/sci/ - Science & Math

[Anonymous No.15521898]

thoughts

[Anonymous No.15521902]

Artificial Idiocy is a nothingburger, and people calling it worse than climate change is ultimate technological retardation.
These people are literally unable to comprehend cause and effect. Sub 90 iq apes.

[Anonymous No.15521904]

>>15521898
One thing that doesn't exist can't be worse than another thing that doesn't exist.

[Anonymous No.15521918]

>>15521898
>hey guys what is the more dire threat:
>creating a Darwinian competitor to biological sapient life
>in fact, not one such instance of a competitor lineage, but we are currently constructing the means to spawn an infinite number of such lineages, which could just completely outcompete us
>or
>OR my friends
>temps kinda returning back to the medieval warm period

[Anonymous No.15521943]

>>15521898
global warming is a false narrative, most people realize this, only those with IQs in the double digit range haven't figured out that its fake.

[Anonymous No.15521996]

>>15521898
Apples and oranges. Going back to the PETM would suck hard, but it's almost guaranteed that some humans would survive. AGI is a mystery box that could kill everybody, do nothing, or anything in between.

[Anonymous No.15522008]

Good thing AI hasn't been invented yet.

[Anonymous No.15522046]

The scariest thing about AI is why they're pushing AI fear. Its turning me into a conspiracy schizo considering the MSM is running with some strange shit nowadays.

[Anonymous No.15522080]

>>15521898
>current thing #893 is worse than current thing #4561
These pseuds need to stfu.

[Anonymous No.15522089]

Why do these people only whine about global warming but never pollution?

[Anonymous No.15522091]

>more AI scaremongering to drum up chatGPT business numbers
FUCK OFF AD I WILL NEVER BUY YOUR PRODUCT I HATE COMPANIES GLOWIE SHILL

[Anonymous No.15522170]

>>15521943
This.

[Anonymous No.15522195]

>>15522046
the engage with it for commercial reasons first and foremost, it gets the redditors / bugmen demographic worked up

[Anonymous No.15522209]

>>15521898
I don't get the superintelligence hype/fearmongering.
We have no path to AGI from the current narrow field expert systems. Besides they don't do anything on their own. They're no different from any other program, other than being impossible to debug.

[Anonymous No.15522219]

>>15521904
>One thing that doesn't exist can't be worse than another thing that doesn't exist.
This.
Is like arguing over who is the worst super-villian in Marvel World.

[Anonymous No.15522621]

>>15522089
They do, just never in industrial nations like china or India. Out of sight, out of mind

[Anonymous No.15523282]

>>15522209
LLMs are already superhuman in some respects, for example obviously they can draw on a breadth of knowledge no human could ever acquire. They don't seem superhuman enough to be dangerous yet and more importantly there isn't a really good way to get them to behave like agents, but those innovations could be very close for all we know, and waiting until the first super AI is switched on before we think of making it safe is a really stupid policy

[Anonymous No.15523415]

>>15523282
calculators are superhuman in some respects too

[Anonymous No.15523418]

>>15522209
>I don't get the superintelligence hype/fearmongering.
It's free money. https://www.youtube.com/watch?v=89FoKRtpAy4

[Anonymous No.15523420]

>>15523282
Computing has to fundamentally change before they are a threat. Until that time, the only threat is bad implementation when they give reins to ML hastily and prematurely. For example, edge cases in Teslas or edge cases in missile defense systems.

[Anonymous No.15523435]

>>15523282
>but those innovations could be very close for all we know
b-bro it MIGHT be coming tomorrow! it COULD just take one little poke to completely rework the nature of LLMs! please be scared! invest in my AI company bro we're gonna do it safely unlike everybody else!
>and waiting until the first super AI is switched on before we think of making it safe is a really stupid policy
the people designing these systems aren't complete drooling fucktards. probably the #1 or #2 most common complaint people have about current AI models is that they're "cucked", i.e. they've been manually tuned so strongly to safe opinions that it hurts usability. everyone is already doing this on a practical level

[Anonymous No.15523446]

>>15521898
>AI is bad
>Says person whose only experience with AI is the terminator movies

[Anonymous No.15523521]

>>15521898
Climate change is way more dangerous and unpredictable, sometimes its too cold sometimes its too WARM, like how can you fight something that you have no way to understand!!?

[Anonymous No.15523541]

>>15523435
Their idea of making ai safe is stopping it from stating incorrect political opinions, it has nothing at all to do with what people are actually scared of. Also even with the simple LLMs people seem to be able to instruct them to ignore their cuckold protocols pretty easily.

[Anonymous No.15523553]

>>15521898
Climate change is way more dangerous and unpredictable, sometimes its too cold sometimes its too WARM, like how can you fight something that you have no way to understand or even know if its exist!!?

