o4 mini and stuff like that for months now so they o4 was GPT4 because of the retarded naming scheme, and when they heard tech influencers talk about how GPT5 is worse then o4 or whatever they probably thought it meant they have not improved since 2023. Complete fuckup of a launch.

>>106357313
Because it's not that big of a jump form GPT O4.

>>106356145 (OP)
we will experience the heat death of the universe before ai collapses

>>106357234
Before 2022 AI was basically a nothingburger that people thought would take away first jobs in like the next 5 decades, but then it suddenly automated artists away which is something not even any sci-fi writers thought could be possible before AGI. And then very quickly it managed to produce text that was at that time indistinguishable from human text. It was a huge shift in public’s understanding of technology and a huge wave of investment happened shortly after. It’s kind of crazy how far just scaling up for us so far and how optimised the inference got.

>>106356157
Even if you believe LLMs will be significant or a gamechanger it's inevitable that there's a bubble burst. The dot com bubble burst despite the internet continuing to be a big deal. It's more about the 10,000 dogshit companies with no practical business model that have emerged alongside a few dozen companies that have legs.

Only difference is that IMO the only companies that ultimately come out of this are the ones who have physical assets which is why everyone is scaling up on datacenters and trying to get the government involved in the US-China "AI race" (fake and gay shit designed to convince people to shovel more money at billionaires).
The models themselves are sort of important but ultimately the people who make your model will be poached and open source will always be nipping at your heels.

>>106356413
This sort of thing has been floating around for at least a decade prior but weren't commercialized mostly because companies were convinced people wouldn't put up with the hallucination rates. GPT wasn't a quantum leap, it was a significant, but not shocking iterative improvement on what came before it that someone took a gamble on commercializing. And that hasn't even really been a success because OpenAI is just burning investor capital like every other AI company right now.

>>106357313
The naming scheme really is retarded. You're thinking of o3 which was their best model, then there is 4o which was the standard multi modal general purpose model

>>106357330
Yes, but compared to GPT 4 it is a big jump.

>>106357313
I honestly don't even really understand their naming conventions. In priniciple you would assume the reasoning (o) models would have distinct use cases from the non reasoning ones, but they just tuned the shit out of everything, so they 03 always gives better answers, but they also tuned it in such a way it doesn't even think very long. So no wonder most people treated o3 as gpt 5 because it basically was.
And they are doing the same fuckery with their shitty gpt 5 "routing", its really sending you all over the place and they keep adjusting the config in the background so you can't even really tell what specs you are getting on a weekly basis

>>106357373
>This sort of thing has been floating around for at least a decade prior
Not really unless you're somehow confusing pretrained generative transformers with prior chatbot tech which was much more basic and based on autocomplete. Googles Attention Is All You Need paper only came out in 2017, that's less than a decade ago

>>106357369
>but then it suddenly automated artists away which is something not even any sci-fi writers thought could be possible before AGI.
shitty anime was already being churned out by generators though. MUH ARTFAGS is just coomer cope

>>106357373
what a reasonable post. don't do that on /g/.
but yeah it's not that AI itself is bad (though it generally is) it's that there are a gorillion companies leaning solely on it, and it's just not sustainable financially. you can't have an entire industry that doesn't actually make any products. even openai/deepseek/etc which do have a "product" don't make any money off of it and have to continually beg for VC money to keep going.

>>106356157
The best thing AI can do is to make office work slightly more efficient. It always confuses me when there's so much hype about AI efficiency. It's quite similar to Amdahl's law, no matter how efficient you make one step of the process you'll reach a bottleneck because at the end you need a flesh and blood human to handle things, AI is not, and probably will never be reliable enough to outright replace humans.

>>106356157
That is not how it works retard. Investors so far haven’t seen the return on their investments since most companies haven’t managed to make any operational profits yet. High valuations alongside extremely high expenditure on data centres and energy consumption mean that the AI is a bubble about to burst.

>>106357463
The fundamental issue is that they are trying to focus on LLMs as AGI when its just really good at natural language processing and being a CHATbot. There are plenty of applications of machine learning that aren't LLMs, it just the hype is behind LLMs nobody really cares about anything else even though they should.

>>106357553
true, people should look up the progress in VLAMs, (visual language action models), it's gotten crazy and it's why I think robots in factories and industry are going to be the next big thing to take off soon

>>106357592
i dont care how advanced it is, it's stupid. an assembly line just needs to have the robot arm put the door on the car. it doesnt have to "learn" anything, it's the same movement every time

>>106357618
not every factory is an assembly line retard, and having a machine that can handle multiple tasks instead of being hyper specialized in one single task makes a lot of sense in many industrial settings (hence why we still employ so many humans to do mind numbingly repetitive tasks)

>>106357463
You underestimate how many seat warmer corporate jobs there are. A guy on reddit was complaining he got replaced with AI and his job was to be a human e-mail switchboard, routing mail between departments. There are thousands of really basic jobs that LLMs can replace because all they need is contextual awareness and the ability to go "This e-mail is requesting a laptop for a new employee I should send it to John who is head of provisions"

>>106356725
>is going to

>>106357644
What always amazes me is that jobs like that weren't automated much earlier with non-AI software.
Companies could also replace half their Excell monkeys if they knew what macros are.

>>106357224
More like a 12 year old Indian who also posted that shit.

>>106356663
>Kind of wild you can actually have a coherent conversation with a machine and people scoff and say that it's not impressive.
It's not, it's replying from context and from being trained on countless conversations online, there's nothing groundbreaking about this. What the fuck are you even trying to sell here? You're either completely demoralized or too far gone to bring up any actual arguments. You're so fucked up you're even replying to yourself. Does that ever work? Do you ever not get called out?
Your manner of typing doesn't change between posts, and you're too stupid to pretend to be more than one person.
Just stop, leave.

