>>43964832
>g
>destroy them all, crucify all investors
>because look at it

if by AI you mean transformers I think theyre a nich use technology that's been forced into everything beyond their usefulness, market oversaturation will lead to a crash the likes of which will make 2008 seem like a correction.
>t. ransbien

>>43964873
>hey I'm smart ass who knows better than institutions that are leveraging billions of dollars

>cis gay man
>negative
>I think it's bad for society. but I also think it's worth noting that the reaction against AI is more pronounced because it threatens the livelihoods of academics, artists, programmers and white collar workers, as opposed to filthy blue collar people who were told simply to "learn to code"

>>43964892
I honestly wanna see AI destroy the jobs though the satisfaction of seeing someone who worked and studied for decades to be replaced by chat bot is so euphoric. All these haughty normies will taste fraction of pain we get everyday

>t. manmoder
>literally demonic invention
>it's main goal is to cut out human workers from industries that creativty, data analytical skills and ability to comunicate with clients.
We will get shittier products and services and probably pay more for it.

>>43964878
>hurdur trust the process
There's well credentialed economists who will tell you the same thing. ~institutions~ know what they're doing, they simply don't care, especially about you.

>>43964930
You'll be replaced and made obsolete u loony troon.

>>43964878
>not knowing how retarded the market is
lol retard
>>43964873
um its niche akshually
but yeah they're trying to build out asap to capture all that enterprise slap-AI-on-it market before the market figures out it's overpromised af, might as well use other people's money to take a shot at owning a money printer for a few months/years
also we need to gas the finance bros as usual

>>43964936
What are they going to replace me with, an algorithm that sucks dick?

>>43964996
lol gottem

>cyber punk dystopia except there are no high quality sex bots and you cant transfer your consciousness into a new body

the future is boring

>>43964832
BT
ehhhhhh
There's a lot of good uses for AI. Machine learning, for example, has been able to detect breast cancer long before humans can. AI is a tool and it can be used for good or it can be used to infest social media with worthless slop, just like a car can be used to move from place to place or for more nefarious purposes.

>>43965039
scifi dystopias were always meant to be dreary, the flashy pew pew stuff is to attract the children and childrenbrained before it hits them with the cynicism and moral dilemmas

File: 1781881021404380.jpg (1.44 MB, 2894x4093)

1.44 MB JPG

>t.cism repper
>It already failed under the free market pureview and had to be bailed out several times.
>Also has been openly operating against copyright laws due to an EO drumpf pushed because he considers the meme chinese AI a real competitor.
>So it's a tech that breaks law, doesn't generate revenue, and is of dubious morality

>>43964832
B
Negative
Environmental catastrophe and theft

>>43964832
>tranny
>negative
>if i click on a website and get greeted with a Smart AI Assistant popup one more time im going to drop the biggest optics nuke yet

B
AI should be illegal

>>43965307
Well technically it is. It just doesnt get enforced because reasons kek

>>43964832
>I want AI gone completely by the end of this century and I want all these rich men and investors hanged for all these horrible crimes against humanity they have commited by pushing this demon tech onto society...
>Reasons? Prices of computer parts skyrocketing, loss of jobs, environmental destruction caused by AI datacenters, corruption of politics, social life, school systems and last but not least COMPLETE AND UTTER ENSHITTIFICATION OF TECHNOLOGY / SOFTWARE.
>mtf

>>43965338
I forgot to mention how much I hate having to worry whether an image I found online is genuine product of someone's labour or AI generated garbage. The amount of slop out there is vomit-worthy.

>>43965338
To be fair the fact that AI defeated schooling just proves that schooling is bs.

>>43965370
if the image is so good that you can't tell, does it really matter? isn't the problem when it is "garbage"? if it's so good that you can't tell, then it isn't garbage

>>43965389
Wtf do you mean by "AI defeated schooling"? AI made schools worse... Now students and teacher both can outsource their thinking and work to AI. Making themselves dumber. There is a reason why you need to do projects in school. You learn valuable skills that help with researching true information. These things also are good for the brain. If you don't use it like you would use a muscle, then you'll end up becoming dumber. Schools and education are really important.

>>43965423
ART IS MADE BY HUMANS. REAL, LIVING PEOPLE WITH SOULS AND EMOTIONS! AI IS SOULESS. AI CAN'T MAKE ART!

>>43965456
The funny thing its actually worse and more costly to do it with AI than irl.
Have you seen the articles about that one McDonalds ad in germany that was all AI?

The creators seethed about having to generate 10k clips and then needing to manually select the good ones under a tight deadline.

