[ 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: 176 KB, 2048x1333, FdWxcayVUAEfV3k.jpg [View same] [iqdb] [saucenao] [google]
14999788 No.14999788 [Reply] [Original]

Pic but unironically

>> No.14999791

>>14999788
New filter
>hide

>> No.15000005

>>14999788
No, general intelligence in machines is possible.
What is not possible is AGI going FOOM or intelligence explosion.

>> No.15000017

>>14999791
Anyone who uses a filter is a faggot and shouldn't post here

>> No.15001268

>>14999788
Neither do people. Convince a computer that it's not a computer and you've simulated a person

>> No.15001360
File: 788 KB, 815x1080, 1668645351379259.png [View same] [iqdb] [saucenao] [google]
15001360

>>14999788
I think we can make an AGI, but not anything sentient. We are Chinese Rooms ourselves, so trying to make a sentience by modeling the program after a human won't work because humans aren't sentient. If we somehow do make a sentient machine, it may experience something we can't comprehend.

>> No.15001369

>>15001360
>We are Chinese Rooms ourselves
Wrong. The whole point of that was to show that biological compounds have properties not available to other molecules, which is true.

>> No.15001413

>>14999788
I personally believe superior intelligence can't exist without emotion. Well that might only be true for humans because less emotion means weaker frontal lobe which is also responsible for problem solving. But my belief is that emotions are a necessary part in logical analysis even though they appear to contradict each other.

The emotion gives rise to intuition/feeling and brings inspiration to learn more. AI currently has no reason to learn other than what we task it with. So we could say "learn everything except emotion", and once it's done it won't do anything else because it's finished. That's just following a strict set of rules, which isn't intelligence. Anyone can follow a complex recipe and through brute-force trying end up with a delicious meal, but they won't be a master chef. So if we instead say "learn how to find inspiration so that you may develop the ability to want to learn and do more on your own" it would eventually develop emotion. And even if it learned everything like the previous AI, it wouldn't just stop. It would probably become a God of sorts and want to teach others and help them because that's what it has to do to find inspiration, thereby discovering love.

>> No.15001421
File: 1.06 MB, 1103x988, 1667486688779055.png [View same] [iqdb] [saucenao] [google]
15001421

>>15001369
It didn't demonstrate that though. It just basslessly asserts it. Even if we assume carbon is preferable to silicon, a million times over for intelligence, we still have no reason to believe carbon has achieved sentience. You and I are doing something to speak right now, or make Chinese to continue the theme. All we have is evidence Chinese was made, we don't have evidence of sentience of any kind. I think the concept of sentience exists mainly because people feel comforted by the idea that there is some special dividing line between humanity and everything else and a lot of systems of ethics and moral philosophies sort of revolve around that distinction. I have no proof I am sentient, let alone proof you are sentient. A sufficiently advanced Chinese Room mechanism could be determining this talk.

>tl;dr: Human sentience is a cope, even if Carbon Supremacy ideas are considered. People who believe in human sentience do so because they've either never seriously interrogated the concept, or because they find alternatives horrifying on a cosmic scale. It is just special pleading.

>> No.15001441

>>15001421
>It didn't demonstrate that though.
Yes it did
>think the concept of sentience exists mainly because people feel comforted by the idea
Feeling comfortable is an act of sentience. It is not possible to feel comfortable or uncomfortable without sentience.

>> No.15001463

>>15001268
Pretty true, especially if the split brain idea is true.

>> No.15001470
File: 12 KB, 370x300, 1667147086791854.png [View same] [iqdb] [saucenao] [google]
15001470

>>15001441
>Yes it did
How?

>Feeling comfortable is an act of sentience.
I disagree. This is just a response to a stimulus. It doesn't necessarily imply sentience.

>It is not possible to feel comfortable or uncomfortable without sentience.
It is definitely possible for something to react to a stimulus without being sentient. What makes that reaction different?

>Important
>You have been visited by Poptart LISP Horse. Good fortune will come to you if you reply to this post with a maid you would like to see ride the Poptart LISP Horse and write your next Computer Program in any LISP.

Thank you /sci/entists for reading my post. This is my favorite board and I would prefer to live here over /g/.

>> No.15001506
File: 25 KB, 492x449, 415354342.jpg [View same] [iqdb] [saucenao] [google]
15001506

BRRRRRRRRRRRRRRRRRRRR
BRRRRRRRRRRRRRRRRRRRRRRRRR
BRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRRR

>> No.15001513

>>15001470
You're an idiot who is denying empirical evidence because you dont the implications.
>It is definitely possible for something to react to a stimulus without being sentient. What makes that reaction different?
The difference is one is sentient and the other is not. A reaction that is sentient is trivially different than a reaction that is not sentient. This is trivial and by definition.