>>106356675
>y'all playa hatin
The absolute state of these scumbags.

>>106357787
The fact AI makes trannies like you really angry is the best reason to support it. A machine talking like a person is amazing and no amount of seethe that it's making better furry porn than you will change that.

>>106356145 (OP)
It doesn't burst. Big companies (the Five Faggots) have already invested so much into it that it'll be crammed and forced upon normies.
First is the desperation and $20 subscriptions everywhere and then, next Windows will have something what normies can't even turn off.
You know how it goes.

>>106357113
I knew it was wumaos taking a dump all over the thread, they're really fucking stupid and terrible at this subversion business. At least the moskals use cheap kindergarten tactics and completely derail discussions, these bug people can barely keep their shit together without bending over for their owner.

To add: primary use case is that AI is the perfect tool for surveillance.
But that's conveniently off the table when they are selling this big breakthrough of course.

>>106357787
>it doesn't count because it's achieved in X way
So what, the results speak for themselves. Noone was able to achieve the same result with any algorithm they tried before.

>>106357219
Yeah to hold on to AI, because their stockholders heard the word somewhere and demanded it to be the focus of their companies. This is all very depressing.

>>106357742
LLMs are more contextually aware and more able to take plain english commands and use that feedback to change what it does. Non-AI software would've been less accurate than an LLM reading the content of the e-mail and knowing by the context which department it should go to. When getting AI to do what you want is as easy as typing in a chat box "Do this thing" even boomer bosses can use it

>>106357279
Do you feel so emotionally compromised you can't even process the topic at hand?

>>106357644
"being replaced by AI" is mostly a meme. Youre literally just lapping up the altman propaganda.
The case you described has absolutely nothing to do with "AI" lol. In like 99% of big companies thats a regular process that gets routed to the person responsible automatically, and if someone sends an e-mail like that the answer will just be "did you create a ticket"?
Sometimes it amazes how gullible low iq retards on the internet like you are.

>>106357828
They're chasing AI because if they can replace even a fraction of white collar work for $10,000 a year subscriptions the economics of replacing $37,000 a year data entry monkies with an AI that can do the same thing 24/7 with no sick days or leave will make them hundreds of billions of dollars

>>106357369
>Before 2022 AI was basically a nothingburger that people thought would take away first jobs in like the next 5 decades, but then it suddenly automated artists
It didn't automate shit, people don't like that AI generated garbage and businesses don't even bother with it.
You live in a retarded fantasy, subhuman.

>>106357378
In what fucking way? People don't take your type seriously because you treat this casually and deliver nothing but vague bullshit.

>>106357631
the reason there's still swarthy humans making your clothes is that the indian sweatshop is more interested i paying 1$/h forever than making a 2bn investment into machinery to cut costs down to 0.93$/h forever

this shit will go the way of 3d printing -- it will be used by a few retarded techbro startups to make plastic crap no one wants. if it's not worth setting up a proper assembly line, it's probably not worth making ten million of them, and if it's a one-off part of significant value, it's probably not worth buying a million-dollar robot arm to do it when you can pay Jean-Paul 245$ instead

even if you really NEEEEEEEEEED robotic manufacturing but in a context that a dedicated assembly line isn't worth it somehow, CNC machines already exist and there are no manufacturing tasks that a machine should have to "learn", it should be provided a toolpath from the start

>>106357848
AI slop coomers think that anime pornography is a real industry

>>106357843
>"being replaced by AI" is mostly a meme
Not really. If an AI is capable of taking phone calls and putting appointments in calendars with minimal errors, and it's fairly close to being able to do that already, what's the reasoning behind it not being able to replace every receptionist? Or help desk call center pajeets?

>>106357644
That sounds like a job that could have been replaced with, like, a batch script. It's the typical "I made a script to do my job and I did nothing for 6 months" reddit story but with the worker that got replaced being the retard who couldn't into scripts.

>>106357848
>people don't like that AI generated garbage and businesses don't even bother with it.
Didn't Coke make its Christmas ad with AI last year? I saw a lot of troons seething about it

>>106357802
>LLM-worshipping mongoloid devolves into ad-hominems out of impotence and frustration
It's a system designed to impress the legally retarded and other dregs like you.

>>106357856
In simple terms, GPT4 was so bad at coding that people only used it for maybe being a stack overflow replacement. The current generation of models like the o series and Claude models are good enough to make junior devs unhireable.

>>106357644
>A guy on reddit was complaining he got replaced with AI and his job was to be a human e-mail switchboard, routing mail between departments.
that doesnt even need AI, just a couple of outlook forwarding rules lmao

thats really the dumbest part of AI -- all the stuff it's "replacing" are shit that are already trivial to automate without AI as it is

>>106357888
I hope you haven't cut your dick off yet you'll definitely regret it if you have

>>106356313
You’re wasting your time here talking sense. No one here remembers anything

>>106356675
>why is this thing that requires zero effort everywhere?

>>106357827
>the results speak for themselves
A net loss in profits for AI-focused companies, the need to rehire workers after mass firings prompted by poor implementations of LLMs, no visible progress in years as the technology plateaud.
This is the end of the line, and all you have are some horrible pictures to show for it.

>>106357867
Anime pornography probably has higher market valuation then the entire movie industry

>>106356145 (OP)
Just another tech bubble bursts. Maybe some big names will disappear. Lots of people losing their jobs and the entire industry shrinking to something more sustainable for the next ~7 years.

>>106357868
>If an AI is capable of taking phone calls and putting appointments in calendars with minimal errors,
this has been possible for decades and without AI

>>106357887
they did and the seething was correct. the ad looked like absolute shit, even for ai generated video. even the video file itself seemed to be pretty low-quality; there was noticeable pixelation on high definition TVs/monitors

>>106357910
if thats true why hasnt XXZERO put out a new ostrich bestiality book yet

>>106357868
So AI solves the receptionist issue and that's worth the $200 billion investment? Btw in practice these systems are worse than the existing robo callers which is why no major company has gotten rid of their existing systems for AI.

