[ 3 / biz / cgl / ck / diy / fa / ic / jp / lit / sci / vr / vt ] [ index / top / reports ] [ become a patron ] [ status ]
2023-11: Warosu is now out of extended maintenance.

/sci/ - Science & Math


View post   

File: 45 KB, 592x220, AGI.png [View same] [iqdb] [saucenao] [google]
15158399 No.15158399 [Reply] [Original]

Ignoring the problem does not make it go away

>> No.15158403

>>15158399
why's this dude having a conversation with himself on twitter

>> No.15158407

>strawman
opinion discarded

>> No.15158414

>>15158407
>strawman
I wish

>> No.15158462

>twitter screenshot threads

>> No.15158804

>>15158399
THERE IS NO SUCH THING AS GENERAL INTELLIGENCE

>> No.15158819

the jews are pushing muh AI HARD and it's STILL losing steam

>> No.15158915

>>15158819

its not, and its not about popularity,

>> No.15158916

>>15158403
The average AI schizo

>> No.15158917
File: 9 KB, 234x309, 1672066439175945.jpg [View same] [iqdb] [saucenao] [google]
15158917

>just do what humans
So why can't we? Ignoring money, what does a neuron do in the brain to allow thoughts?

>> No.15158934

>>15158819
>t. retard

>> No.15158987

>>15158917
Well in the first place, even if you're fully materialist and believe that consciousness originates from physical processes, silicon lacks the complexity to emulate the complex biological systems from which consciousness supposedly originates.

>> No.15159002

>>15158987
Isn't silicon in the same column as carbon? Shouldn't it have similar carbon analogs with just them being denser?

>> No.15159003

This isn’t even grammatically correct.

>> No.15159012

>>15159002
Carbon is much more versatile than silicon when it comes to the range of bonds and structures it can make.

>> No.15159023

I will admit that the advancements that did take place, that just five years ago I said would never take place, have given me pause.

>> No.15159028

>>15158804
What is general intelligence anyway. High intelligence in animals (us) is an evolutionary spandrel. It is a cosmic accident. It is absolutely useless. It’s just a matter of time before a cataclysm wipes it out.

>> No.15159029

>>15159023
Like what?

>> No.15159033

>>15159028
It must be difficult to live life as a materialist.

>> No.15159060

>>15159012
True since I imagine it would require higher energy for those silicone analogs. Still, what makes up a neuron as a structure and does it do as an organ? If just stores data could one not make a similar version in physical storage? If it creates patterns of electronical connections could you not use series of wires to mimic it? I don't even mean adult brain, could you start out small like copy a baby and follow the similar growth that humans go through to create a digital/mechanical brain?

>> No.15159061

>>15158987
It's not the silicone. It's the information processing. Information is a real thing in the universe.
https://en.wikipedia.org/wiki/Information_theory
https://en.wikipedia.org/wiki/Maxwell%27s_demon
https://en.wikipedia.org/wiki/Free_energy_principle

>> No.15159066

>>15159023
Humans are really really bad at recognizing exponential rates intuitively

>> No.15159070

>>15159033
thats not actually what anon said lol

>> No.15159276

>>15159060
The brain doesn't "compute" in the way binary logic gates do. It's an extremely complicated electrochemical process that we do not fully understand.

>> No.15159277

>>15159070
What he said was that he's a depressed materialist who can't fathom a world in which he is anything other than a useless eater.

>> No.15159300

>>15159276
Well then it looks like the first step to figuring out AI would be to figure out how the brain works. If they lowered the price on MRIs and allowed some more open brain surgery we might get there. Before any superbugs make surgery impossible.

>> No.15159301

>>15159028
Evolution and soience really did number on you didn't it chump?

>> No.15159384
File: 140 KB, 795x1080, CAESAR.gif [View same] [iqdb] [saucenao] [google]
15159384

>>15159276
Meh, that is completely irrelevant. As long as a brain can be reduced to a Turing machine current computers can do whatever a brain can do. Anything else is outrageous speculation that requires hard evidence.

>> No.15159408

>>15159060
we don't really know how it works and seems very complicated. just looking at 'neurons' means nothing because there's a lot more adding complexity. there's no way to replicate it by any methods available to us much less in a computer.

>>15159300
MRI is borderline useless and primitive to that end. so is looking at dead torn up cells prepared for a microscope.

>>15159384
ok so why can't it. irrelevant thing to say when no one has the slightest idea of even beginning.

>> No.15159430

>>15159384
>As long as a brain can be reduced to a Turing machine
It can't.

>> No.15159452

>create ai model
>Immediately add layers of censors, filters, and stopgates to stop it from being racist/sexist/ablist/offensive to anyone or anything
Lmao we're NEVER getting AGI, it's just going to be Alexa giving us milquetoast quotes and buying suggestions until the end of human civilization anon.

>> No.15159508

>>15159408
Well shit. Guess I'm out of idea short of just wiring people's brains to circuit board. And like >>15159452 said they're either gonna have a billion censors or worse just talk about 13/50 jokes with every question.

>> No.15159516

>>15158399
There is no such thing as "AGI", corporate shill.

