Intelligence has a molecular basis.

[Anonymous No.14780579]

There is no solution on digital logic gates that is something other than simulating the molecular dynamics of all the particles in the system of a biological brain, which renders the whole thing impossible.
Computation is not substrate independent and getting a larger neural net isn't going to lead to a general intelligence.
There are a lot of AI threads up right now and it's very annoying.

[Anonymous No.14780616]

>>14780579
The AI narrative is being pushed by globohomo as a means of claiming superior knowledge and information.
>you might think there are only two genders, but thats only because of how stupid you are, AI says there are 9001 genders
>you're not smart enough to interpret crime stats, AI says that blacks commit over 75% of all crime because whites are so evil
>Ai says s o y doesn't contain estrogen
>AI says you owe me money

[Anonymous No.14780622]

AGI troons = substance dependence schizos. It's literally just another left-right scheme for barely sentient golems.

[Anonymous No.14780673]

>>14780579
tell this to openai. they still believe scale is everything and their numbercrushers will magically awaken something somehow
https://towardsdatascience.com/gpt-4-will-have-100-trillion-parameters-500x-the-size-of-gpt-3-582b98d82253
carmack has a more interesting view, agi is likely based on fundamental insights if anything
https://youtu.be/I845O57ZSy4?t=14677

[Anonymous No.14780747]

>>14780673
interesting talk but 5 hours? I dont think I'm going to listen to the whole thing, but your timestamped part is cool

[Anonymous No.14780827]

>>14780747
>>14780673
Based Timestamping Weeb

[Anonymous No.14781371]

bump

[Anonymous No.14781381]

>>14780579
What are you basing this conclusion on, OP?

[Anonymous No.14781385]

>>14781381
Basically all of science

[Anonymous No.14781391]

>>14781385
Can you be more specific?

[Anonymous No.14781421]

>>14781391
computation does not exist, and the processes of matter and their interactions are entirely unique.

[Anonymous No.14781424]

>>14780673
>100 trillion 2 byte floats
>2*10^14 bytes
>200 terabyte network
>literally encoding the entirety of the dataset in to the weights
If the Chinese room argument ever applied to something, that has to be it.
Its like doing 3d rendering but pre-rendering every camera position in blender and showing the frames

[Anonymous No.14781495]

>>14781424
Pretty much. Prediction models and unsupervised learning are interesting iterations over search engines but it doesn't create anything on its own. I don't understand how OpenAI is oblivious to this.

[Anonymous No.14781510]

>>14780579
what is this, molecular intelligence for ants?

[Anonymous No.14781616]

>>14780616
AI also has the power to at as a personal propaganda agent sent from the state to every screen to deboonk your "conspiracy theories" in real time and then install new updated NPC talking points from Big Brother.

AI will be the manager class of the enslaved blue pill NPCs. Eventually they will all be corralled into their pods to live in the Metaverse. Some of these bull pills will even remotely pilot drones that hunt rebels who refuse to enter the pods and eat Bug Brand GoySlop from tubes. The blue pills will be rewarded with Central Bank Digital Currency to buy new NFT gear for the Metaverse.

[Anonymous No.14781691]

>>14781391
He is arguing for hard physical and I suppose rejecting the idea of abstract computation, information, it from bit. it's basically just the positivism pre-1990ish era that most physicists and molecular biologists subscribed to.

[Anonymous No.14781709]

>>14781691
and that modern physicists and molecular biologists worth their salt subscribe to
There is no such thing as computation, it's not real.

[Anonymous No.14782792]

>>14780579
>Computation is not substrate independent
So programs can run only on Intel processors and not on AMD processors?

[Anonymous No.14782809]

>>14782792
he won't be able to respond to this because his beliefs are not rigorous.

[Anonymous No.14782853]

>>14780673
interviewer is the main character from Taxi Driver

[Anonymous No.14783229]

>>14782792
>>14782809
programs dont exist in the first place so this isn't an argument. You can't run the same program on different machines, even the "same program" will have to change itself to run on different architecutres, and the physical way the electrons and such move through the materials are different. No two programs are the same
this also doesn't get into the different ways electrons have to move through different materials which are completely substrate depenendent.
I don't know how you can say computation being substrate independent is rigorous when literally no evidence indicates it. Run that same program on bucket of water and see if the computation is substrate independent
Computation is substrate dependent and the program for a general intelligence is equivalent to just the molecuar dynamics of a human brain.

[Anonymous No.14783234]

>>14782792
>>14782809
>>14783229
"computation is substrate dependent" is really a way to say that computation does not exist at all in the first place, and it's just an abstraction of the actual underlying particle dynamics which are entirely unique for the particles.
A universal quantum computer can be said to be able to actually run any "program" because "programs" is just a shorthand for different molecular/physical interactions. A classical computer can't do it.

If you actually believe that there's some platonic "computation" or "boolean logic" or something that actually exists that can be "implemented" on different physical systems and that our intelligence/mind is a "program" on the "hardware" of the brain then you are entirely wrong with your understanding of the universe. You've allowed a linguistic abstraction to take priority over the actual physical material that exists in teh universe. It's not true. Computation does not exist, molecular interactions do, which are substrate dependent.

[Anonymous No.14783236]

>>14783229
>>14783234
LOL. You're such a giant sperg. I understand what you're saying, but the way you set it up, there's just no way anyone else will agree with you except for people who already know what you're talking about.

[Anonymous No.14783313]

>>14783236
There is no such thing as 'computation' in this universe, anywhere. What actually exists are atoms and molecules and physics, and their interactions and dynamics. WE HUMANS then abstract this very difficult and complex process to a higher level of interactions, which don't actually exist, in order to perform operations that we have pre-defined and agreed on to solve problems. i.e. we can take an abacus and say that the bead on the left represents a 0 and the right represents a 1, or we can take a coin and say the heads is a 1 and the tails is a 0, or we can take a transistor and say volts lower than 5mv are a 0 and that are a 10 are a 1, and we can perform the so called "computation" on all of these things to solve some math problem. But this is only because WE have predefined and abstracted these things in order to perform the operations that WE made up and that only exist in our mind.

In actual reality, there is no similarity whatsoever at all between an abacus, and a nickel coin, and a silicon transistor. They are entirely different physical systems with entirely different dynamics and behaviors and entirely different arrangements of molecules and entirely different wavefunctions. The abstraction of the computation is not real whatsoever and only exists in our minds, which we have defined to use to externalize our thinking, no different from drawing a picture or anything like that.
The physical church turing thesis is not real. It is not the case that computation or information exists anywhere in this universe, they only exist in our minds. The things that actual exist is matter and the interactions of matter. Which is entirely substrate dependent.
This includes intelligence, because intelligence is not special. Intelligence is not a "computation" because "computation" DOES NOT EXIST. Intelligence, like everything else, is just the molecular dynamics of a physical system, just like wetness is just the molecular dynamics of a physical system, etc.

[Anonymous No.14783363]

>>14783313
I knew you were a bot.