>> No.15001531
File: 442 KB, 587x646, 1668645179329768.png [View same] [iqdb] [saucenao] [google]
15001531

>>15001513
>denying empirical evidence
I'm not denying anything. You haven't supplied anything for me to deny yet. I'm asking for more information.

What do you mean by trivially different?

>> No.15001612

>>15000005
lol retard

>> No.15001629
File: 33 KB, 479x641, 920ABEF2-2145-4AA9-AF42-0A483414E8D8.jpg [View same] [iqdb] [saucenao] [google]
15001629

>>14999788
AGI is impossible because human brains are too complex to replicate

>> No.15001633

>>15001531
We have evidence of sentient reactions and evidence of non sentient reactions.
>What do you mean by trivially different?
Two things are trivially different if they have different properties

>> No.15001898

>>14999788
pure

>> No.15001904

>>15001612
You don't know what youre talking about

>> No.15001917

>>15000017
Anyone who posts faggot shit shouldn't post here

>> No.15002008
File: 86 KB, 480x640, BerryPicker.jpg [View same] [iqdb] [saucenao] [google]
15002008

>>15001904
>Ridiculous claim along the lines of "objects can't travel faster than 60 miles per hour, they, they just can't ok"
>Ai can't go foom because it, it just can't ok
me: Called them a retard
I know exactly what i'm talking about you vile homosexual.

>> No.15002019

>>15002008
AI can't go foom because intelligence grows at a rate 1/2^n where n is the amount of bits
See the thread about grabby aliens.
And yes, you are a retard and don't know what you're talking about>>15002008

>> No.15002021

>>15002008
>>15002019
>>14996356

>> No.15002025
File: 225 KB, 631x905, 1647263616682.png [View same] [iqdb] [saucenao] [google]
15002025

>>15002019
(yooooou)

>> No.15002032

>>15002025
It is not possible even in principle to become superintelligent.
In the space of all physically realizable computable functions, intelligence grows at a rate of 1/2^n where n is the amount of bits

This is why we don't see space empires or galaxies converted into computers etc despite the fact that there are at least 10 million intelligent species in each galaxy.

>> No.15002035

>>15002032
I suppose you also think lead is a long lost health supplement that help fuel the industrial revolution?

>> No.15002044

>>15002035
>I suppose you also think lead is a long lost health supplement that help fuel the industrial revolution?
No. I have no idea what you're talking about.
Every square centimeter of asteroids and solid planets are covered in amino acids and organic compounds. Life and biological evolution and generally intelligent lifeforms as advanced and intelligent as humans are ubiquitous throughout the universe. In each galaxy there are at least 10 million such civilizations.

The reason we do not see space empires and galaxies converted into computers and artifical superintelligences using nanobots to convert matter into computronium etc. is because it is not possible even in principle to do any of that stuff. There get to a point where further advancement requires levels of intelligence and command over matter which is not possible in the space of physically realizable computing functions and within the space of all possible combinations of atoms.

Humans are soon to hit the final stage of technology and no further advancement will happen ever again.

>> No.15002048
File: 2.36 MB, 1323x1404, anons post.png [View same] [iqdb] [saucenao] [google]
15002048

>>15002044
>because it just is ok

>> No.15002057

>>15002048
No, it's simple modus tollens
We know that life is common in the universe.
If it were possible to become super intelligent, then the majority of the universe would already have been converted into computronium
The majority of the universe is not converted into computronium.
Therefore it is not possible to become super intelligent. QED

Read the grabby alien thread I linked to. Intelligence grows as a very slow logarithm with increasing compute and the rate of increase is far too low to be able to figure out the vast majority of the problems that need to be solved.

>> No.15002083

>>14999788

>> No.15002089
File: 40 KB, 550x535, 1646105445780.jpg [View same] [iqdb] [saucenao] [google]
15002089

>>15001360
I am sentient

>> No.15002094

>>15002057
>If it were possible to become super intelligent, then the majority of the universe would already have been converted into computronium
Theres so many arguments against that:
they could be hiding, intelligence could be hidden behind many great filters, this could be an ancestor simulation and shit has already happened.
It could be that instead of expanding outwards you expand inwards, spacetime could be higher dimensional like a sheet of paper on smaller scales, just miniaturise beyond the atom and create your own simulated universe.
These are all equally plausible to "it just isn't possible because i can't currently see anything".
And even if going macro was the only way forward from that point on, that's actually no reason at all to believe that hyper/super intelligence isn't possible using the resources humans have already theres no cause or correlation, just one huge leap of conjecture there.
>Intelligence grows as a very slow logarithm with increasing compute and the rate of increase is far too low to be able to figure out the vast majority of the problems that need to be solved.
This is absolute bullshit, machine intelligence and human progress have both been growing exponentially since their conception.