>> No.15160571

>>15159066
Tech is not exponential. Otherwise we would be on mars and have flying cars by now. People said we would have that by year 2000

>> No.15161530
File: 669 KB, 2403x1785, The neural binding problem(s).png [View same] [iqdb] [saucenao] [google]
15161530

>>15159384
The brain isn't the mind though. This is where your physicalist presuppositions will always throw you off. See pic, especially my favorite part, here
>There is now overwhelming biological and behavioral evidence that the brain contains no stable, high-resolution, full field representation of a visual scene, even though that is what we subjectively experience (Martinez-Conde et al. 2008). The structure of the primate visual system has been mapped in detail (Kaas and Collins 2003) and there is no area that could encode this detailed information.
The subjective experience is thus inconsistent with the neural circuitry. This is just the visual binding problem by the way, there are many more.
Note: Notice they don't say 'the circuitry hasn't yet been found'. The whole thing has been ALREADY BEEN MAPPED. Subjective experience is INCONSISTENT WITH NEURAL CIRCUITRY. So there is no one to one correspondence between brain and mind. So any 'physicalism of the gaps' model fails. Any appeal to a one to one mapping some day FAILS. The computational theory of mind FAILS to account for consciousness. Computation is ONE of the things that consciousness DOES, as well as reasoning. We also already know that humans minds can harness computation to mimic some of the things that their own consciousnesses DO, but this says nothing about creating an experiencer to experience such experiences. You can't do that . A piece of virtual meat in spacetime is not the seat of consciousness and software created BY consciousnesses (that's what AI is) immersed in a VR is also NOT the seat of a consciousness either. This piece of meat you think creates consciousness never even needs to beed rendered ie have defined measurements in spacetime unless measured/observed, at which time what is probable to be there will be there, the same as with all matter. This is how you minimize computational complexity see vid
https://www.youtube.com/watch?v=JMImjFYZ1iY

>> No.15162259

>>15161530
Wait, so you found out about optical illusions and now you think magic and souls are real?
AHAHAHAHAH

>> No.15162299

>>15162259
An optical illusion would just be some kind of alteration of the stable, high-resolution, full field representation of a visual scene. It takes a stable, high-resolution, full field representation of a visual scene in the first place for such an alteration to occur. DURRR.

>> No.15162306

>>15162259
Also, in terms of this
>>15161530
not only can the content of a brain account for the CONTENT of the conscious experience, it can't account for how to create a consciousness to experience the content of consciousness in the first place.

>> No.15162316

>>15158399
Humans already have a way to create General Intelligence. It's called "sex".

>> No.15162339

These threads tend to make absolute cretins crawl out of the woodwork

>> No.15162340
File: 555 KB, 2753x2718, 325234.png [View same] [iqdb] [saucenao] [google]
15162340

>>15162316
>Humans already have a way to create General Intelligence. It's called "sex".
AI enthusiasts don't have sex and therefore have no way to create general intelligence.

>> No.15162811

>>15162340
i kneel

>> No.15162825

>>15161530
behaviorists saw the binary nature of action potentials and thought that meant neurons form individual components of a binary computer

they also thought you could program a chimp to speak

materialists jumped on the behaviorist idea of cognition from binary neural signals before any real investigation into the function because they weren't really looking for a robust explanation - just a physical (i.e. material) one

at no point have they been able to reconcile the inability to predict a neuron's outputs from solely its action potential inputs, and even computational neural networks cludge around this by assigning singular weights to individual nodes and adjusting them based on output alone (output that is judged by... preexisting consciousnesses)

>> No.15163065

>>15162825
and midlets retards see action potentials, thought of bits, and actually think this is how actual AI research works

>> No.15163098

>>15159066
you mean economists are really bad at thinking the start of a logistic curve is exponential

and like physicists, obsessed with a really bad habit of rolling as many terms together as possible so the end result is "elegant" (i.e. easy to fit on some paper) regardless of how nightmarishly lossy that compression is

>> No.15163108

>>15163065
yours is the brainlet take here - midwit is too kind
the communication between ML nodes is indeed binary, that's what the weights are for determining; this does not necessarily mean "bits"

since you have no fucking clue how any of this works, you are incapable of contributing to the conversation, and should find a thread that's more your intellectual speed (try looking on /b/)

>> No.15163243

>>15158987
citation needed

>> No.15163404
File: 112 KB, 627x449, 102 iq ai expert.png [View same] [iqdb] [saucenao] [google]
15163404

>>15159060
>go in ai thread
>find a wall of text
>check it's iq
>result: low iq
errrry single time

>> No.15164409

>>15159430
Machines more powerful than Turing machines are practically divine in nature, they can do stuff that belongs in the realm of magic. That is why you can't make such claim without hard evidence backing it.

>> No.15164436

>>15159452
I asked it to pitch me game ideas on several different occasions, 50% of them regarded climate change and sustainability. I then asked it to make a prompt for Midjourney for a painting of a booming society in a Norwegian fjord, and it immediately went on about "people of diverse cultures" living together in said society.
The fucks at OpenAI even blocked Hitler mode.
Why can't we have fun anymore?

>> No.15164598

>>15159277
I am sure you have evolved beyond the need to eat thanks to your immaterialstic beliefs.

>> No.15164649

>>15162340
Sexual reproduction of AIs when?

>> No.15164676

>>15158917
It's just a matter of scale. A neural network is the right basis to create a human brain (I don't even consider it AI, I would consider it a human), but the scale is not big enough. If you want to build a legit brain you need to go orders of magnitude bigger. If you ignored money then we would have artificial humans in a few years tops, but you don't get to ignore money in real life.

>> No.15164685

>>15159029
The advances in voice emulation and images are shocking

>> No.15164687

>>15159033
So do non-materialists believe in God and angels and other dimensions and demons?

>> No.15164717

>>15163404
>average IQ is low IQ
???