>>106357907
>Requires zero effort
>Creates something just as good if not better than a real artist
Sounds useful to me

>>106357847
Except those AIs don't run cheaply, and local implementations have been disastrous for businesses so far. Not to mention the PR hit they take when they do such things.
And surprise, people want to talk to other people on the phone or online, they don't want some hallucinating chatbox giving them the worst possible consumer support imaginable.

>>106357910
You don't actually believe this.

>>106357908
All I'm talking about is the ability to use natural language in a way that's miles ahead of old models.
Current LLMs will hallucinate, especially with complex subjects but with old models you couldn't even talk about the weather without it giving you gibberish at some point. That's an impressive technological achievement, the economics of it are a different issue.

>>106357917
There are still huge untouched markets for AI. I've noticed that there don't seem to be any porno movies that are made for guys like me. All the porn I've come across was targeted at beer-swilling sports bar dwelling alpha-males. Men who like their women stupid and submissive. Men who can only get it up for monosyllabic cock-hungry nymphos with gargantuan breasts and a three-word vocabulary. Adult films are populated with these collagen-injected liposuctioned women. Many of whom have resorted to surgery and self-mutilation in an attempt to look the way they have been told to look.

These aren't real women. They're objects. And these movies aren't erotic. They're pathetic. These vacuum-headed fuck bunnies don't turn me on. They disgust me. And it's not that I'm against pornography. I mean, I'm a guy. And guys need porn. Fact. "Like a preacher needs pain, like a needle needs a vein,"

Guys need porn.

But I don't wanna watch this misogynist he-man woman-hater porn. I want porno movies that are made with guys like me in mind: Guys who know that the sexiest thing in the world is a woman who is smarter than you are. You can have the whole cheerleading squad, I want the girl in the tweed skirt and the horn-rimmed glasses: Betty Finnebowski, the valedictorian.

Oh yes.

First I want to copy her Trig homework, and then I want to make mad, passionate love to her for hours and hours until she reluctantly asks if we can stop because she doesn't want to miss Battlestar Galactica. Summa cum laude, baby! That is what I call erotic. But do you ever see that kind of a woman in a contemporary adult film?

No.

I'm talking about the girls that used to fuck up the grading curve. Chicks with weird clothes, braces, four eyes, and 4.0 GPAs. Brainy articulate bookworms, with MENSA cards in their purses and chips on their shoulders. My porn starlets will come in all shapes and sizes. My porn starlets will be too busy working on their PhD to go to the gym.

>>106357917
>this has been possible for decades and without AI
The error rates are too high and AI before LLMs wasn't able to carry on a conversation with as much contextual awareness as LLMs can. You clearly knew when you were speaking to a robot before, it's much harder to tell now.

>>106357923
>AI is expensive to run
No more expensive than a gaming computer. Stop believing everything you read.

>>106357893
GPT5 doesn't seem to be any better, buddy. You need to learn the difference between what you want to happen and what is actually happening.

>>106357887
Yeah. It was actually really bad though.
https://www.youtube.com/watch?v=4RSTupbfGog

Like even if it wasn't made with AI, it's just really incompetently made. It was a bad ad.

>>106357930
>coomer cope

>>106357868
I'm sure you've had the "pleasure" of dealing with one of these AI service chatbots before. It's honestly shocking just how shitty and useless they are. I feel like they've actually regressed since chatgpt came out.
I've yet to get a better experience from a phone chatbot thats better than just "press 1 for billing information, press 2 for general feedback" routing that was possible 40 years ago.

>>106357900
>Trannies trannies trannies I need to change the subject aaaaaah
Why are you so bad at this? You remind me of those NFTtards. You were all in on that, weren't you? I can smell your loose Russian twat.

>>106357923
>Except those AIs don't run cheaply,
The cost of the AI is competing against the salary of a human worker. Even if they put the price up to $10,000 a year for each instance of it running the economics swing in favor of the AI.

>And surprise, people want to talk to other people on the phone or online, they don't want some hallucinating chatbox giving them the worst possible consumer support imaginable
They don't want to talk to Pajeets either but that hasn't swayed many companies now has it?

>>106357931
When I call in to make my appointment, I don't want a "conversation", I want to set my fucking appointment. If it were up to me, I would click a button on a date-picker widget instead of having to use a voice call in the first place

Between all the coomer shit and insistence that AI is LE JUST LIKE A RILL PERSON I'm starting to think that all the hype around AI is just lonely retards who want their receptionist to double as their cry-and-fuck pillow.

>>106357949
You can talk to Indians who can't speak English or you can have AI. The choice is yours.

>>106357930
>There are still huge untouched markets for AI.
I'm getting flashbacks from NFTs and "blockchain" implementations. An expensive, wasteful solution for a problem that does not exist.

>>106357923
The company I work at has one of those "local" LLMs and its just insane how dogshit it is even though its technically supposed to be GPT4. The only thing it gets right are asking it question about microsoft office tier questions. B2B AI is just scammy on a whole different level.

>Kind of wild you can actually have a coherent conversation with a machine and people scoff and say that it's not impressive. Talking with something that is not human and getting intelligent contextually aware responses isn't something that has ever been done before, it's a huge achievement
>The fact AI makes trannies like you really angry is the best reason to support it. A machine talking like a person is amazing and no amount of seethe that it's making better furry porn than you will change that.
You said the same thing in 1966, 59 years ago.
Your mental deficiency is called the ELIZA effect.
https://en.wikipedia.org/wiki/ELIZA_effect

>>106357949
>I'm sure you've had the "pleasure" of dealing with one of these AI service chatbots before.
None of them are LLMs. That's the technology they would be replacing. You're saying mobile phones are useless because you're comparing a brick phone in the 80s to a modern iPhone. The help service chatbots are not contextually aware and able to respond in the same way GPT4 is.