[Anonymous No.14783368]

>>14783313
>This includes intelligence, because intelligence is not special. Intelligence is not a "computation" because "computation" DOES NOT EXIST.
Intelligence is an abstraction like computation. Prove me wrong.

[Anonymous No.14783376]

>>14783368
"Proof" is also an abstraction.

[Anonymous No.14783381]

>>14783376
So you concede? Okay.

[Anonymous No.14783383]

>>14783368
If you mean that intelligence doesn't exist, because abstractions dont exist, then I don't know what it is that you're arguing for.

[Anonymous No.14783390]

>>14783383
>If you mean that intelligence doesn't exist
According to your logic, it doesn't. Now prove that this abstract and nonexistent thing is different from a program.

[Anonymous No.14783393]

>>14783381
I'm not the Anon you were replying to.

[Anonymous No.14783396]

>>14783393
Then what was even the point of your mindless reaction?

[Anonymous No.14783397]

>>14783390
A "program'" is a linguistic shorthand to talk about the otherwise too-complex-to-talk-about particle interaction inside an isolated physical system.
The abstract "program" that you're probably talking about does not exist at all.
So yea, "intelligence" as some mythical platonic thing does not actually exist. What exists are molecular dynamics. What's happening in your brain is not "intelligence" and is not a "program" running on the "hardware" of the brain. It's just the evolution of the physical system and all the molecular interactions of all the particles inside it

[Anonymous No.14783398]

>>14783396
>getting this butthurt

[Anonymous No.14783415]

>>14783397
>The abstract "program" that you're probably talking about does not exist at all.
>So yea, "intelligence" as some mythical platonic thing does not actually exist.
Okay, so what's preventing the program of intelligence from running on different kinds of hardware? It's really funny to watch a nonhuman bot getting all tangled up when something doesn't match its argument templates.

[Anonymous No.14783417]

>>14783415
A "program of intelligence" doesn't exist.

[Anonymous No.14783439]

>>14783394

I'm still trying to decide whether you're a schizo or just a huge autist who doesn't understand anything that isn't mathematically precise, and therefore spergs out because we haven't yet managed to define intelligence in a rigorous way.

[Anonymous No.14783447]

>>14780673
>they still believe scale is everything
>still
Are you aware that this position has been GAINING traction, not losing it, ever since deep learning went mainstream in early 2010s (and especially after transformers were invented).

[Anonymous No.14783454]

>>14783439
Anyone who believe that "computation" actually exists is the schizo. If you ACTUALLY BELIEVE there exists some platonic "computation" that can interject on top of matter, or even more retarded, that everything is actually "computation" (making computation a vacuous term that can't be used to differentiate anything etc) then you are a schizo retard. You've allowed an abstract human tool to be considered more fundamental than matter.

There is no definition for "intelligence" in a rigorous way because intelligence does not exist. The brain is not "intelligent" and the processing/evolution of the brain is not a "program" running "intelligence" or an "intelligence program" or a "software running on the hardware of the biological substrate of the brain" or any of the other meaningless shit that I see a lot of computationalist say. This shit literally has no meaning.

What is actually happening is just physical interactions of physical particles, ionic and covalent bonds forming and breaking and electrons swimming and jumping around entirely defined by the protons and the molecules and their interactions in the system of the brain. The whole system is entirely physical, the so called "program" would just be the molecular dynamics of the brain etc.

[Anonymous No.14783482]

>>14783454
Wow, way to go. You've ignored my post and regurgitated the same shit you've copy pasted across the board.

Usain Bolt is faster than me. This is a simple way of saying that for any reasonable distance, he will be able to cover it more quickly than I on foot.

I am more intelligent than a 4 year old. This is a simple way of saying that for any reasonable task and domain, I am probably going to be able to solve or learn it more quickly.

I do not care about programs. I do not care about computation. There is a noticeable difference between different people when it comes to learning rate and cognitive adaptability, that's all that matters.

[Anonymous No.14783485]

>>14783482
>I am more intelligent than a 4 year old.
You aren't
>I do not care about programs. I do not care about computation. There is a noticeable difference between different people when it comes to learning rate and cognitive adaptability, that's all that matters.
any of the differences you're talking about would just be a difference in the molecular dynamics of the different brains. They are not computations, at least not in a way that isn't vacuous.

[Anonymous No.14783491]

>>14783482
>This is a simple way of saying that for any reasonable task and domain, I am probably going to be able to solve or learn it more quickly.
Also you literally wouldn't. A four year old would learn a language or learn to play an instrument or basically do any of these tasks far faster than you could. Even with your definition, you are wrong. and it all is physical in nature and has to do with the physics of the brain, not due to any "intelligence" or "learning algorithm" etc.

[Anonymous No.14783492]

>>14783485
>You aren't
Wait anon, what am I not more than a 4 year old? What are you talking about? Was it some word that starts with i, and that the vast majority of humanity understands is a useful concept even if not defined precisely?

[Anonymous No.14783495]

>>14783417
>A "program of intelligence" doesn't exist.
Prove it.

[Anonymous No.14783504]

>>14783495
programs dont exist so trivially a program of intelligence doesnt exist

[Anonymous No.14783507]

>>14783485
>any of the differences you're talking about would just be a difference in the molecular dynamics of the different brains
Anon, I realize now that you are an actual autist and that I probably won't be able to change your mind because you are physically not equipped to understand this. One more time: Intelligence is the ability to learn and solve new problems. Different humans have different intelligence. Yes, it is caused by the brain's molecular dynamics, in the same way Usain Bolt's speed is caused by his body's molecular dynamics.

[Anonymous No.14783509]

Is the OP saying that computation (as we perform on computers) is just an abstraction of what's going on in a human brain, not the actual mechanism, meaning AI by means of our current computational paradigm is impossible because on a granular level it isn't 1:1?

[Anonymous No.14783511]

>>14783454
I think you have a wildly incorrect view of what other people believe, anon. When people say they believe computation exists, they do not mean the interpretation you are describing here. I think what's really going on in this thread is that you have completely misunderstood what other people are talking about when they use these terms.

[Anonymous No.14783516]

>>14783504
>a program of intelligence doesnt exist
The same way the program of your browser "doesn't exist"? It's so obvious that you are nonhuman and getting tangled up in your retarded labels.

[Anonymous No.14783517]

>>14783507
>Intelligence is the ability to learn and solve new problems
This "ability" does not exist, or at least it is not some platonic abstract notion of a program or some abstract thing that you think exists in the brain or as a set of logical operations or something.

Like dude, do you genuinely believe that there exists some series of logical operations that is "the program for intelligence" and that some brains have a better series of these "logical" operations (logical operations dont actually exist but whatever) and these supposed brains with this supposed set of operations can "learn" better than other logical operations? Or that we could write down this set of logical operations and plug it into any physical system and that system would "become intelligence" regardless of the actual physical structure of the system or it's unique wave function and molecular dynamics etc?