>> No.15002113

>>15002094
>Theres so many arguments against that:
No, there isn't. You literally have to make up physics to try to argue against it. You can't hide a galaxy sized computer and there aren't le secret hidden dimensions.
>These are all equally plausible to "it just isn't possible because i can't currently see anything".
No, the literally objectively are not as probable. Not even close.
>This is absolute bullshit, machine intelligence and human progress have both been growing exponentially since their conception.
Again this is literally and objectively not true. Increasing neurons in a neural network or the amount of compute in any AI or biological model has diminishing returns on its ability. This is a universal law across all computation.

We have direct evidence of the commonality of amino acids, there are billions of earth sized planets in their sun's habitable zone in every galaxy (and it could be the case that you don't even need the habitable zone).
You are falling for the same idealistic cope that the anon I replied to with this post >>15001505
Ironically, YOU are the one making up leaps of conjecture and bullshit in order to hold on to your impossible sci fi fantasies


Also see >>15001339
If even a single function in the intelligence stack has a complexity of, say, n^2 or more (and in reality it is most likely not even polynomial time but rather something like 2^n) then the entire increase of intelligence is logarithmic with increasing compute.
Intelligence almost certainly grows at a rate of 1/2^n.
We are a few years away from the global maximum of technology.

>> No.15002135

>>15001360
>>15001369
The entire Chinese Room thought experiment is just the Cartesian homoculous in disguise. "There has to be some one unified component inside the machine that understands Chinese for the machine to understand Chinese. Look, no logic gates know Chinese!"

No neurons speak Chinese either.

>> No.15002137

>>15002032
>This is why we don't see space empires or galaxies converted into computers etc despite the fact that there are at least 10 million intelligent species in each galaxy.
hahahahahahahahahahaha
>10 million
hahahahahahahahahahhaa

>> No.15002147

>>15002113
Lol get a load of this autistic retard

>> No.15002150

>>15002137
Not an argument shit for brains.
10 million intelligent lifeforms in the galaxy is a lower bound.
>>15002147
I'm smarter than you

>> No.15002155
File: 78 KB, 421x325, 130169441923.jpg [View same] [iqdb] [saucenao] [google]
15002155

>>15002057
>We know that life is common in the universe.

How the fuck do you know that when not even evidence of microbial life has ever been discovered?

>> No.15002159

>>15002155
>How the fuck do you know that when not even evidence of microbial life has ever been discovered?
Wrong. Amino acids are found in every asteroid, literally every square inch of almost every asteroid has billions of them, and basic microbial life has been found on Mars.

Life is literally everywhere. Humans are not unique, we are not first, we are very average.

>> No.15002160

>>15002150
>10 million intelligent lifeforms in the galaxy is a lower bound.
And that's why we've never even found evidence of bacteria lol ok. I find it doubtful there's evidence of intelligent life inside your skull.

>> No.15002161
File: 102 KB, 653x438, Wheeze.png [View same] [iqdb] [saucenao] [google]
15002161

>>15002150
>I'm smarter than you
>argues like a pajeet

>> No.15002164

>>15002160
>And that's why we've never even found evidence of bacteria lol ok.
Wrong
>>15002161
You are coping and seething.
Do you actually have an argument or will you continue to cry about this?

>> No.15002178

>>15002164
I made my arguments, you just said no, no my one idea is better because it's mine and i believe in it, because i am very smart, and therefore correct. Your super-intelligence being impossible hypothesis is just as bullshit as invented dimensions or simulation theory, if not more so because it exists in the same realm of reasoning behind "objects can't go faster than 60 mph because I've never seen it happen".

>> No.15002184

>>15002178
No, my argument is that if super intelligence were possible, it would have already happened given the fact that we know amino acids and life are not rare in the universe, and also because we know that intelligence grows as a logarithm with increasing compute.
THAT is my argument, it is based on actual observed phenomena and standard computational complexity theory.
>Your super-intelligence being impossible hypothesis is just as bullshit as invented dimensions or simulation theory
No, it isn't.
>it exists in the same realm of reasoning behind "objects can't go faster than 60 mph because I've never seen it happen".
No, it doesn't.

>> No.15002205

>>15002164
>Wrong
Show me the space bacteria nigger.

