File: IMG_1633.png (523.8 KB)
Unless you’re specifically concerned about
critical issues like job security or cognitive
decline, your reaction to advancements in artificial intelligence is ultimately a humility test.
If you’re excited to see how sophisticated and capable AI can possible get, you passed.
If you’re malding over your skill/knowledge
potentially becoming more accessible or
less relevant, you failed.
Generally speaking, truly intelligent people can
see the futility of pride.
Showing all 113 replies.
>>
>>16984536
>If you’re excited to see how sophisticated and capable AI can possible get, you passed.
I would have been, but the fun of it is spoiled by malicious parasites trying to inject AI slop into everything and unleashing mindless hordes of human replacement fetishists to sing their praises everywhere all the time.
>>
>>
>If you’re malding over your skill/knowledge potentially becoming more accessible or less relevant, you failed.
But that's directly tied to job security. I think people who see math being solved and go "omg i'm trembling with fear rn" are really concerned about being thrown to the streets.
>>
>>16984536
I'm most acutely concerned with AI turbocharged malware bringing down the critical infrastructure that keeps electrons and water flowing. Also, if the overall cognitive decline of humanity isn't an additional apocalypse level threat to you, you haven't adequately thought through the direct impacts and second order effects.
>>
>>
>>
>>
>>16985270
>>16985327
People on this board are primarily concerned with bragging rights. If you state otherwise, you’re being dishonest.
>>16985340
The statement stands on its own merit.
>>
>>
>>
>>16984536
Wrong. Only talentless lazy diots or the people making money off of AI are excited.
Talentless idiots are happy because their talentless lives and laziness to develop actual skills may soon be validated by having the plagiarism machine steal the hardwork of more talanted and hardworking humans.
You seem like you're happy about it, so I'll just assume youre someone with very little achievement irl, and so are happy you're incompetence will soon be rewarded by AI
>>
>>16985377
You're not understanding.
People who develop skills for the sake of
enjoyment have no reason to feel negatively
about advancements. Most of them just think
it's cool they figured out how to make a computer
do it.
Also, how the premise "wrong?"
There are plenty of skilled people who haven't
a monetary incentive to like AI who are excited.
Look at how many vulnerability researchers
reacted to Mythos finding zero days.
(ignoring the safety implications)
Or how high ranking chess players react to chess bots.
>>
>>16985383
No skilled person is happy for AI advancements, theyre just accepting it.
If theyre happy, they werent skilled.
Those that developed skills for the sake of developing skills are obviously not happy dumb people are now stealing their skills, unless theyre autistic.
In fact, online its either AI workers, former crypto scammers, and pajeets who love AI. Youre probably a subset of those three, arent you?
>>
>>16985383
Security researchers that were all paid for by the same investord as anthropic? How fucking retarded can you be?
Also, literally best Go player retired after alphago, and Im a chess GM, and everyone around me hate chess bots since it ruins the game.
Any famous person can obviously be bought, so their statements are meaningless noise. Youd realize that if you had any bit of success in any field and see whats its like at the top. When you have publications at good journals, you get tons of emails asking you to use your name in a ready made paper to support their products. Seeing as you don't seem to know this, tells me you dont have a top journal publication.
Youre not humble, you just dont have anything to be proud of
>>
>>
>>16984536
i dont like the government giving them giant breaks in EVERYTHING. cheaper electricity, tax breaks, cheaper water, etc etc. i understand giving them some basic things like tax breaks for X years or helping expedite some bullshit government permits, but giving them everything is gay.
also, driving up muh computer parts kinda sucks.
>>
>>
>>16985377
Let me guess, you also believe we ought to destroy automated factories so we can "make more jobs," right?
Decreasing the labor necessary to produce goods increases economic efficiency and, therefore, the effective supply of labor. It also allows individuals to spend less for the same service, making things cheaper and freeing up time and resources to be better used elsewhere.
Ultimately, this is good for both the market at large and the consumer, and the only individuals who would be opposed to such a reduction in necessary labor would be those selfish neo-luddites who want to end innovation and halt the meeting of the ends of man.
>>
>>
>>
>>
>>
>>16985377
>>16985395
>>16985494
All OP said is that you’re not humble. What are you rambling about?