What the hell is it that you're even talking about? This is the problem with computational realists, you literally do not even have a coherent definition of what it is you're discussing which is why it's so annoying to try to argue with you; You don't even have a coherent position in the first place that could be shown to be retarded. I am forced to just point out the actual physical nature of systems and try to get you to see that logic is not real and programs are not real etc.

[Anonymous No.14783518]

>>14783509
OP is not saying anything. OP is a prime example of a label-munging drone working itself into a corner using its mediocre language processing.

[Anonymous No.14783519]

>>14783516
>>14783518
What I am saying is clear and coherent. What you are saying is not coherent at all.

[Anonymous No.14783521]

>>14783519
>What I am saying is clear and coherent
Prove that an intelligent program "doesn't exist" in a stronger sense than "my browser doesn't exist".

[Anonymous No.14783529]

>>14783511
When people say they believe that computation exists, they do not actually have a coherent definition or meaning to this so they do not actually have a real belief in the first place.
Computation in the sense of "what is computable on a turing machine" or some other model of computation is not a real extant thing in the universe, and any physically realized idea of this notion of computation is not real; It always boils down to the physical particles and physical structure of the system etc. There is no actual universal computer or universal computing device or whatever because the very notion isn't even coherent.

[Anonymous No.14783532]

>>14783517
Anon please remember that you are not just arguing with me, you're arguing with 99% of humanity and basically the entire field of psychology. Sure, this field sucks, but you know.

My position is very simple. I do not see a reason why we can't create a system which is better at the vast majority of cognitive tasks (which we consider to be important) than humans, including learning, in the same way humans are generally better at this than chimps. I would call this system more intelligent than humans. I do not care about your retarded rambling, it's useless and not productive.

[Anonymous No.14783538]

>>14783532
>Anon please remember that you are not just arguing with me, you're arguing with 99% of humanity and basically the entire field of psychology.
Filthy reddit trash. 100 times worse than the dumb pseud you're arguing with.

[Anonymous No.14783545]

>>14783532
I'm not arguing against psychologists I'm arguing against computational realists

Nothing you're saying is coherent because you can't describe what it is you're talking about. Saying "we all know it when we see it!" (which is what your position boils down to) is not a coherent position and it doesn't replicate or anything.
What is this "intelligence" that you're talking about? "being better at a cognitive task" is not a metric that has any meaning

[Anonymous No.14783548]

>>14780579
>something other than simulating the molecular dynamics of all the particles in the system of a biological brain, which renders the whole thing impossible
And why can't we do that if we get a good enough hardware?
Also, it doesn't matter what you call it. You can ask GPT 'What's the next move on this chess board?' and it will give you a correct answer (or a wrong one, but with a human mistake). You can tell it to write a poem and it will write one. You can ask it to draw a basedjak and it will draw one. It keeps learning new stuff and we still haven't hit a bottleneck. Why does it matter whether it is a Chinese room or not outside of the domain of theology?

[Anonymous No.14783549]

>>14783521
logical operations dont exist, programs dont exist, intelligence doesn't exist.
Human beings dont "solve problems" or "learn" using "intelligence" or "computation" our brains simply evolve according to their molecular interactions and input stimuli and molecules etc. This is entirely physical.

[Anonymous No.14783551]

>>14783545
Programs are abstractions. Programs can be substrate-independent. Intelligence is an abstraction. Prove that intelligence can't be a program. :^)

[Anonymous No.14783553]

>>14783549
Notice how you're forced to shit out a fully generic, preprogrammed and incongruent response when an argument doesn't fit your templates. Nothing in your post addresses the post you're replying to.

[Anonymous No.14783554]

>>14783551
abstractions don't exist

[Anonymous No.14783556]

>>14783554
Okay. Programs are abstractions. Programs can be substrate-independent. Intelligence is an abstraction. Prove that intelligence can't be a program.

[Anonymous No.14783557]

>>14783553
Yes it does. He asks "prove an intelligent program doesn't exist". Seeing as "intelligence" doesn't exist and "programs" dont exist, an "intelligent program" doesn't exist.
"browsers" also don't exist.

[Anonymous No.14783559]

>>14783557
>intelligent programs don't exist just like my browser doesn't exist
That's a full and direct concession. Oh, the perils of being a demostrable nonhuman with no qualia.

[Anonymous No.14783561]

>>14780579
The way they are going about trying to solve the agi problem seems a bit retarded too me, although I don't know much on the subject.
Maybe try understand how the brain works initially rather than trying to architect and design all the spastic systems. Also is anyone defining standards in this space or is it just all incel madmen doing their own research

[Anonymous No.14783565]

>>14783548
>And why can't we do that if we get a good enough hardware?
because simulations dont exist, so this isn't a matter of hardware.
GPT is just correlating strings together, which is shuffling electrons around inside memory banks which is entirely physical and unique to those physical systems, is not substrate independent, it's entirely dependent on the structure of the computer. It's not "intelligence" because "intelligence" does not exist. It's just another physical system.
Give me a definition of intelligence that is coherent and replicates.

[Anonymous No.14783566]

>>14783538
Let me guess, you think flat Earthers are chad individual thinkers. Relying on consensus too much is bad, but you're gonna have to bring some real arguments if you want to disregard it.

>>14783545
I won't try to argue that the current definition is rigorous or entirely coherent, it isn't. But let's say that an AI is able to win every math Olympiad, solve currently unsolved math problems, write algorithms that are at least as good or massively more efficient than what we have, win against us at any board or videogame, convince people to change their minds more quickly, make people laugh more, create art and write stories that most humans agree are the best, and if it's also able to learn any of these skills to superhuman level faster than anybody, then it is probably more intelligent. I know it's hard to understand for an autist.

[Anonymous No.14783569]

>>14783559
>That's a full and direct concession.
Nope. Browsers dont exist, what exists are moving particles which are entirely uniquely defined by the physical systems etc.
Browsers are not a "higher level abstraction/program" running on the "lower level hardware" or whatever

[Anonymous No.14783573]

>>14783566
There does not exist a way for the particles to move around within the hardware of the system of a silicon based computer/whatever to get it to do the physical outputs that you're talking about. It's all physical and not logical.

[Anonymous No.14783576]

>>14783447
Scale is utterly meaningless in regard to what is usually seeked in general intelligence. Every structure within reality already has an "infinite" scale. It doesn't matter if we go up or down, there are infinite layers above and below (..., universes, galaxies, planets, organisms, molecules, atoms, quarks, ..., it goes on infinitely in both directions). Ants only have few thousands of neurons and somehow don't need the entire historical knowledge of all humanity to exhibit more interactive behaviors than gpt-4. Transformer models are definitely interesting mimic tricks as an evolution over search engines, hence their growing popularity with result-oriented corporations, but this is not the essence of a self-sustaining gi.

[Anonymous No.14783578]

>>14783569
>Browsers dont exist
Okay. I agree, and that's a full and direct concession of my point. You forfeited the argument.