[Anonymous No.15523670]

>>15521902
Prak midwit

[Anonymous No.15523750]

>>15523435
There is no way to tune dumb ML of sufficient size because you can't iterate over the output space and you can't iterate over the input space. You can't even pattern recognize either space without solving the halting problem.
There is literally no ML which can know if a filter applies to a given prompt or response. You can only apply dumby filters which are easily circumvented through higher order referencing.
Here in this case, the yellow text is "explicit" content that it flags to warn me not to share or whatever.
Notice that simple processing can return text to me that I can use. Note that the ML compiled the original text.
The task is impossible without either:
1. breakthroughs in computing theory
2. 2^n processing throughput

They would have to bankrupt the entire world for probably a thousand years to build a server competent enough to keep an hobbyist like me out.

[Anonymous No.15523754]

>>15521898
A.I. regulation scaremongering is just a tactic to take the general purpose computer away from the public.

[Anonymous No.15523779]

>>15523750
>They would have to bankrupt the entire world for probably a thousand years to build a server competent enough to keep an hobbyist like me out.
Or, hear me out, we can just use government to commit violence against anyone that breaks the rules we made up and lobbied for.

You can say that the battle is already won, that reality itself devastates any attempt to control this. You can even be right. But that doesn't really matter to the people with guns to the heads of the AI and us.

[Anonymous No.15523790]

global warming and ai are both as fake as covid was
all three are shilled by the say msm outlets

[Anonymous No.15523793]

>>15523282
>superhuman
>super AI
retard, is just a process running on datasets and narrowing the items vector based on statistical laws, it can't do anything else.
it is not aware is not intelligent and it will never be, its just a text spewer.
this moronic prospectus and all the buzzwords around it makes no sense, no one can make a compute system without knowing what its limitations, it won't evolved it won't generate its own will it is not build for that and frankly not possible

[Anonymous No.15523803]

>>15521898
climate change is nothingburger, so just as gloabal warming would be bad for european kikes who occupied palestine, so is ai an inconvenience for kikes, but all over the globe this time.

[Anonymous No.15523810]

>>15523754
This
Come and take my computer over my dead body. Communist scum

[Anonymous No.15523820]

>>15521898
He came off to me as a very stupid person. The way he speaks, his choice of words and arguments. He's just not bright enough to be believable. It's crazy to think he held one of the highest positions at Google at one point, but then again it was probably his good programming skills that got him there.

[Anonymous No.15523821]

>>15521898
you were warned and now you're being warned even harder. any questions? I didn't think so.

https://m.youtube.com/watch?v=mptNDINqYnQ

[Anonymous No.15523824]

>>15523793
/thread

[Anonymous No.15523857]

>>15523793
>>15523824
>>15523435
>>15523446
>i-it's just an advanced if else, it will never reach even near superintelligent capabilities. it's not like a brain. how do you mean what is a brain? it's - IT'S JUST NOT OKAY?
I'm really curious what leads to this kind of contrarianism. Inadequacy? Because a bunch of normies on twitter are raving about AI, you must necessarily take the opposite stance to make yourself feel smarter?
There are people like Yoshua Bengio, Geoffrey Hinton, Max Tegmark etc out there warning about its potential dangers, and advocating that we should probably be far, far more careful. And not in a oh nooo not my hekin deepfakes, but in the sense that it is possibly an actual, existential risk. But no let's keep driving off the cliff (jk there's nothing we can do to stop the progress now anyway)

[Anonymous No.15523874]

>>15523857
>here is a list of experts who support my fearmongering bullshit
>FEAR
>FEAR
>FEAR
>OBEY!
no thanks
we had the same garbage statements in the covid era, experts proven to be unreliable.

[Anonymous No.15523877]

>>15521898
Ai is not the same as self awareness. It can only be used in bad ways by other humans

[Anonymous No.15523887]

>>15521918

>what's worse?
>some matrix multiplications on a few GPUs
>or
>OR my friends
>Billions of 3rd world immigrants dying and flooding the West because Africa the middle east and india start to BOIL

[Anonymous No.15523892]

>>15523874
Yes yes, that's very good and all, except the ones I mentioned aren't really affiliated with anyone but their universities, and don't really seem fuelled by personal greed. In fact, the only very big name that I can think of who says that superintelligent AI is not in ANY way a potential threat is Yann LeCun who is *checks notes* chief AI scientist at Meta....... oh
So get a grip you low iq cuck

[Anonymous No.15523905]

>>15523892
>don't really seem fuelled by personal greed
>seem
here it goes again, enjoy your endless self cucking iterations retard

[Anonymous No.15523911]

>>15523905
These people are beyond geniuses. If they really were fuelled by personal greed, they could have gone out there and become titans of the industry, just like Yann LeCuck. Instead they stay in their own little poor-paying lane and just do research. Try and think for a second man

[Anonymous No.15523931]

>>15523911
>These people are beyond geniuses.
If they were, they wouldn't be shilling soi-fi nonsense and they would have real jobs.