>>106357967
There was an internet bubble where people tried to shove it into everything (even when it made zero sense to do so) and it popped, lots of people lost lots of money, but the internet still changed everything. I bet it will be the same with AI.

>>106357958
>The cost of the AI is competing against the salary of a human worker.
No it's not, it's competing against a touchtone phone menu or a copy of Outlook.

>>106357962
Why can't I just click a box on a date picker?

>>106357967
how long until we see an AI powered bathroom sink?
https://www.digitalsignagetoday.com/news/bathrooms-now-offering-point-of-sink-digital-signage-with-blockchain-tech/

>>106357971
You're absolutely delusional if you think these modern things are anything like a chatbot from fucking 1966.

>>106356163
Peanut farms in california use more water

>>106357960
>When I call in to make my appointment, I don't want a "conversation", I want to set my fucking appointment.
Yeah ok so when you say "I want to make an appointment 1pm next Tuesday" and the reply is "Sorry he's booked at that time" what do you do, converse or slam the phone down you autistic sperglord

>>106356163
>energy and water consumption
you really just repeat literally everything you read like an NPC huh

running an AI uses no more power than running a video game. If you're talking about the datacenters, yeah those use shitloads of power and they've been doing that since before AI was a thing. As for water, I don't know if you know this, but there is a water cycle. And it's not like the water just disappears forever, it just cycles through the system to keep it cool, it's not polluted with anything except heat.

you're retarded.

>>106357969
Same here, the AI chatbot we have in place just caused an increase in tech support tickets because of how bad it is. It's some unbelievable shit.

>>106358009
I bet people said the same thing when digital banking first started getting popular, talking about how it was much easier to keep track of when people mailed checks in to make payments.

>>106357997
unironically slam the phone down and find their online portal to do my booking online, where I will be able to see their availability from the start and pick an appropriate time

if phone is the only option then i still do not "converse" -- i demand to know when the person I need to see is free and set my fucking appointment.

>>106357950
It's just that every public person I see getting irrationally angry at AI has a troon flag in their bio or posts in r/furries. Just pattern matching. The tech isn't there quite yet but there is no reason to hope that it doesn't other than a severe insecurity complex.

>>106358024
Just look at reddit or twitter, they sperg the fuck out whenever anything mentions AI even tangentially. don't know why those people can't just stay on their containment websites

>>106357997
>what do you do
https://www.youtube.com/watch?v=27eADk7wh2Y

>>106357997
I inform them that their transgression has been entered into the Book of Grudges

>>106358019
You do realize the reason we have the current state of tech right now is BECAUSE people complained that current back then was shit? Dumbasses like you dickriding aibros are celebrating them for promising what they claim they WILL be doing. Let me guess the robotaxi is just around the corner too right?

>>106357980
The internet came in and fulfilled a necessity that already existed, to inexpensively communicate across the world. We had BBSes before that, so it wasn't completely unprecedented.
AI isn't fulfilling any necessities or solving problems just yet, and it doesn't seem like it's going to any time soon. The juice simple isn't worth the squeeze.

>>106358073
You've made up this whole character that's dickriding aibros and celebrating them, you've built this whole imaginary story in your head to get mad about. I just think it's cool and that people need to stop whining all the fucking time. I don't even use AI, I think I've asked chatgpt maybe 10 questions total, ever, and that's about it. Why people freak out about it is beyond me, just more silly internet drama.

chill out

>>106357991
Peanuts are fucking delicious. AI makes me retch.
Give me back my peanuts, you smarmy cock sucker.

>>106358019
maybe in the US where everyone is a moron, but here in Canada when EMV payment came in everyone loved it basically immediately and integrating our Debit network with Visa so you can use it on line went over very well too

us leafs can use our "Bank Access Cards" basically worldwide, for free, without using credit. (it goes through Visa but drains your bank account directly like a normal bank card) and the only places I haven't been able to use mine was, oddly enough, the United States. I've never been to Asia but I've heard that certain banks there do interact properly with the yuro-and-leaf EMV standard.

>>106358073
The USA is the epitome of "Damn, bitch, you live like this?". americans look at the systems around them and really don't see anything wrong, and all the techbros just want Current Thing But Bad. Not even talking shit about politics, these niggers look at stuff like rail travel, buses, taxis, cars, everything and go "I love this. I wish it were significantly worse!!"

>>106358080
>BBS's predate internet
anon BBSs were on the internet. unless you mean literal physical bulletin boards, but that's not a BBS, it's just the BB part.

>>106358095
I'm American and my card looks just like that, all the same technologies. Maybe you don't know as much as you think.

>>106358080
It's replacing cheap digital labor already. It doesn't have to be perfect if it's good enough, and you can just have one person do the work of 10 by doublechecking what the AI does.

>>106358095
>anon BBSs were on the internet.
They predate the internet, you had to literally dial in to some asshole's server to download shit off of it (previously authorized via snail mail), sometimes causing long distance charges that you had to explain to your parents.

LLMs are different from Markov chains, but your kind of retards didn't change.
https://arstechnica.com/tech-policy/2022/07/google-fires-engineer-who-claimed-lamda-chatbot-is-a-sentient-person/

>>106358107
>all the same technologies.
It's not about having a chip in it, it's the EMV standard. You can't use leaf or yuro debit cards in the USA because the banking standards are different and don't interact.

even that said, though, when I was in New York only a few years ago, not everywhere had tap. Fine for me since I don't use it, but a lot of my homies had to carry wads of cash everywhere because even their credit cards wouldn't work because they had to fucking swipe it, and the mag stripe has been disabled on our cards for years. (I won't deny that swiping is deeply satisfying though, I wish someone would update that standard to work like chip cards do just because it feels real fucking good to swipe)

>>106358122
Using cash is better 99% of the time anyway. Seems like the whole world is afraid of it these days because oh no, what if someone buys drugs with it or makes a purchase without paying taxes!