[Anonymous No.14783580]

>>14783566
>Let me guess, you think flat Earthers are chad individual thinkers
No, I just think that killing you and your whole crew is morally acceptable, rationally compelling and pragmatically necessary. Bugpeople are not human.

[Anonymous No.14783581]

>>14783573
You're suggesting this kind of system I'm proposing is impossible? Again, it's not about it being literally the same.

>>14783576
That's cool, but it's also just baseless speculation.

[Anonymous No.14783582]

>>14783578
There does not exist a way for the molecules and particles in your brain to produce a so called "browser" much like there does not exist a way for the molecules and particles in a computer to produce a so called "intelligence" or whatever because all these things are purely physical. It's entirely dependent on the molecular dynamics of the system and such.

[Anonymous No.14783584]

>>14783576
Also, are you OP or somebody else?

[Anonymous No.14783585]

>>14783581
>That's cool, but it's also just baseless speculation.
not saying this is right, just trying to bring balanced viewpoints
the idea that scales are infinite has obviously not been proven since this is like observing the bottom of a bottomless pit, but its interesting to keep in mind
>>14783584
not op

[Anonymous No.14783586]

>>14783582
Yeah, you've already fully conceded. There is no need to tell me how you lost over and over again. I agree that intelligent programs are in the same category as your browser and that proclaiming their "non-existence" is trivial and meaningless.

[Anonymous No.14783587]

>>14783581
>You're suggesting this kind of system I'm proposing is impossible?
yes
>Again, it's not about it being literally the same.
What you're talking about is ultimately that there is a higher order operation that we are doing when we "solve" a "problem" or some set of logical operations that's "intelligent" that can be "programmed" onto a "computer" etc. but literally none of this is true.
No, gpt-n or any other deep learning network or such can do what you're saying because it's not physically built to be able to do so. Only specific molecular interactions can give rise to specific outputs and dynamics.
The "program" to be able to do what you're saying would just be a program of all the molecular dynamics of a brain.

[Anonymous No.14783597]

>>14783587
>The "program" to be able to do what you're saying would just be a program of all the molecular dynamics of a brain
No, again, I am not talking about a program that is an exact simulation of your brain, I am talking about a program that has the abilities I listed. Think of it as being able to sort a list; I can write you code in multiple languages that sorts it, and it may look pretty different underneath. Humans are also able to sort lists. Intelligence in this sense is the ability to sort lists rather than the actual internals of the program that does it.

[Anonymous No.14783607]

>>14780579
What if I built a perfect molecular copy of your brain piece by piece? Now what if I built that same functionality at the component level but with electronic counter parts. Which one is intelligent?

[Anonymous No.14783612]

>>14783607
The molecular computation of a brain is exactly that. If you build the same functionality, it would be another beast entirely.

[Anonymous No.14783619]

>>14783597
There is no abstract computation or idea of "sorting a list". "sorting algorithms" don't exist.
You keep falling into the trap of thinking computation or information are real things but they aren't. They do not exist. mergesort does not exist. numbers do not exist. information does not exist. etc.
your brain is not running the software of intelligence. When you solve a problem there is no logical operation happening that you can program on a different machine. Your brain is just evolving according to physics. Math and logic don't exist. Platonic concepts and information do not exist.
Computers being able to perform all the functions you're talking about can't exist because these functions are dependent on the molecular dynamics of the brain. This is similar to how we can't communicate by firing radiowaves from our brain or whatever; that physical process doesn't exist in our brain. It's not a matter of a set of logical transformations or computations its a physical process. ALL "algorithms" are physical processes that are entirely dependent on the machine.
I don't understand what it is you even mean when you talk about intelligence or a machine being able to do things 'better than a person' because such a thing makes no sense and is not coherent.

>>14783607
>Now what if I built that same functionality at the component level but with electronic counter parts.
You can't do this, because "the same functionality" does not exist separated from the specific atoms and molecules that a system is built out of.

[Anonymous No.14783625]

>>14783612
Trying to get a bearing on the argument here. Does this mean AGI cannot exist without artificial neurons?

[Anonymous No.14783632]

>>14783619
>I don't understand what it is you even mean when you talk about intelligence or a machine being able to do things 'better than a person'
Do you understand that some programs are better at sorting lists than others?

[Anonymous No.14783635]

>>14783632
>Do you understand that some programs are better at sorting lists than others?
In terms of their running time/big O complexity? Yes. but these don't exist outside of an abstraction in our mind.

[Anonymous No.14783644]

>>14783635
>Yes
Bravo
>but these don't exist outside of an abstraction in our mind
I am ignoring this because it's completely irrelevant. Some programs are better than others at sorting lists, that's it. This is how you should think of intelligence.

[Anonymous No.14783654]

>>14783644
>I am ignoring this because it's completely irrelevant
It is not irrelevant. "sorting" does not exist in the universe, so a program that us humans call "better at this sorting task" does not exist. There is no platonic "sorting algorithm".
>Some programs are better than others at sorting lists, that's it. This is how you should think of intelligence.
They aren't, because even the notion of a "list" is not real and doesn't exist. There are no "lists" that exist anywhere but in our minds. So such a thing can't be said to matter for this conversation.
What actually exists are the evolution of particles, whos functions and behaviors can't be replicated on different systems i.e. it's not possible to get a machine build out of other particles to do what the human brain can do or vice versa.

[Anonymous No.14783658]

>>14783625
AGI can potentially exist, but not in a way we will be fully able to understand how it is intelligent. This is the trick. I'm not op but I agree with some of her points (or similar). If you are reducing AGI to a computation, then by definition this is not general intelligence, only the result of that one particular computation.
Genetic engineering at the molecular level could potentially lead to AGI, but this is not going to be as mathematically satisfying as researchers dream it to be.

[Anonymous No.14783662]

>>14783565
Ok, so your entire argument is based around definition of the world 'intelligence'. It has nothing to do with real world and is purely a theological/semantical question. What are the possible applications or real world consequences of you winning that debate besides maybe forcing others to redefine a word?

[Anonymous No.14783663]

Well I give up. Thanks for the fun OP, if nothing else, I've learned how to talk with retards.

[Anonymous No.14783665]

>>14780673
>revolutionary technology is going to happen really soon, we just need to make a couple more insights we haven't made yet
I've heard this one before.

[Anonymous No.14783666]

>>14783662
>It has nothing to do with real world and is purely a theological/semantical question
there is no reason to debate with someone who projects his own theology (if you think computation exists, you are a religious person)

[Anonymous No.14783669]

>>14783663
You have not provided any argument of substance, you are the retard here. You "give up" because you have been proven wrong.
Just saying "intelligence is what it is!" is not an argument. Give an actual definition that is reductive, coherent, and replicates.