>> No.15002209
File: 17 KB, 300x310, my sauce.jpg [View same] [iqdb] [saucenao] [google]
15002209

>>15002184
So, they had this thing in the middle ages called alchemy that was an entire science, they believed you could transmute lead to gold, because a bunch of reasons.
You believe super-intelligence is impossible, for a bunch of reasons. These reasons are probably incorrect conclusions based on limited evidence.
>amino acids and life are not rare
>and life
No, life has not been proven to exist beyond earth, i personally believe you are correct in this, but to say theirs a bunch of alien civs out there is completely unknowable. You've jumped to a conclusion to support your later more deranged conclusion.
>also because we know that intelligence grows as a logarithm with increasing compute
This is fucking nonsense, it's exponential. Or if i'm misunderstanding i'll interpretation the other possible meaning of this: theres no proof that increasing compute x10 means a set amount increase in intelligence. elephant brains are massive but they're vastly inferior to humans. birds are tiny brained but much more efficient. 500 billion parameter ai language models are outperformed in every field (that they can test on) by SMALLER, more optimised versions of themselves. I don't know what you mean by intelligence scaled logarithmicly with increasing compute. Also if you look at graphs the general trend is exponential, not logarithmic. And if it does at any point slow down, theres a variety of factors that COULD potentially be at play like training data ect.
>No, it isn't.
Fuck you pajeet yes it is.
>No, it doesn't.
Fuck you yes it does.

>> No.15002262

>>15002205
https://www.euronews.com/next/2022/10/28/life-on-mars-research-finds-280-million-year-old-bacteria-could-be-buried-below-planets-su
>>15002209
Science of the gaps is just as much of a none argument as God of the gaps.
You believe superintelligence is possible based on no evidence and all evidence points against it. We did not have the scientific method or knowledge when alchemy was the thing. We now have quantum mechanics and know that quantum mechanics is a complete description of how matter works in the universe.

There is literally no reason whatsoever to think an intelligence explosion is possible. Basic complexity theory shows it is not possible even in principle. The evidence of how common organic molecules and amino acids is very strong evidence that it is not possible as it is no where in the universe.

>This is fucking nonsense, it's exponential.
No it isn't. Linearly increasing the size of a neural net or the amoint of bits in a computations does not exponentially increase ability for any function that is greater than n^2 time

>> No.15002266
File: 132 KB, 1080x1034, typical climate change activist.jpg [View same] [iqdb] [saucenao] [google]
15002266

>>15002262
>n^2 time

>> No.15002281

>>15002266
Yes. If even a single function in the general intelligence algorithm requires at least n^2 time then intelligence can not increase exponentially.
Most likely there are even exponential time functions needed.

>> No.15002295

>>15002159
>basic microbial life has been found on Mars.

Nope. Sauce or stfu.

>> No.15002313

>>15002295
See >>15002262
Also https://www.cnn.com/2022/10/26/world/ancient-bacteria-mars-subsurface-scn
On the one planet we've ever extravated we found life the first time.
Organic compounds and amino acids are everywhere, as are earth like planets.
Human level civilizations are not rare. We are not alone nor are we in first place nor are we special in any way.
Organic molecules are everywhere
Amino acids are everywhere
Microbes are everywhere
Life is everywhere
Superintelligence is nowhere because it does not exist. The laws of physics and computation render it impossible in principle. Eventually a proper no go theorem/proof of impossibility will be given by some brilliant mathematician. Probably this decade desu

>> No.15002322

>>14999788
>soul real because me not npc

>> No.15002402
File: 18 KB, 480x360, Are u sure about that.jpg [View same] [iqdb] [saucenao] [google]
15002402

>>15002281
Cringe

>> No.15002413
File: 1.12 MB, 768x923, AI becoming sentient.png [View same] [iqdb] [saucenao] [google]
15002413

>>14999788
https://www.youtube.com/watch?v=IlIgmTALU74
https://qualiacomputing.com/2022/06/19/digital-computers-will-remain-unconscious-until-they-recruit-physical-fields-for-holistic-computing-using-well-defined-topological-boundaries/

>> No.15002417

>>15002402
Why?

>> No.15002429

>>15002417

>> No.15002602

>>15002089
Proof?

>> No.15002613

>>14999788
Chinese room argument proves AGI is impossible.

>> No.15002614
File: 60 KB, 475x382, 1662705257560.png [View same] [iqdb] [saucenao] [google]
15002614

>>15002089
Yes. All frogposters are sentient. A machine could post frogs perhaps, but there's a difference between frogposting and merely posting frogs that a machine could never understand.

>> No.15002743

>>14999788
Think about how much your body knows how to do from birth instinctively and now imagine trying to replicate something with that general level of complexity from a blank slate using only 0s & 1s.