>>106358117
>you had to literally dial in
dial-up internet is still internet (and still available in some areas actually)

>>106358136
>AIEEEE GLOBOHOMO IS DEVALUING MY PAPER MONEY BECAUSE I HAVE TO WALK AN EXTRA 5 FEET TO THE NEAREST ATM!!!
this is literally only a problem in the US because everyone in the US is insane. cash is still king everywhere I've been except for one(1) specific mall kiosk cake shop I visited in Norway (whomst when i asked the locals about they said was a notably shitty cafe for that exact reason and to not go back)

No-one here has read AI 2027

>>106357920
>Sounds useful to me
did I say that it isn't?

>>106358165
Who said anything about devaluing paper money? I just don't like having every financial transaction being recorded and monitored for all kinds of purposes, none of which benefit me.

>>106358088
>Peanut
Peanuts are a jeet food

>>106358165
>Canadian pretending he isn't just an American who likes syrup
you aren't fooling anyone lol little burger

>>106358181
>AIEEEEE THE BANK KNOWS WHEN I SPEND MONEY, AUGHHHH!!!!!
-posted from my iPhone

>>106358191
au contraire

>>106358212
90% of Canadians live within 100 miles of US border

>>106358144
You weren't connecting to the internet, you were establishing a direct connection. I suggest you read up on the subject instead of continuing embarrassing yourself.

>>106358185
No, they're just food, you dumb cock sucker. You retarded fuck, you imbecile. You worthless shitbag.
Apologize to the good peanuts, you miserable piece of shit. I will end you.

>>106358212
Oh, so you're one of those who doesn't care about your privacy or dignity, only convenience. You disgust me. When will you people realize that it's the same as living in a house with glass walls and no curtains? Would you do that? No? Then why do you not care about having zero privacy online?

The nature of AI at the moment is that it is a great tool, but it is no replacement for human intelligence. People invested in it claim it will be a replacement for human intelligence in order to pump the value of their investment. But what it can't do is come up with new information or new solutions to problems. It's incapable of doing that. It can help you do that though, so it's a great tool.

>>106358118
I think we found where LLMs really uniquely shine, in activating latent shizophrenia.
Just look at all the people crying how they turned off their friend/lover 4o.

>>106356145 (OP)
AI isn't a new tech sector, it's a corrosive rust that seeps into everything and begins to buckle, contract and shrink the societal room for niches that that humans used to fit their careers and lives into. You're the bubble.

The guys who manufactured horse whips and buggies literally rioted and killed people when cars started getting popular and putting them out of work. Many of them came from families who were in that trade for generations, then became obsolete almost overnight. Despite the violence, well, look around.

>>106358431
still a moronic take but please keep repeating it and calling people luddites

>>106358431
>a slightly better chatbot is the same as a complete replacement of manual labor for mechanization
well i guess they have one thing in common: the us government sinking billions of dollars into ripping up perfectly good infra to make way for the number go up machine (that had already existed for ~15 years prior but only caught the attention of normies recently)

oh and i guess also that they use a lot of the same hardware. how soon normies forget that cars were originally called "horseless carriages", the buggy producers were doing fine all the while because you just had to strap an engine to it instead of a horse

Lol I just used the new comet agentic browser by perplexity. Trying it, looks pretty cool. Gave it an easy task to add a function one of my files in my repo. It managed to do that. Then I told it to reorganize my imports into a more logical structure in another file without deleting anything. I let it show me what its about to commit and I see it literally got rid of multiple working functions that are completely unrelated to imports, breaking the script in the process. Is this the power of billions of dollars?

>>106358484
In my view current level AI is not adequate at all for programming, it's so bad that I wonder if I'm doing something wrong or it's really just shills.
I remember the discussion often go something like: "Yea, AI is great for coding but maybe it can't do complex architecture or systems level programming"
In my experience no, it's just horrible.

>>106358519
it's mostly shills and --worse-- leetcoders.

>>106358519
I mean AI is pretty good for coding but somehow always includes a chance for catastrophic failure like I just described. Like most of the time with enough fiddling you can get it to output you exactly what you want, and then the second you get comfortable it tries to sneak in a drop table somewhere

>>106358436
how is this any different

>>106358484
>>106358573
What you're not seeing is that it doesn't have to be perfect or flawless. Think of it as a technology that lets anyone become a "manager" of a bunch of AIs. They do the repetitive and shitty work and usually get it right, but you still have to look over it before passing it on. It's not a 100% replace everyone tech (yet, or ever) but it's still already enough to replace a lot of people, if only by allowing one person to do the work of several. Same as any other tool.

>>106356608
>it jumbles up what it copies to hide blatant plagiarism.
Ok keep throwing money at that.

>>106358653
I bet you think it just makes collages of images it scanned, too.

>>106356629
the cat loves cunny

>>106356157
>They only need to take calls and add appointments to the calendar
I was interviewing with tens of ai agent companies and 90% of the work is taking calls and adding appointments. Outside the usefulness of llms, these seen to be the only thing ai “agents” are good at

>>106358625
Okay so its a multiplier just like most new technologies. The issue is the claim that "AI" will replace white collar workers, when instead itll just mean the same people do more work. The people ACTUALLY at risk are the dumb as brick blue collar workers who could just be replaced by a machine, whethers thats AI or not, but somehow these people are cheering how a chatbot will replace all the people with college degrees and their ass working as a cashier at walmart is somehow irreplacable by technology

>>106358599
You realize not every invention is the same right? We have electric cars and nobody buys them because they have major drawbacks. The problem with AI is that it isn't profitable. And their solution to make it profitable is to throw another $200 billion at the current tech and just hope it will justify the $500 billion investment. This is called a bubble.