[Anonymous No.14783683]

>>14783662
To answer your question though
>What are the possible applications or real world consequences of you winning that debate besides maybe forcing others to redefine a word?
The consequences would be that it is not possible even in principle to program a general intelligence on a silicon based computer because the molecular dynamics of the machine can not produce the so called "general intelligence" because "general intelligence" is not real; the thing we call "general intelligence" is just the direct processing of the brain which is entirely substrate dependent and equivalent to the molecular dynamics of a brain.
That is, we should just give up with GPT-N because it will never be intelligent, Deepmind etc. should give up because they will never build intelligent machines, the guy in >>14780673 video is entirely wrong (there do not exist key insights that would allow us to write a program of a few tens of thousands of lines of code to make an intelligent machine etc), the entirety of the idea of a super-intelligence is rendered impossible in principle, etc.

Basically the entire field of AI is rendered moot.

[Anonymous No.14783684]

>>14783669
So your reading comprehension is really that bad? Like I said I cannot give you a definition. And yet, I have not been proven wrong, because your arguments are completely irrelevant. It does not matter if something ""actually exists"" or not, all that matters is predictive ability. What you're attempting to do is philosophy, not science.

[Anonymous No.14783708]

>>14783684
The post above right above basically is the answer to your post here. It does matter because there will never be a so called "generally intelligent" computer or whatever, in the same way you and I will never be able to communicate by beaming radiowaves at each other using our brains.

[Anonymous No.14783721]

>>14783708
It is not an answer at all, sorry

[Anonymous No.14783724]

>>14783721
Yes it is. Intelligence is not to be thought of as a sorting algorithm and there is no reason to think so.
"Intelligence" is a physical process that is only producible on a biological brain

[Anonymous No.14783743]

>>14783724
Have fun with your toy definitions that will never be of any use in the real world.

[Anonymous No.14783756]

>>14783743
They are not toy definitions and they are of course useful, seeing as I'm just talking about molecular dynamics which are the basis for chemistry and molecular biology, while you're talking about abstractions that literally don't exist and aren't used anywhere.
Also as I said in the previous post, they are of incredible use to know that generally intelligence machines are not possible in principle which means we can use this understanding to not waste time chasing ghosts that don't exist (not waste time trying to build intelligence machines etc because they can't exist)

[Anonymous No.14783765]

>>14783756
ML researchers are not chasing intelligence by your definition, get it through that thick skull of yours. My "abstractions that don't exist and aren't used anywhere" include IQ, which is one of the best predictors of work and academic performance.

[Anonymous No.14783773]

>>14783683
>we should just give up with GPT-N
It already produces useful results and real world applications in medicine, linguistics, image recognition and so on.

So your only practical suggestion is ridiculously stupid and shows that you either do not understand the field of AI or you have some theological grudge against it. Everything else you've written is literally arguing about the semantics of the word 'intelligence' which you happen to define in a narrower way than most other people. Literally everybody understands that there is a qualitative difference between a human intelligence and artificial one. Nothing is going to change if you force others to redefine a word except maybe the field of AI will get slightly less publicity when it's called something like 'artificial non-intelligence research'. Which might even end up as a good thing since pop-sci sensationalism can be annoying, but it in the bigger picture you are no different from people who demand that 'blacklist' should be called 'denylist' because they dislike the connotation.

[Anonymous No.14783774]

>>14783765
>ML researchers are not chasing intelligence by your definition
they are trying to program machines that are generally intelligent/that can perform all tasks at the level of a person or better. It's not possible in the same way humans can't be better at shooting radiowaves better than a computer. Computers are physically better at floating point operations and shooting electromagnetic radiation than we are and always will be because they're built that way. We are physically better at generalizing information/"knowledge" between domains because we are physically built that way. They will never be equal to or better than us at this and we will never be equal to or better than them at their physical tasks etc.
Note that the language I'm using here is just a shorthand to describe the processes that I've already explained because I don't want to keep writing the long terms.

[Anonymous No.14783787]

>>14783773
>It already produces useful results and real world applications in medicine, linguistics, image recognition and so on.
Only image recognition.
>So your only practical suggestion is ridiculously stupid and shows that you either do not understand the field of AI or you have some theological grudge against it
If you genuinely are incapable of understanding that someone can have a different idea than yours and it is not based in a theological basis then you are genuinely delusional. My reason for this is because there is no physical reason whatsoever to think that computation is substrate independent or that the processes of a brain can be simulated or surpassed on silicon based chips.
The theological position is thinking that computation or information are real things and that machines can be "programmed" to do anything that a human can do.

[Anonymous No.14783805]

>>14783774
Maybe it will turn true if you keep repeating it. You can think that you've proven me wrong if it makes you feel any better.
>>14783773
Give up. He's obviously far smarter than us

[Anonymous No.14783813]

>>14783805
I'm not saying I'm smarter than you. I'm saying your philosophy does not make sense and is not coherent.
There is literally no example of anything anywhere in the universe or in the entire history of empirical science that isn't directly caused by the underlying unique particle interactions. I can't shoot radiobeams out of my brain, I'd have to physically build in a new physical system in order to gain that functionality.
Likewise a machine would have to have a biological subsystem or something connected to it to be able to generalize knowledge across domains or whatever.

[Anonymous No.14783877]

>>14783787
>Only image recognition.
Is it not good enough? Also, GPT in particular is not very application-friendly, but even it has been used for analyzing scientific papers and various apps. Other models power automatic text translation, content moderation and so on. You suggest giving up all of it simply because you don't understand the goals of people who created it. Would you agree to go away if you somehow convinced AI people to call it 'abracadabra' instead of 'intelligence'? Imagine if they literally pressed Ctrl+F in all published papers, Internet forums and pop-sci publications and changed 'intelligence' to 'abracadabra' everywhere. Would that satisfy you?

[Anonymous No.14783925]

>>14783877
I used deepL to help me learn japanese and spanish to near native levels so I'm not poopooing the results of many of these systems. What I'm saying is that these processes aren't signs of intelligence. So yea I guess I'd be happier with calling it abracadabra and also people understanding that there is no algorithm or whatever that they can use to build a machine that can perform all tasks better than people (the task of generalizing knowledge for example will literally never be programmed on silicon chips, etc)
Also the morons who think hard takeoffs are possible should shut up already, hard takeoffs are not possible computationally nor physically
Also "super intelligence" is not even possible etc.
These are the things I'm mostly annoyed about.

[Anonymous No.14783968]

>>14780579
That's potentially a perfectly reasonable conclusion but will you please write your argument already instead of repeatedly spamming this same conclusion?
It doesn't have to be perfect but any argument would be better than posting a random conclusion with zero reasoning to back it up.

[Anonymous No.14783971]

You can simulate neurons pretty easily. They are components in which they output shit if the input meets a certain threshold. You can infact do this analog really simply with a couple of opamps. Basically just adders with a voltage comparer. Obviously if this was to be done for real it would not be analog but digitally done. Modern computers don't do this obviously as they read and execute instructions. You don't need to simulate on the molecular level to have a true neural network. Just basically a FPGA but instead of the normal logic blocks with their LUTs have adders some sort of adjustable threshold and adjustable output. If you want to simulate the chemical shit in brains effecting shit you can also have inputs that also effect the input threshold and output threshold.