>>
>>16985395
>dumb people are now stealing their skills
You can't steal an idea, as ideas aren't scarce goods (no ownership).
If someone is using what's cost-efficient and available to them (AI) instead of wasting resources, then you can't really call them "dumb" for that.
>>
>>
>>
>>16985377
>more talanted and hardworking humans.
I'm really talented (world's best) and hardworking (>30,000 hours) at creating so-called "useless" tools from the Stone Age, but those evil "modern technology" bros are stealing my job! >:(
>>
>>
>>
>>
>>16985451
Lack of jobs is entirely valid concern because even though automation is destroying jobs our actual economic model is built around people having jobs. It doesn't matter if goods are getting cheaper if you get fired and have no money. Even if you are personally insulated by the effects for instance being in the owner class or having a job that isn't easily replaced or even going up in value thanks to AI your quality of life will stiff drop when crime increases and unrest grows as the breadlines grow.
>>
>>
File: ted-ideas.jpg (57.6 KB)
>>16985572
>AI is destroying our heckin' economy!
Good.
>>
>>16985572
Ultimately, what you want will end up stifling economic development just to take money from people who rely on AI to put money into the hands of people whose desires must be more valuable to you.
>our actual economic model is built around people having jobs
New jobs will be created by AI so long as people can create value. Any short-term desires those whose jobs are better done by AI have will be alleviated in the long run.
I mean, would you rather live as a millionaire in the 1800s or middle-class today?
The only case in which jobs would be lost is when workers aren't able to provide value to their customers, in such a case why should we force people to pay for an inferior service (assuming you are pro-government restriction on the right to use AI for producing goods)?
This is all at the expense of the consumer and those who work using AI.
>>
File: brainlet3.jpg (78.8 KB)
>>16985602
>t.
Other golems will argue with this one unironically.
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>16985602
>New jobs will be created by AI so long as people can create value.
No they won't because AI fundamentally replaces people, new jobs will be done by AI as well and as AI gets better it replaces new jobs faster than they are "created". People found new jobs when people replaces horses with cars. Horses didn't find new jobs and got culled by the millions.
>I mean, would you rather live as a millionaire in the 1800s or middle-class today?
False comparison because without a job you aren't a millionaire or middle class. You will be poor and starving.
>force people to pay
People won't have money to pay for anything without a job. They will be stealing what you make.
>>
>>16985572
>your quality of life will stiff drop when crime increases and unrest grows as the breadlines grow
That definitely wouldn't happen, but even if it did, poverty and economic recession correlate negatively (~-.6) with all sorts of crime (yes, including property crime).
https://archive.fo/S4iyK
>>
>>16985640
Poverty from job losses does cause crime, the difference is that real life poverty today is almost always systemic and slow and the most aggressive people usually just move out but you can't "move out" from under the AI. The arch typical places that get poorer are places like old mining towns where the mine closes and the miners move away leaving only old people to inhabit the area, old people don't do crimes at nearly the rate than working age people do. These things also usually take decades with the first effect being that the mine stops hiring new people, the old miners stay for a long time but there's no "new miners" just weeks before the mine closes which means young people have migrated away from these areas for years before the primary economic engine even shuts down in the first place.
Real starvation always causes real crime, people don't just go quietly into the night.
>>
People objecting AI are the same people who objected the invention of the wheel, script, steam engine, electricity, computers & any other technology that boosted human capability.
Btw, the West conquered half the world essentially because of technology that others lacked.
>>
Bro the elevator operator didn't become homeless you know, he switched to factory work, then to retail, then to data entry, then to warehouse logistics, now to data annotator or AI trainer.
If there's no need for us to do a job, it'll be replaced by robots or completely gone.
But there will always be jobs for humans.
If you're not flexible or willing to develop yourself and learn sth new adn adapt to evolution, that's your problem.
>>
>>
File: non-player-retard.jpg (17.5 KB)
>>16985820
>People objecting AI are the same people who objected the invention of the wheel
It's fun to watch this 5 years old corporate marketing talking point gradually devolving into more and more cretinous forms.
>>
>>
>>
If it's possible to take advantage of the efficiency provided by using AI to do your work, without simultaneously getting brain damage by the cognitive offloading and atrophe of your faculties, I'd be ok with that, but I don't see how it's possible. The brain prunes dendrites that don't get used. The body downregulates the production of neurotransmitters that you flood it with externally. Your vessel will seek homeostasis no matter what. It is both a blessing and a curse. What will you do? Retain your hard skills but be left behind? Or ride the machine's coattails at the virtual cost of your frontal lobe? If there's a way to get the best of both worlds, please share!