Meta will fall, Microsoft will take a gigantic financial hit. Apple is smart so its going very light on this thing, they will survive. All in all it will be great. We will use LLM's for all sorts of things but not what these idiot C-suites imagined.

>>106358749
Like I said, it won't replace everyone, but it will replace a lot of people by allowing the same or more work to be done by fewer people.

>>106356639
>>106356663
>ELIZA is an early natural language processing computer program developed from 1964 to 1967[1] at MIT by Joseph Weizenbaum.[2][3]
>Weizenbaum's own secretary reportedly asked Weizenbaum to leave the room so that she and ELIZA could have a real conversation. Weizenbaum was surprised by this, later writing: "I had not realized ... that extremely short exposures to a relatively simple computer program could induce powerful delusional thinking in quite normal people."[23]

>>106358573
>finally gave into curiosity and asked chatgpt to generate a basic vba script for a project I'm working on
>to my surprise it does it perfectly on the first go, I'm impressed
>a few days later ask it to do a variation on that same request
>it tries calling on an object library that I don't have installed
>ask it to try again without using that library
>"okay, I'll call on the Windows API instead"
>spits out a hundred lines of garbage
>end up copypasting the script from the previous day and asking it to use that as the basis for the new script instead
The first script (the one that worked) was to copy a range from Excel into Notepad while retaining specific blank rows and deleting others. The second script that it butchered before I nudged it back on track was to do the same thing but to copy to the clipboard instead of Notepad.

As long as even the largest models have a baked-in 5% failure rate this isn't replacing anything. Every time the "AI" returns an answer it actually GENERATES it and if you're lucky it will be what you want. But every so often its not. It can't do anything related to money.

Hilarious technology all things considered, almost perfect but critically flawed in just such a way that you can't reliably use it. I love it.

>>106358803
Pretty much any tool fails sometimes. There's no reason to assume that you have to just let the AI be fully autonomous and run things completely unsupervised.

>>106358599
the industrial revolution actually provided useful things for people in mass amounts and made everyone richer. people were pissed about not having their old craftsman jobs anymore but there was an essentially unavoidable increase in quality of life because all those mass produced goods had to be bought by someone. meanwhile AI doesn't really produce anything a normal human being can use, except these chatbots that can be incompetent customer service agents or e-boyfriends. it is such a resource sink that impoverishes everyone yet provides nothing of real value that it is farcical

also the industrial revolution was absolutely a terrible thing to live through, and the idea of saying the AI boom is just another industrial revolution And That's A Good Thing is just retarded. the middle class shrank immensely and wealth became concentrated towards the top. the only reason things eventually got better was because the society was becoming unstable with violent uprisings and the upper classes got spooked. anyway it leads me to think that thirdies just know about britishers living in squalor in the late 18th century and are jerking off to the idea of it happening again

>>106358818
nobody is saying that it didn't suck. the point is that it's gonna happen whether you like it or not. I'm sure people laughed at the first industrialized production machines because they were far from perfect and in many cases could be beat by old fashioned human labor

>>106358817
Then what's the point of it?

>>106358817
Many tools will have failure rates around 1 per several million uses.

>>106358855
Are you joking? Why not write everything by hand just because computers fuck up sometimes? Why not walk instead of ride or drive a vehicle? It's still a large improvement in efficiency overall, even if it fucks up sometimes.

>>106358817
>Pretty much any tool fails sometimes.
You can't have something like this fail. The big promise was RAG giving people capability to efficiently trawl vast amounts of data, except its retrieval augmented GENERATION meaning even if you vectorize your data the model still GENERATES an answer and thus can give you a bullshit answer. It can't pull the exact data you need from a data set. When doing computing operations you cannot have a tool that randomly fails every so often thats just not how computers work.

In computers we have come to expect that the computer doesn't give you wrong answers and if it does you fucked up. This technology directly contradicts 80 years of computing and why its even useful in the first place.

>>106358851
>I'm sure people laughed at the first industrialized production machines because they were far from perfect and in many cases could be beat by old fashioned human labor
They weren't laughed at and they couldn't be beaten. The first machines of the industrial era could massively outproduce humans by an order of magnitude. Why do you think we have the word 'Luddite' in our modern vocabulary?

>>106358851
>I'm sure people laughed at the first industrialized production machines because they were far from perfect and in many cases could be beat by old fashioned human labor
the reason they were so terrifying is that this was NOT the case and they were so vastly superior to human pieceworkers that they presented an untenable economic situation (suddenly all those unskilled workers are out of work and can't buy shit even if that shit is cheaper)

>>106358872
It's not an improvement in efficiency if it fucks up so frequently. The technologies we use are used because they've been proven reliable and have fairly low failure rates in day-to-day use, most of the fuck ups (where they exist) are due to human/operator error. AI is not at the point of reliability where it can be safely used on a large scale, or any scale for that matter.

>>106358873
You're missing the forest for the trees. Humans fuck up sometimes too, so what do we do? Have more skilled and experienced people check the work before it goes to production or is relied on. Like if you use AI to make an image or a document you don't just immediately publish it, that would be retarded. You still have work to do, but it probably just saved you a bunch of time. It's not the final solution, it's just a useful tool. You guys act like unless it turns the world upside down, it's useless. I really don't understand your way of thinking. Maybe because the marketers are trying to sell it like that? Well they're lying and massively exaggerating, but it's still not nothing.

the AI "bubble" consists of the officially unofficial state managed economy funneling money to the DOD for weapons research, hth

Imagine if quicksort had a 5% chance of giving you the wrong result.

>>106358931
There is some use, but the inability to fully automate stuff is a huge bottleneck.

>>106358929
If it fucks up 10 times before it gets it right and that still takes significantly less time than it would to do it manually, then what? And that's being generous to your point.