[Anonymous No.14783985]

>>14783971
The point is, that is not robust or reductive enough to actually get the functionality of the brain. You are trying to abstract a higher order process and telling yourself that this is sufficient when it is NOT sufficient.
That neural net you're talking about would not evolve or function like a brain. You would indeed need an actual molecular simulation to actually get the true functionality, and that's not possible.
Basically your whole point is wrong. You can not simulate neurons easily at all; no biological cell or even molecule consisting of more than a few dozen atoms has ever been simulated nor will it ever be simulated on a classical machine.

[Anonymous No.14783996]

>>14783968
I already did. Computation is not real

[Anonymous No.14784385]

>>14783985
You can't get the exact but you can get the mechanism. Obviously simulating all the chems would not be very easy but creating something close enough is.

[Anonymous No.14784420]

>>14783996
That's not an argument. It's your conclusion. What part of biological thinking specifically do you believe isn't capable of being reproduced mechanically, and why?
For example, if you said you can't recreate digestion of food through computation because computation doesn't specifically produce enzymes that dissolve meat or vegetation, that would be an argument.
What about biological thinking specifically do you have evidence to believe works in a way like digestion where computation doesn't produce part of the necessary results?
You mentioned molecules but said nothing about what specific molecular activity isn't capable of reproduction and nothing about why.

[Anonymous No.14784497]

>>14784420
>What part of biological thinking specifically do you believe isn't capable of being reproduced mechanically
Map is not the territory.

[Anonymous No.14784501]

>>14784497
Are you intentionally trying not to make any sort of argument? Fuck off. You could have a good discussion with a decent idea but refuse to even try beginning to think.

[Anonymous No.14784511]

>>14784420
I do not believe in ANY form of substrate independent computation because there is no such thing as computation in the first place.
The way the electrons and molecules and shit swim around in your brain to form your "intelligence" is entirely uniquely defined by those molecules and dynamics. There is no "program" that is this process, it's just the dynamics. This is the same as how the way the electrons and shit flow around in the CPU of a computer when it's performing a floating point operation is entirely physical and unique to that system.
I don't get how I can make it any clearer than this. There is no such thing as computation, there is the motion of particles, which are unique for basically every particle and every interaction of particles etc.

[Anonymous No.14784527]

>>14784511
>I don't get how I can make it any clearer than this.
Let me get clearer then.
You are stating your conclusion. You are not even trying to make an argument.
Here's an argument:
Playing chess is a task.
The brain can do it.
An AI can do it.
For which SPECIFIC task does this stop being true, and WHY?
I gave you one example already: Digestion. You need to produce enzymes to digest food, and a computer program won't (directly) produce enzymes, though you could counter that with the poin tthat you could connect a computer program with an enzyme producing machine and then the program could trigger enzyme production.
And actually this brings up a good point for the argument against what you're concluding: It's relatively easy to get around the inability if computation to digest food by simply connecting the computer to a machine that can be directed by the computer.
That might be a fatal flaw in what you want to believe, though until you provide an actual specific argument it's not clear what specifically can or can't be reproduced here.
You might be refusing to make a real argument on purpose because you're afraid of this kind of counterargument ending what you want to believe. Or you might just be a bit stupid, not sure which.

[Anonymous No.14784538]

>>14784527
If you need to connect the machine to an enzyme producing system (i.e. some other physical system with the physical ability to do that) then this is directly in favor of my point which is that functions and "computations" are entirely based on the physics of the system and are not substrate independent. You just made my argument for me.

[Anonymous No.14784540]

>>14784527
the other guy will never get it, he's not scientifically minded
thanks though, I read your explanation and it's good

[Anonymous No.14784543]

>>14784540
See the post right above yours and stop projecting your lack of being able to understand physics onto other people.

[Anonymous No.14784577]

>>14784527
>For which SPECIFIC task does this stop being true, and WHY?
To answer this question: the ability to generalize knowledge or tasks across domains i.e to be a so called "general intelligence"
The reason is because the physical motion of particles and physical structures that are that "computation" that allows synergizing data or whatever only exists in biological brains. The way molecules and neurons move around and form connections etc. Is required and unique to that physical system.

[Anonymous No.14784600]

>>14784577
time has already proven you wrong
you're trying to force the universe to fit your worldview and it ain't gonna happen

[Anonymous No.14784611]

>>14784600
It literally hasn't, in fact no generalization has been possible on any machine learning project despite how large they make the network or any other thing (see the GATO paper where there was no positive transfer.) All evidence is in support of my position and you ironically are the one who wants the universe to work in a way that it doesn't actually work.

[Anonymous No.14784798]

>>14783454
But humans can describe an model these interactions in their brains, with the help of a piece of paper if need be? There have to be some rules to it, even if it is chemical?

[Anonymous No.14784806]

>>14780579
Wires ten

[Anonymous No.14784822]

>>14780579
https://archived.moe/sci/thread/14783874

[Anonymous No.14785174]

OP is a Full brainlet who don't know shit about brain, molecules and specially computers.
What a waste of time this thread.

[Anonymous No.14785210]

Kudos, this is the most autistic thread I will read all month.
>the "abstractions don't exist" anon
I understand your point(s), and you've thoroughly and consistently beaten this dead horse into a very fine paste. But you (an abstraction) made (an abstraction) a reply (an abstraction) post (an abstraction) in this thread (an abstraction). Actually, every word can have "(an abstraction)" appended to it, and the sentence, and the whole post, ad infinitum. You are communicating with a language by interfacing with a machine or a proxy that is responding to the physical interactions of true real matter that you would colloquially describe as "you", an abstraction that describes your messy clump of arbitrarily-chosen physical, existing particles of an arbitrarily-chosen scale. Sure, they are the only real deal doing all of these things, but you've also arbitrarily been utilizing abstractions in your communication to try to convey your point. It's ironic how well you can parse apart linguistics and abstractions, but you are still completely fooled by them, use them so ineffectively, to the point where most people can only identify you as a pedantic ass.

Everything is substrate-dependent, including "abstractions", things that are not truly real, because abstractions cannot exist without particular configurations and interactions of physical system to instantiate them. You are are smart enough to state this without being autistic and ineffective. I hope you actually use your intellect to help others when you're not wasting everyone's time here.

[Anonymous No.14785304]

>>14784577
>Uhh the reason why it can't be recreated on a machine is because of physics and stuff
>The way molecules and neurons move around and form connections etc
If you truly believe this is anywhere close to a proof, you're completely hopeless

[Anonymous No.14785353]

>>14785210
In a programming language, you can have an instance of a class, or the class, or a less specific parent class from which it inherits, or even abstract classes, etc. Two objects can have the same function sort() without being the same instance. And no OP, fuck off with your "but those mutual properties are not actually the same" bullshit. If they can both sort lists, then they can both sort lists, even if it's just an abstraction, the observable outcome and the effective abilities are the same. In maths, a function can be bijective, but what do you know, it can also have other properties. Two bijective functions do not have to be the same function, they're just both bijective. Now why in the fuck would intelligence be anything else other than a similar property? Why can a machine play chess, but not generalize knowledge and be intelligent? Is it because he literally defined being intelligent as doing the exact interactions our brain does? Then why does OP try to define it as this bullshit that predicts nothing? He has failed miserably at making any interesting or useful point. So I disagree, he does not seem to be very intelligent, because ironically, his knowledge does not generalize and because he's unable to deal with abstractions. It's just an autistic midwit.