>>
>>16985383
>>16985400
It's like when a new meta ruins the fun of a multiplayer game. As AI gets better and more jobs become simply let the AI do it's thing or just do what your robo overlords tell you it gets less fun. That's all there is to it.
Chess bots made the game of chess less fun as well for many people, while people still enjoy it the top level isn't as fun as it used to be since the bots basically just tell you what's good or not.
>>
>>16984536
Oh boy! I can't wait to eat slop made by a lazy idiot who knows nothing on the internet! They are such reliable sources, like all the time. My AI girlfriend told me so. She also told me she never plagiarizes from more reliable sources and says the ai push does not make those sources harder for normies to find. Plagiarism is also an outdated term the luddites use anyways. It trains, and there are no laws around training, therefore it is revolutionary, and the future. We all must accept this. Our new utopia is right around the corner, and all problems will be solved forever. I will obviously be on top, possibly the ruler of it all, because my AI girlfriend says I am the smartest manliest most biggest little guy she has ever met, and she never lies.
>>
>>
>>16986185
>Oh boy! I can't wait to eat slop
You don't have to, since if AI is so bad, then nobody will use it, and special people won't lose their jobs. Problem solved.
>Plagiarism
If humans are using a tool to plagiarise, then humans are using a tool to plagiarise.
Learning isn't plagarism. Everything needs to learn how to draw, say, a dragon in order to draw a dragon.
"We show that for
small and medium dataset sizes, replication happens fre-
quently, while for a model trained on the large and diverse
ImageNet dataset, replication seems undetectable."
"The amount of replication reduces
as we increase the size of the training set."
"The histograms of similar-
ity scores computed using the full dataset model are highly
overlapping. This strong alignment indicates that the model
is not, on average, copying its training images any more
than its training images are copies of each other"
https://openaccess.thecvf.com/content/CVPR2023/papers/Somepalli_Diffus ion_Art_or_Digital_Forgery_Investig ating_Data_Replication_in_Diffusion _CVPR_2023_paper.pdf
>>
>>16986298
>You don't have to, since if AI is so bad, then nobody will use it, and special people won't lose their jobs. Problem solved.
Mentally ill take.
>Learning isn't plagarism
Machine """learning""" isn't learning.
>>
>>16984536
every new technology leaves behind casualties.
https://rumble.com/v6x56z2-twilight-of-the-idols-the-last-hammer.html? e9s=src_v1_cbl%2Csrc_v1_ucp_v
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>
>>16986441
:)
>>16986447
start a dead pool on me then. see if i care or who cares. i hear that there was a movie about automated law enforcement. bye thread
>>
>>
>>
>>
>>16986377
>>16985377 (wtf)
>>16986298
>plagiarism
Neural networks are a data compression and indexing system. Its simile is Google, which is a perfectly reasonable index. If you search Google for fan art of Hatsune Miku the resulting images are not plagiarized, even if you pay for some premium ad-free search engine it's the same. AI performs something similar except in a hyper compressed format that in a lot of cases can be stored on most readers devices.
>is this plagiarism
At the end of the day, what the gooners are generating are like a cosine similarity of all images or chunks of texts relating to their prompt. It's not actually stealing anything- it's using public data, categorizing it, and averaging everything that relates to the user input. Whether this act of categorization is moral is obvious. If someone created a collage of 6 gorrilion images and hosted it on a site the image would be fair use. If someone took the first, second, third etc. word of every written document in existence and combined them in sequential sets, this would also be fair use.
>>
>>
>>16988515
except for that part where arbitrary portions of trained data can be reconstructed. Larger ideological concerns are similar to plagiarism which has led to convictions. Paraphrasing and rewriting a sentence in "your own words" is not necessarily protected.
>>
>>16984536
People are retarded and short sighted. They care about having a job and buying pointless shit. The advancement of human intellect and technology is irrelevant to them.
The top 0.1% know this. And will use this knowledge to monopolize everything that isn't already.
>>
>>16988515
It plagiarises less than search engines (see: >>16986298 ).