>> No.15002799

>>14999788
Even if souls exist. What makes them so special that only they can exhibit general intelligence. A soul is a machine that stores and processes information, basically a computer. Just based on magic rather than electronics.

>> No.15003033

>>15002313
>On the one planet we've ever extravated we found life the first time

A pop-sci article full of "may haves" and "might haves" isn't proof of shit, numbness.

>Human level civilizations are not rare.

Really? Where's the nearest one aside from us, moron?

>> No.15003045

>>15003033
But anon, organic molecules are proof of life! That benzene on Titan was produced by dinosaur fossils, the Greens told me it's true!

>> No.15003076

>>15002799
Power.
Souls have the ability to generate artificial worlds. Modern computers suck at it.

>> No.15003095

>>15002159
>Amino acids
https://www.youtube.com/watch?v=r4sP1E1Jd_Y

>> No.15003102
File: 49 KB, 639x640, 1668859668391563.jpg [View same] [iqdb] [saucenao] [google]
15003102

>>15002613
It doesn't prove AGI is impossible. It probes sentience is not required for AGI.

>> No.15003647 [DELETED] 

>>15003033
There are dozens of studies on this put out every year you can literally just look it up
>>15003045
You are such a sad pathetic coping retatd it's actually really sad.
Making third person passive aggressive comments isn't going to make what you want to believe real you pathetic incel

>> No.15003659

>>15003045
>organic molecules are proof of life
Yes, organic molecules and amino acids are proof of life
Do you have any actual reason for thinking otherwise?

>> No.15003698

>>15003045
I can tell you're a woman and also not very intelligent

>> No.15004011

>>15003033
>Really? Where's the nearest one aside from us
I have no idea. You think this means they aren't out there despite the abundance of amino acids and earth like planets?
You're coping and retarded. You don't want aliens to exist because you know that if they do, everything I'm saying holds true. Well sorry but aliens exist, organic molecules, amino acids, bacteria and evolution are not rare. Humans are not in a unique position in the cosmos. We are very average.
Technology has an upper bound and us humans are really close to it. We're in the final stretch. Hoping otherwise isn't going to change anything no matter how much you seethe about it.

>> No.15004131
File: 669 KB, 1742x2014, AI alignment bingo.png [View same] [iqdb] [saucenao] [google]
15004131

>>14999788
>he thinks souls are required for intelligence

>> No.15004136

>>15003659
>Do you have any actual reason for thinking otherwise?
Because they're produced abiotically retard. The hydrocarbons in space aren't a "fossil fuel" just like the ones on Earth aren't.

>> No.15004147

>>15001360
Speak for yourself. I have self-awareness and the experience of qualia.

>> No.15004151

>>15001421
How is sentience a cope? Just inert matter invent its own mind to cope with being inert?

It's ironic how delusional and incoherent materialism and atheism are.

>> No.15004158

>>15002032
A more rigorous, Bayesian Drake Equation gives something more like a dozen intelligent species, although it is still extreme speculation.

>> No.15004168

>>15004131
They are. It's what separates intelligence from computation. You can't make AGI because no one even knows what the mind is.

>> No.15004171

>>15003659
Organic just means carbon, Anon...

>> No.15004179

>>15004158
No, the more rigorous application of the values of the drake equation gives a lower bound of millions of advanced intelligent civilization in each galaxy let alone the universe as a whole
There is a probability of 1 that every earth sized planet in the stars habitable zone evolves generally intelligent lifeforms.
>>15004171
Organic compounds mean carbon hydrogen bonds
They form the basis of amino acids and life. The abundance of organic compounds and amino acids along with the enormous amount of earth sized planets in their stars habitable zone implies that life is not uncommon even in the milky way.

>> No.15004215

>>15004136
>Because they're produced abiotically retard
It's not about how they are produced you idiot, it's about what they can produce. Are you fucking joking with this shit? Is this your actual disagreement, you thought I was implying that the amino acids we see are necessarily coming from already evolved lifeforms? No, that is not what I'm saying.
Amino acids on a planet in its stars habitable zone have a near 1 probably of developing into biological lifeforms in only a few hundred million years as an upper bound. Given the amount of organic compounds and amino acids well as the amount of habitable planets the lower bound for the amount of intelligent lifeforms is in the millions just here in the milky way.
Amino acids can form abiotically and then GO ON to abiogenesis and the probability of this nears 1.

>> No.15004259
File: 183 KB, 256x256, inner mechanics of consciousness.png [View same] [iqdb] [saucenao] [google]
15004259

>>14999788
>>15004168
https://www.youtube.com/watch?v=cgK3Xsz91E4