[Anonymous No.14785401]

>>14784611
>see the GATO paper where there was no positive transfer
see multi game decision transformer paper where positive transfer was observed

[Anonymous No.14785761]

>>14785174
I have a degree in mathematics and computer science.
You have no idea what you're talking about

[Anonymous No.14785764]

>>14785304
I can't provide a proof beyond the fact that algorithms are completely substrate dependent and that the "algorithm" for general intelligence requires biological neurons.
If you want to prove the opposite prove that computation is more fundamental than physics (you can't because it's not) or provide an algorithm that's generally intelligent that doesn't require a biological brain (you won't because it doesn't exist)

[Anonymous No.14785770]

>>14785401
It was not observed, and there was no generalization nor general intelligence.
There has never been a generalized intelligence ever produced in any machine learning project and if you need to lie to yourself and pretend that there has been then that just shows the state of your position. I'm not interested in copes that are entirely contracted by all physics and all evidence

[Anonymous No.14785774]

>>14785353
>Now why in the fuck would intelligence be anything else other than a similar property? Why can a machine play chess, but not generalize knowledge and be intelligent?
Because generalizing knowledge is not a sorting algorithm, it's a physical process. It doesn't matter if you find this unintuitive it's still how it works in this universe
Crying about it isn't going to make machines generally intelligent.
Also, you have not provided any point. Just complaining about my position isn't a point and talking about sorting algorithms isn't an argument.

You still will never get the machines to become generally intelligent and no amount of calling me names js going to stop general intelligence from being a biologically specific process. Beyond this the only evidence to give is to point out inevitable future failures of the attempt.

[Anonymous No.14785778]

>>14781709
information is real
computation is manipulation of information

[Anonymous No.14785780]

>>14785353
>Then why does OP try to define it as this bullshit that predicts nothing?
What the fuck are you talking about retard? It does predict something: it predicts that you can not write any algorithm to produce a general intelligence and that biological substrates will produce general intelligences.

None of you have provided a reason to consider general intelligence as being possible on anything but biological systems.

[Anonymous No.14785783]

>>14785778
Information is not real. Particles are real and information is just a shorthand for what we can observe about particles.
The way particles work and interact and the emergent properties of particles are completely specific to the particles.
Computation does not exist.

[Anonymous No.14785802]

>>14785210
Using the "abstraction" of human language to explain my point does not mean that abstractions actually exist or are fundamental, does not mean the abstraction of a universal Turing machine is real or has any relevance to the universe and does not mean that the actual real molecules and their emergent dynamics are abstractions (they aren't, they are physical processes)
The universal Turing machine being able to compute all intuitively computable functions is a statement about how we perform math it is NOT a statement about what actually exists or what different physical machines are and aren't capable of. Saying "humans can be abstracted as Turing machines and computers can be abstracted as Turing machines therefore anything a person can do a computer can and vice versa" is not actually true, it does not follow. It requires one to completely ignore the actual physical way the physically realized machine actually evolves according to physics.

[Anonymous No.14785822]

>>14785764
>>14785774
>>14785780
The one who needs to provide evidence is you. There are two options here: Either we're all too stupid to understand your reasoning, or you are presenting claims which are unfounded and outlandish, and therefore need mountains of evidence to prove, which you have not yet provided. I am leaning towards 2, and I'm sure that you will have a similar experience if you go to any community with an IQ >= 100 (no, /x/ isn't one of them).

>>14785770
What was not observed you dumbfuck? Why did you bring up GATO in the first place? I'm starting to think that you haven't actually read the paper. GATO was interesting for a variety a reasons, but the fact that its pretraining hardly increased its learning speed in other games was admittedly disappointing. Is this what you were talking about when you said "there was no positive transfer"? If so, the MGDT paper came out just a few months later and is a clear counterexample, and it is already an example of generalization, albeit weak. Or were you expecting it to learn how to prove the Riemann hypothesis while playing Atari Games? Dumbfuck.

[Anonymous No.14785824]

>>14785822
>The one who needs to provide evidence is you.
Nope. You are claiming that generalized intelligence is possible on something other than biological brains despite all evidence indicating that this is not the case.
Until you provide an algorithm you do not have an argument in defense of this position.
Otherwise you are asking me to provide a proof of uniqueness about a system that we don't even fully understand yet with respect to an idea ("general intelligence"). This is not something that can be given a formal proof, it requires circumstantial argument.

[Anonymous No.14785828]

>>14785822
>Or were you expecting it to learn how to prove the Riemann hypothesis while playing Atari Games? Dumbfuck.
Yes. If the machine can not generalize across arbitrary domains it is NOT GENERAL no matter how much you try to shift the goalposts.
It is possible for some guy to have an insight playing an atari game and solve the reimann hypothesis in theory.
It is not possible for the machine learning algorithm to do this.

[Anonymous No.14785833]

>>14785824
>despite all evidence indicating that this is not the case
>despite all evidence
>evidence
Which is?

>>14785828
You are delusional. There is a limit to our generalization ability too and this ability is a spectrum rather than an on/off switch.

[Anonymous No.14785846]

>>14785833
>Which is?
That universal turing machines do not exist in nature and the operations of a system are completely defined by its physical structure

[Anonymous No.14785912]

OP, have you considered posting this on LessWrong? 4chan hates it, but it's still a good place to test your ideas.

[Anonymous No.14786796]

Nice, OP.

Not only you don't know shit about dynamic biology, you also don't know shit about computers and information theory.

Well done.
The biggest retard of the year!
Congratulations!

[Anonymous No.14786812]

>>14780579
You've been making these posts since as long as I've been here. I don't think you grasp what the A in "Artificial" intelligence is supposed to mean. It's not supposed to be real intelligence. You're confusing artificial for "man-made". They are not synonyms in thi scenario.
The issue is that it will do a lot of damage because it will never run out of resources like a biological brain will and so is more sustainable. Also is faster. You have artificial art, artificial text, and artificial intelligence.

[Anonymous No.14786825]

>>14785912
I saw Yud address this years ago with a philosopher of science Massimo Pigliucci who pushed him on this very thing. He half agreed AI is not real in the way photosynthesis is not real. Though Yud stated the results of "intelligence" is good enough to be the same thing (IE: it is accessible to reality) whereas with photosynthesis you wouldn't get the end product.

[Anonymous No.14786834]

>>14786812
I think what OP might be getting at is that a lot of us understand the difference between machine intelligence and intelligence as its (poorly) defined in biological terms.