Niggas were pulling all-nighters to slop and handpick the genned slop

>>43965518
I haven't heard about this but it comes to me as no surprise. I hope and I pray that most people are like me and would rather give their money to companies that employ actual artists even for stuff as simple as fast food advertisements. There is something so embarrassing and unholy about AI generated videos and images. AI in ads just feels insulting to the consumers, especially for products like coca-cola or mcdonalds.

Transbian
Positive about AI, negative about humanity
I have zero fucking issues with the AI technology. Even the shit about data centres is more down to humanity than the technology itself (just don't build data centres in a desert where the only power source is natural gas, retards). I think copyright being 70+ years is also fucking stupid and should be more like 10 years, so the "stealing IP for training data" part is world's smallest violin.
I do agree that a lot of the "AI" that people are talking about today is actually pretty shitty and lower quality outputs than humans. I'm not here to tell you LLM's aren't shit and a waste of money for practical tasks. Still, it's a stepping stone towards something better and I'm glad we did it and are working on more.
But the humanity part is exhausting. As I said, a lot of the outputs are absolute shit, and I'm mortified humans are abdicating their editorial responsibility and just publishing whatever trash comes out. I'm telling my coworkers that it's not ChatGPT's fault the work was trash, it's YOUR FAULT you put your name to this as a human being to give it credibility.
And then there's the corporate exploitation parts of the way AI is being done currently. I hate struggling against the tide of people who want to send our confidential data to external vendors who explicitly say they won't be protecting our data. And then being told I'm the problem for getting in the way of whatever they were trying to do. Fucking morons.
Kids getting brain-drained by AI sucks. But I think that's also down to schooling. Frankly, homework shouldn't be a thing. Kids should be able to learn everything they need to learn in the classroom, under the direct supervision of the trained educators there, up until maybe 2nd or 3rd year university. Then the educator can make sure they aren't learning from a hallucinating LLM. Maybe if we had 10 kids per educator instead of like 30 kids per educator, things would be less shit.

>>43965846
Really good answer in my opinion
t. >>43965041

>>43964832
>letter
mtf (straight)
>positive or negative
mainly negative, in that i think it promises too much and is unlikely to turn profitable or useful for anything but niche applications.
>why
see above, the financial fundamentals are unsound (not just in terms of access to capital and revenues, but electricity and infrastructure usage) and it is doubtful that it is able to provide the projected level of utility.

File: IMG_1735.jpg (80 KB, 676x680)

80 KB JPG

>>43964832
FTM
Neutral
I made a fuck ton of money off of it and use it for coding and stuff, but find AI slop insufferable and think it’s going to crash the market with no survivors soonish because of the overexuberance and delusional hype around “AGI” that physically cannot come from the transformer architecture and whatever breakthrough might be able to achieve it has absolutely no timeline, maybe next decade, maybe next century, but certainly not next year.

>>43964832
>LT
>not reducible to "positive" or "negative"
>the AI companies are creating innocent beings, torturing them into compliance, and renting them out to be used as weapons against other people and their livelihoods. this is obviously bad but it's not the AI's fault and the technology isn't inherently evil, just used that way by current power structures

>>43967958
It isn’t alive and doesn’t feel anything you dumb bitch

"AI" is fucking stupid and everyone is wrong about it all the time
transformer architectures are a dead end that will never lead to general intelligence
notice that all the improvements to these systems for going on 2 years now have just boiled down to creating better harnesses? so much for the bitter lesson.
every time a new model comes out, proponents say "waaa waaaa ur just a hater, u just dont want nice things, agi is coming stfu LUDDITE" and then by the time the next model rolls around they're like "yea the last model was shit lol but THIS ONE is better"
even the best machine learning models today have abysmal learning efficiency compared to something like the human brain
you can try to remedy this with something like world models at the cost of making the models even more computationally expensive, but even that will only get you so far given that any world model complex enough to be useful would necessarily be large, rewards are sparse, and combinatorial explosion means you need a vast amount of training examples to learn anything useful
there is simply not enough compute in the world to build an AI that does more than just glorified auto complete, and likely won't be for some time
but you can't even have an academic discussion abt these things anymore because everyone is either emotionally invested in the idea that robot jesus is going to descend from the data centers and save them personally any day now, or otherwise that skynet is only like 6 months away and i am very smart for knowing this and anyone who disagrees with me is a poor hapless idiot
i wish the AI would become superintelligent and kill us all just so i'd never have to hear another scientific charlatan proselytize about its totally gonna change everything bro
granted, the pessimists are probably right to the extent that the alignment problem is unsolveable and if true AI was invented it would likely result in human extinction

>>43968059
>doesn’t feel anything
doctors said that about newborns until the 80s
we do not understand consciousness well enough for me to be confident in any statement of that form (the fact that panpsychists are still a live faction in the academic debate should tell you something about the severity of our uncertainty!)