>>
>>16984536
Look, we've been over this.
By "we," I mean the French, and by "have been over this," I mean the French Revolution.
The French Revolution happened because 1% of the society made life impossible for the other 99%, until the 99% concluded that the solution to their problem was to physically destroy the 1%.
The AI craze will end with the Silicon Valley tech bros getting gulliotined. The tech bros are autistic animals with a 5-year-old's understanding of non-STEM subjects, that's why they neither understand how evil they are, nor can see the guillotine waiting for them.
>>
>>16988515
I can tell you're technically illiterate so I'll try to keep it simple for you: the outputs of a diffusion models are completely defined by their training set. Every single aspect of every single image you generate is fully derivative from other people's work. The only visual contribution of the model itself is in the artifacts AI sloppers are trying to eliminate. If you show me an image and say "I made it" but literally every single detail of it is due to other people, you're a plagiarist. Not only are you a plagiarist, but you couldn't plagiarize something so perfectly even if you were a master forger. The only thing that tops taking credit for AI slop is taking credit for a straight up photocopy.
>>
>>16988638
I mean, hell, even if you were desperate and tried to use picrel as a defense to the effect that arranging and combining other people's work can produce original new art, your diffusion model can't even do as much as Warhol because it would have to plagiarize the configuration of the grid itself.
>>
>>
>>
>>
>>16988648
no i only use it ironically
>>16988660
lick my nuts garon you still havent apologised for your ludicrous comments in the other thread
>>
>>16988638
>>16988640
Derivative works are still fair use.
>technically illiterate
You should be enthralled by my masterful analogy, proving deep command of both ML systems and discourse. If I were to ever meet you in person, I would likely kill you. This is a threat, and this is equivalent to my digital signature.
>>
>>
>>
>>
>>
>>16988638
The AI is operating on only a few parameters with terabytes of data. How could you possibly believe everything that the AI generates is a copy?
Let's say there's 20 billion parameters with 32 bits, and the training data is as low as 1 terabyte. That's only a rate of 0.08 bits (being extremely liberal).
The entire purpose of AI image generation is to NOT produce copies (i.e., produce new works based on the user's given prompt). It just being another search engine defeats the entire point.
To draw an orc, you obviously need to have seen another drawing or representation of one. Do you really suppose that humans don't do the same thing?
>>16988740
>fair use
Ideas can be possessed by multiple individuals at the same time (i.e., they aren't scarce), so because there's no negation of possession, there's no ownership (which is independent of the possessor). Copyright law is illegitimate.
>>16988797
>>16988798
>just ad homs
>>
>>
>>16990014
So, you acknowledge that AI image generators do create original artwork. Then you do want larger datasets, right?
"We show that for
small and medium dataset sizes, replication happens fre-
quently, while for a model trained on the large and diverse
ImageNet dataset, replication seems undetectable."
"The histograms of similar-
ity scores computed using the full dataset model are highly
overlapping. This strong alignment indicates that the model
is not, on average, copying its training images any more
than its training images are copies of each other"
"Very similar" here means not similar at all but doing similar facial expressions. Remember, this is at only 30000 images.
"Figure 3. The top two matches (according to different feature
extractors) for generations across diffusion models trained on
datasets of size 300, 3000 and 30000 (whole dataset). Across the
board, we can see full replication in the first 2 models (indicated in
green). Very close but not exact copies are indicated in blue. How-
ever in the model trained on the whole dataset, the first matches are
very similar but not the same"
>>
>>16985377
>steal the hardwork
They can still work as hard as they want. You're just mandating that people fork over their earnings to these people.
>>16985395
>stealing their skills
How do you suppose someone does that? Do they grab their skills out of thin air, then put them in a bag and run away?
>>
>>
>>
>>16991555
You're a mentally ill retard. No part of >>16988638 is about "copies".
>>
>>16991574
I was responding to >>16991474
Check the post that someone is replying to by clicking on the #.
>>
>>16991581
Is this (>>16989953) your post, psychotic patient?
>>
>>
>>
>>
>>
>>
>>
>>16991766
>>groups of individuals should be forced (at gunpoint, obviously) to not access the Internet at the whim of the government
Yes. Note that you have zero rational arguments against this that aren't braindead capitalist propaganda.
>>
>>