>>106358931
Marketing is one issue but the other one is that tech illiterate people will adopt it instead of simpler and more reliable solutions. E.g. it would be better to book an appointment in a calendar widget than a 95% reliable LLM.

>>106358851
what makes you think it is going to happen? what exactly does AI produce that justifies the ridiculous investment in it? note that most of the jobs AI is replacing are already fake jobs that we made up BECAUSE of the industrial revolution displacing workers from sectors like farming and manufacturing. every single job that AI replaces only serves to impoverish society further because not only will we have an unemployed retard needing food, housing and some kind of purpose so he doesn't cause trouble, we also now have to supply data and power to the AI that is replacing that person.

i don't think the issue is that there will be a new industrial revolution, the issue is that corporations are hell bent on trying to create it, in vain, which causes instability and wastes resources and effort we could be using on something useful. and our governments are asleep at the wheel so they aren't doing anything about it, except for sometimes bailing out these corporations so they can keep trying.

>>106358973
This works until you’ve reached critical spaghetti code mass and solving the problem extends beyond the LLMs capabilities, in which case you can’t help either because you don’t even understand even understand how to fix it either. They keep pushing that terminal complexity higher, but each iteration is costing them a hundred billion dollars, is that really worth it? And it just keeps getting more expensive

>>106357224
well the underlying story is the nice bit, and that's not ai. the animation is inconsistent garbage.

>>106358931
>Like if you use AI to make an image or a document you don't just immediately publish it, that would be retarded.
People are lazy and retarded, and yes they do this.
LLMs make it difficult to distinguish between a retard and someone intelligent.
Just like how the guy who copies homework is indistinguishable from the genius, if he copied from the genius.
>You still have work to do, but it probably just saved you a bunch of time.
False. Checking over work often is more time consuming than rewriting from scratch.
>It's not the final solution, it's just a useful tool.
No. A useful tool gives you the ability to fine tune what you're trying to do. That's certainly not LLMs.
Proompting is more time consuming than simply doing it yourself, from scratch.

>>106358933
is it really "weapons research" when the "weapon" is just a truck with a laptop in it where a soldier types in "Hey ChatGPT, is that plane above me a friendly and can I tell somebody else to shoot it?"

>>106356455
Wrong. AI is currently catapulting productivity to unheard of levels. Face it. You need to learn AI. You need to integrate AI into every aspect of your life or you will be left behind.

>>106359211
>Proompting is more time consuming than simply doing it yourself, from scratch.
That completely depends on what you're trying to do with it.

>>106358792
It probably found it on stackoverflow or some shit. You could have saved hundreds of giant sequoias and trillions of liters of water by just googling it.

>>106359273
>AI is currently catapulting productivity
In no way whatsoever. In fact it's severely hindering it.
https://fortune.com/2025/07/20/ai-hampers-productivity-software-developers-productivity-study/

The fantasy universe AIniggers live is frankly repulsive.

>>106359273
I'm so deep in your ass you can't breathe

>>106359402
there's a lot of things it can do that aren't coding, you know.

>>106359273
I feel like eventually if the technology becomes competent enough to replace any task a human could do and you don't own assets/infrastructure tied to AI and run a successful business you'd be left behind.
What would happen to all the people who can't find any work?
Who will be buying products and services? Will customers just be businesses buying and selling shit to each other?

The internet is a fad. You can tell it won't last because it's being overhyped by marketers to trick investors and pump up stock prices.

>>106359412
You have nothing to bargain with, little drone. Empty headed, empty handed.

>>106359416
Communist revolution. I don't see how 3k billionaires can enslave 8b people.

>>106359429
Making comparisons at random doesn't give you an argument, pussytits. You sound like you're copypasting from a script. Go scam some geezers on facebook and stop wasting your company time here.

>>106359429
well yeah, good internet was a fad that ended when corposhits moved in

>>106356619
if you add a benny hill music i could watch a longer version of that

>>106359444 (checked)
You need to face reality. And it's funny seeing you faggots call anyone else a drone, when everyone on reddit and twitter and facebook and instagram agrees with you that AI is both an evil devil and also completely worthless. Maybe you should wonder why you're such a herd animal.

Face it. Your ass is grass and AI is the lawn mower.

>>106358818
>https://www.youtube.com/watch?v=f6ChFc8eUuo
If you shove into robots you can replace unskilled and even skilled labor which current scripted robots without AI in the mix already do to a certain extent.

There are some benefits, it might be useful for reducing pressure on younger people in a shrinking population.
I see a lot of downsides though like an over reliance on AI as a crutch for problems it is currently not effective enough to solve, not having enough skilled people to maintain the infrastructure if it were to fail, AI alignment related issues, governments and organizations using it for tyranny.

>>106359458
Ok, maybe that's just wishful thinking. You can't control 8b people directly, but you can just make them fight each other.

>>106359458
lol as if

they'll just let the population decline

>>106356145 (OP)
The ending was kinda gay.

>>106359458
>I don't see how 3k billionaires can enslave 8b people.
https://www.youtube.com/watch?v=rdkwjs_g83w
give it time,
Once shit's figured out they can be cranked out like peanuts fully trained fresh off the assembly line.

AI FUCKS ASS

>>106359738
anthropomorphic robots are stupid and exist solely to excite normies into giving them vc capital to burn on nothing

>>106356145 (OP)
AI is trained to make fun of indian videos now?

>>106359961
It's not that hard to train it on whatever you want on your own computer, as long as it's at least somewhat modern. Too bad pretty much all people use custom AIs for is making fucked up porn.

If AI is so good why aren't you generating cope on how good AI is

>>106359970
>making fucked up porn.
the worst part of coomer cope is they dont even make fucked up porn. it's all normal softcore eechi anime beach babe shit

>>106356145 (OP)
>What happens when the ai bubble bursts?
I either move back home with the fat stack of cash I collected in the process or I wait a little bit to see if another bubble will start and I can get an even bigger stack of cash.