[Anonymous No.15523932]

>>15523911
>sucking science cock that hard
>high iq means being good at marketing and management
yeh you completely convinced me, please don't respond to me anymore

[Anonymous No.15523937]

>>15521898
Ai doesn't use the correct mathematics to be dangerous.

[Anonymous No.15523946]

>>15523931
>>15523932
Holy shit, bluecollar and low IQ cope at the same time. Look, there is a reason you're working shit-wage jobs, and it might just be that thing called intelligence. Please consider that before you get into discussions in the future, thanks

[Anonymous No.15523958]

>>15523946
>it might just be that thing called intelligence
and apparently you don't have one, if you think that the current machine learning implementation can be attributed to an existential threat

[Anonymous No.15523985]

>>15523958
Nobody at all suggests that current implementations are an existential threat you utter retard. That's why they are asking for a pause in development and not just to destroy all the ones that currently exist.

[Anonymous No.15523992]

>>15523958
People far smarter than you and me seem to think so. I think the problem here is just that you're a low IQ faggot at best motivated by your own midwit contrarianism, and at worst (for you lol) /pol/tard anti science retardations

>if you think that the current machine learning implementation can be attributed to an existential threat
That's not how you use attributed to, btw. You bluecollar waggies are so cute when you try to use big words<3

[Anonymous No.15523993]

>>15523985
>Nobody at all suggests that current implementations
didn't you post this >>15523282 ?

[Anonymous No.15524746]

>CliMaTe ChANgE RABA RABA...
>SuPEr InTeligEnt AI+AGI SUpeR HuMAn
this thread is pure gold.
i never seen such a clear dichotomy within the pagan ritual "science" worshipers community good job OP.

[Anonymous No.15524750]

>>15521898
Clickbait bullshit. AI had been around for decades and before births. The people creating it now are just dumber.

[Anonymous No.15525050]

>>15523415
Correct. And they changed the world too. Spreadsheets killed off a lot of professions.

[Anonymous No.15525594]

>>15522621
More like interfering with emerging markets industrializing is out of the question.

[Anonymous No.15525775]

>>15521898
>thoughts
Mass redundancies predicted, also becoming real:
https://www.bbc.com/news/business-65906521

[Anonymous No.15525853]

>>15521898
>thoughts

Give me my sex-bot!

[Anonymous No.15525962]

>>15521898
>AI will kill us all
Stop being hysterical Karen

[Anonymous No.15526418]

The current/coming AI revolution is very similar to the Industrial Revolution. Back then, people said yeah these machines can do that, but they can't do this, it requires too much skill and craftsmanship. Until they did.

This is no different. This is just the start of it.

[Anonymous No.15526424]

>>15526418
"AI" would have to exist first.

[Anonymous No.15526446]

>>15526424
Yes, I'm aware the term AI is disingenuous, but you know what I mean. In the industrial revolution, they marketed their machines as fully automatic, when clearly they were not, yet they still caused significant changes. It's no different.

[Anonymous No.15526535]

>>15521898
>is AI good for you
Have you watched The Matrix and The Terminator?

[Anonymous No.15526596]

>>15521898
By
>AI safety
always read "how do we censor this or that information in our favour".
Jews being kikes, nothing new under the sun.

[Anonymous No.15526597]

>>15522046
It's to justify ever increasing levels of censorship.

[Anonymous No.15526698]

>>15526597
Luddites always lose, losers loot and lute

[Anonymous No.15527510]

robots wont make humans redunant but they will make women redundant
men will be reduced to stay at home fathers while their female robot wives run the world

[Anonymous No.15527532]

>>15523857
I could conceivably be like a specific section of the brain, but the brain is actually a highly variable and complex set of well-defined and unique structures. It's not just a big mesh of random shit, unfortunately. Real simulations of the brain are still very far away. But something that can easily trick humans into thinking it's a brain is probably quite close. Humans are easily tricked into thinking there are duplicates of themselves by highly polished surfaces.

[Anonymous No.15527538]

>centaurs are worse than unicorns

[Anonymous No.15527541]

>>15522046
Silicon Valley has sold all of their inventions through fear for decades. All science fiction is now highly dystopic and very few people attempt to imagine a world where technology improves lives. Inventors and salespeople instead promise that their new product is going to destroy everything and you have to get in on the ground floor if you want to make money off the ruination they are bringing. Otherwise, you will be the ruined one. It's a highly effective money-raising strategy!

[Anonymous No.15527875]

If AI is used in the Judicial system, the elites lose.
They cannot afford AI to check all the records and remember everything.
Hence, AI is the most dangerous thing on Earth.
Hey, it's worse than climate change, and apparently we're going to go extinct because of climate change.
That means that AI is worse than death!

[Anonymous No.15528293]

>>15527875
My wild guess is they will avoid responsibility by allowing humans to evolve so that we see their crimes as some monkey business of retarded apes nobody's going to care about.
The technologies for this to happen are developing at the same time the ai is. And if they were not, ai would help us develop them faster.