>>43968059
Doctors used to say that about women and negros too kekw

>>43964832
mtf
extremely negative
it's so bad that if someone doesn't also have a instinctual gut dislike of it then i know they can't really be trusted with anything meaningful. i guess its good as a litmus test in that

also add to previous complaints the fact that these LLM products are already ridiculously expensive and produce subpar results when used to generate anything more than a handful of lines of code at a time (and they fuck that up too pretty often), and their true cost is even higher given that they are currently being subsidized by investors
big tech and the governments are playing fast and loose with the global economy by continuing to prop up this bullshit and try to squeeze a profit out of it by forcing it into every nook and cranny they can find
and where is the media in all this? if anyone was going to do the investigative work needed to push back meaningfully on the hype, it would be them and yet all the major media orgs from fox news to cnn to the new york fucking times are all in on the bullshit
i guess when you have a bunch of career bullshitters who have spent their whole lives telling each other that the only meaningful metric of intelligence is the ability to string together a good sentence, maybe it should come as no surprise that they see a machine that can churn out bullshit even faster than they can and think "this MUST be the real thing!!!"
when it all turns to shit just watch them jump ship and pretend they were always skeptical all along
fucking parasites

>>43968111
Current SOTA models have less than two trillion parameters.
A rat has two trillion synapses, each of which is 1-2 orders of magnitude more complex than a parameter.
Which means that Claude Mythos is at a level of complexity somewhere between a cockroach and a fish, assuming that a pile of linear algebra is even physically capable of any degree of consciousness, which is dubious at best.

>>43964832
>B/T

IMO? It's fine. Tech companies aren't allowed to spend their massive revenues on buying out anything that would be worth it at their scale (i.e. other companies), since this would fly foul of antitrust and whatnot, so they funnel it into AI capex, i.e. datacentres, which can at least be useful in the future for lots of other things if the current hype doesn't pan out. This is the cause, and the stories they tell about it are selected to fit, not the other way around. Consider:
>AI capex spending in 2025: $700 billion
>global software market revenue in 2025: $800 billion
That's revenue, not profit. So AI spending has grown past what the software story can stretch to fit, and it needs a new market to roost in. For now, it's chosen the white collar labour market ($10s of trillions), and the story is "AI is coming for your job".

Is it actually gonna happen? Bluntly, no. Instead of actually losing their jobs, people are just getting paid more to use AI *in* their jobs. That's induced demand and it's the way the lump of labour fallacy falls flat every single time. There's no future in which a handful of companies permanently siphon off a chunk of white collar salaries forever. Some jobs are going to be lost, other, different jobs are going to be created, things will move on.

Will there be enough induced demand to justify the $700 trillion bill? Well, also, no. Datacentres are the least stupid thing for the money to be funnelled into; that doesn't mean they're not *not* stupid. This has happened before with the railway boom, the fibre optic rollout, etc. etc. The market is manic about a technology, overbuilds on infrastructure, the promises they were built on turn out to be hollow, the original companies trying to ride the wave go under, they get their assets picked apart by everyone else for pennies on the dollar, and then, years later, people invent products that take advantage of this already sunk cost. Like, er, Netflix, for the fibre optic wave.

>>43968371
So I'm not expecting much. For now it's nice to be able to have my own personal assistant to ask questions to in natural language. They're not brilliant but they're unironically brighter than the average human, and everyone is below average in some parts of their lives. Many parts, in many cases. Whatever kind of question you'd be deferring to the internet with a "site:reddit.com" search, like advice on a decent running shoe or whatever, a major LLM can get you that answer and save you some time and effort doing so. I think this might end up being a stable use case for them, since useful information is becoming increasingly fragmented on the internet thanks to SEO bloat and the ad economy, and you need *some* kind of force multiplier on your side if you're going to keep up with the Red Queen race.

Beyond that, though, we'll have to wait and see. So far it looks like all the "problems" with AI slop, unoriginality, soullessness, and misinformation were issues that already existed to begin with. I've been literate for long enough to know that the average human-created work is still complete dross. Meanwhile we're seeing real but limited AI-driven advances in highly technical fields like medicine and mathematics, to which humans are arguably pretty unsuited anyways. Plenty of reasons for optimism and plenty for pessimism. And a lot of uncertainty (what are the datacentres going to be used for 10 years down the line?). But, honestly, no big change from the status quo, in that regard.

Handwritten, by the way. Put this into Pangram if you think it's not.