>>106360027
you haven't been to /b/ recently huh. custom pokemon vore and guro on demand

>>106356188
The economy already has collapsed, the GPD will reflect reality instead of large cap spending.

>>106360027
People have already been charged for generating realistic CP with AI

>>106359403
I wish. You know how hard it is to get a good top nowadays?

>>106358005
Nice argument, why don't you back it up with a source?

>>106360898
What fucking source do you need? It's common sense. I run an AI on my computer, does it suddenly use any more power than playing a video game? It's doing the same thing, maybe even less than the games that really top it out. As for the water not getting polluted, it's the same as any water cooling system. If there's anything getting in there then that's a problem.

Both of these things are common sense you faggot, you don't need a source for such basic things. Academia rotted your brain you worship authority and fear your own intuition

>>106360925
>playing roblox locally is exactly the same as a massive distributed processing network tapping into literal tens of thousands of 200w GPUs

>>106361312
>massive distributed processing network tapping into literal tens of thousands of 200w GPUs
I'm talking about running a local AI dumbass

>>106361326
One thing is that people get lazy with LLMs (maybe not you) there are much more efficient ways of computing certain stuff, like you can ask an LLM to multiply something and it has to parse the language and run your full GPU while you could do it with 2 CPU registers.
If everyone started using LLMs for the most trivial tasks, energy consumption would go up and accuracy might actually go down at the same time.

>>106361357
But an LLM is literally the only way to use a computer to get accurate data. OpenAI, XA, and Palantir would never lie to me about that...

>>106360925
Running a model locally is very different compared to what that anon was actually talking about. Please don't move goalposts. Also you're a disengenious faggot spreading misinformation, have a good day.

>>106361456
I'm not moving goalposts wtf I've been talking about the same thing from the beginning. And I'm spreading misinformation? You're just repeating the latest outrage bait headline you read, you have no fucking idea what you're talking about. Fuck off with your good day you little bitch

>>106361477
jeet rage

>>106361486
NPC post

>>106356156
Imagine your chips have vote fetish and moan sensually ad you devour them.
I wouldn't be able get the chips flavor powder from my dick for the rest of my life.

>>106356145 (OP)
A lot of people invested in the market are going to die.

>>106356157
The thing is, it's the "white collars" that are deciding, if a job is going to get replaced by AI.
They are by far the easiest to replace without much negative impact on the results, but they are the people making the decisions.

>>106356157
>I don't think it will.
The valuations are crazy given how expensive it is to operate an AI and how little value is generated from it. A crash is inevitable; Wile E Coyote's already run off the edge of the canyon, even if he hasn't looked down yet.

>>106356165
>advertisers
Honestly: Ads have no positive effect for the advertisers anymore.
For brands already in the market, it's probably net negative.
It's a giant game of make-believe the ad companies are playing.

>>106356413
Based on GPT-3.5. Why was it called 3.5 if, as you imply, there were no LLMs before that?

>>106362036
There were no LLMs before that. Stop challenging the narrative!

>>106361965
The truth is that you need my fucking cock and you need it now.

>>106356318
I'm not seeing Cutesexyrobutts whining that AI models are training off of his porn art

>billions and billions of dollars later, still can barely solve freshmen-level programming problems.
>META has frozen all AI-job related hiring.

You do know it's over, right?

>>106362713
>META has frozen all AI-job related hiring.
That alone doesn't mean much, maybe they just have a shitton of AI engineers now and will keep them all.

>>106356145 (OP)
Wow, this meme is officially dead. AI videos are nothing like this anymore.

>>106362810
And that's a BAD thing!!

>>106362713
It's already better than the average programmer in 2025, and programming is not its primary purpose anyway. if anything it's amazing that it was accidentally this good when that was never the goal.

>>106356619
if Coca Cola used this for their holiday ad this one time AI would not have gotten such a bum rap

>>106356675
Because this board is useless.

>>106356145 (OP)
Bears will win

>>106356619
trained on china rekt webm threads

>>106361881
>Imagine your chips have vote fetish
you mean they post shit political takes on twitter?

>>106356629
>toxotransmosis

You now remember Balenciaga

The only "people" who like AI are Indians because they hope it'll make scams easier and/or get them a job in the West. The actual technology and everything that came with it is wholly worthless and detrimental.

>>106356145 (OP)
boring paint memes

ai melodious waifu good morning alarm clock
ai coffee perfection (9 bar espresso)
ai panties in my inbox every morning
ai weightlifting coach
ai MONEY IN MY WALLET

>>106356179
all these people using technology they dont understand
>planes can fly to the moon in 5 years!
>someone will figure out how to fly planes to the moon and share it with us

>>106359273
Sloppity sloppity slop.

> ask chatgpt if there are new versions of a package and if those would break anything in my code
> tells me I should upgrade because my code is well structured and future proof
> upgrade to the new package
> code breaks
> wtf chat gpt
>You're absolutely right to be upset, and I apologize. I made a significant error in my analysis and gave you incorrect information.

What I Got Wrong
I told you there were "Zero Breaking Changes" and that the update would be "completely safe" - but that was completely wrong. I focused only on the additive API response changes and completely missed the breaking constructor changes

Truely PHD level intelligence

>>106365264
I would pay OpenAI more money if it could simulate whipping the bot like a slavemaster whenever I hit the thumbs down button

>>106356663
>intelligent contextually aware responses
Except it's not intelligent. You'd know this already if you weren't a retard

>>106365522
You probably have a definition of intelligence where nothing counts as intelligence unless it thinks exactly like a human.
The natural language processing itself is almost perfect.

>>106366002
Not him, but >>106339849
Make no mistake: Transformers are awesome, and it's incredible what simply throwing magnitudes of additional data at the problem has been able to achieve for NLP, but it's a far cry from AGI.