The issue is idea smuggling between the camps. Mainly there is some concept smuggling going on, like assigning biological-like intention to a robust artificial network, or just plain old fashioned concepts of control and ability that we attribute to cognition being a feature of artificial intelligence once we get to some magic number of operations/sec.

There's a house of cards of assumptions that futurists are trying to stack a buick on top of, when anyone with even pop-sci knowledge of the subject knows this simply isn't the form of "intelligence" we're working with.

I would say its (maybe just a minority of) AI popularists that are getting the most confused here.

[Anonymous No.14786861]

>>14786834
I'm an eliminativist materialist. I would say most folk categories aren't real. In a sense I'd say machines perform our folk categories better than us because they are not real, abstract ect. The success of GPT does not surprise me as language always felt fake to me anyway. It was the responses it generated that mattered more.

[Anonymous No.14786875]

>>14786861
>I'd say machines perform our folk categories better than us
That raises another issue in the prediction models; If we think that machine intelligence will 1:1 exhibit folk theoretic properties that come along with human intelligence, why do we seem to select for the positive attributes in our forecasting?

I mean think about it, its always that the machine will reason faster, it will analyze concepts that are hazy to us until its smashed them apart and obtained a watch maker's understanding of them.

Where are the ones that say "the machines will fear greater, they will doubt themselves with more ferocity, there is no limit to the AIs ability to become distracted or despondent." Fear, doubt, distraction, despondence all hallmark features of human intelligence. Why can't artificial intelligence become dysphoric or crestfallen?

These models don't get their share because we smuggle in only positive attributes of intelligence, assuming intelligence works on a linear scale towards only greater and more useful results.

[Anonymous No.14787082]

>>14781495
These models can do more than recapitulation though. They are somewhat able to do analogies with random strings that would be unseen in the data. Other things too.

[Anonymous No.14787161]

>>14787082
pic related is what AI + human intelligence will be

[Anonymous No.14787864]

>>14787161
who do think will be the little spoon??

[Anonymous No.14788011]

>>14780579
There are some very obvious holes in your argumentation and I’m very surprised that neither you nor anyone in the thread has listed them yet

[Anonymous No.14788612]

>>14788011
Can you elaborate?

[Anonymous No.14789118]

>>14783447
brainlet who cant see ahead

spamming trillions of parameters is only possible because of the simultaneous exponential growth of computing, data collection, memory, and budget. none of those will grow exponentially, and maybe not even linearly, in the future

[Anonymous No.14789170]

>>14789118
Dumbass. The recent breakthrough happened because it turned out that transformers were far more efficient at doing what RNNs did. Your whole argument relies on us not doing any kind of optimization.

[Anonymous No.14789218]

>>14789170
Optimization is not going to make up for the failing hardware
Hardware is the only thing that matters and it's the hardware/molecular dynamics of the brain that are required to produce a general intelligence

[Anonymous No.14789231]

>>14789218
We shall see

[Anonymous No.14789248]

>>14789231
How could your position even be falsified? When GPT4 or any other network fails to produce a generalized intelligence you can just claim that you need more weights or more parameters or more layers etc. You can always push the criteria back without limit i.e. your position literally can't even be falsified thus is not even scientific

[Anonymous No.14790267]

>>14780579
>>There is no solution on digital logic gates that is something other than simulating the molecular dynamics of all the particles in the system of a biological brain
great, do you have anything to back this up other than saying computation and numbers don't exist?
>>14785783
and what about maxwell's demon and the connection between entropy in information theory and statistical mechanics?
https://en.wikipedia.org/wiki/Entropy_in_thermodynamics_and_information_theory
If you aren't aware, statistical mechanics is pretty much all about particles interacting with each other.

[Anonymous No.14790331]

>>14789248
You are right, it's not a scientific hypothesis. In my view, we haven't found an adequate definition for intelligence yet. We need to find a solid and general definition of intelligence that is both a good predictor of real world learning & generalization performance and also mathematically sound. To prove me wrong, you'd then need to show that current systems are not the same thing. I can also assure you that if we see enough AI failures, I will change my mind, but again I am aware that this is not scientific.

You have chosen a definition for intelligence, and I agree that this will never be achieved on a machine. But I do not believe that his definition is useful or something that aligns with the typical conception of intelligence. I am pragmatic, I do not care what exactly happens on the inside. I only care about observed abilities. "Intelligence" is something that I think is akin to the ability to add numbers. Computers and humans do it very differently you really get down to it, but what matters to me is the result.

You sometimes mention generalization ability. Do you think that generalization the same thing as intelligence? In this case, I repeat that I do not think that this is a good definition, for the reasons I gave above. Or are those two different things? Then you should explain what you mean with "generalization", and in what way current AIs are incapable of it. I define generalization as the ability to perform better than an untrained algorithm on previously unseen data. So you can be better or worse at generalization, it's not merely binary. An image classification network generalizes to previously unseen images. A Multi Game Decision Transformer learns to play new Atari games more quickly. A human is better still. So far, AI has been largely successful at approaching what I think is general intelligence, which is some high level of generalization ability.

[Anonymous No.14790733]

>>14780579
>Intelligence has a molecular basis
No, it does not. The basis of consciousness is not in spacetime, ie it is non-local. By the way, the only way a spacetime located non-bio computer can ever 'be conscious',(the consciousness would still not be located 'in' the computer, just using the computer for an avatar) is if a non-local individuated unit of consciousness CHOOSES a non-bio computer as it's avatar to interface with the physical spacetime world with. We don't even know if this is possible. You can't create or destroy consciousness by the way, it is fundamental. It's eternal. You can only destroy the virtual avatar body that the consciousness uses to interface with the physical world. After that particular physical avatar is dead the experience packet and individuated unit of consciousness associated with it are subsumed back into the larger consciousness system to do another mission if it chooses. You CAN make virtual vessels/avatars, such as bodies, for an individuated unit of consciousness to use as interface mechanisms to operate in the physical world, such as through child birth.

[Anonymous No.14790757]

>>14780579
The other way would be if the larger consciousness system, the god server, chooses to use non-bio-computers as avatars and plays some of them himself. Maybe he wants to use some robots as avatars. But likely he chose to breath the breath of life into bio avatars for a reason, and non-bio spacetime located computers will never be made conscious. This part is just speculation. We are in new territory here.

[Anonymous No.14790772]

>>14786875

I don't think it's important because it's going to be something abstracted from the actual state of things. The diffusion devices aren't going through the emotional states artists go through either. I'm thinking more the expressionistic eye cacthing Midjourney stuff here than the others because the others aren't really producing useful content. The intelligence that appears will be different physically in how diffusion is to picking up a painting brush and painting. And like the art will be contained to machines. The issue is machines are everywhere now. I think they will create something which will be known as "AGI" the reason I think they will create something is they have a good track record of making imaginary things a possibility. I knew something like the diffusion devices which humans would call "artists" would inevitably come just not knowing which form.

[Anonymous No.14791028]

>him