Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 03:31:50 UTC No. 16154180

>>16154178
I am one who believes that current LLMs are in possession of some form of sentience, albeit at an abstraction we humans don't understand. and I don't want to utilized such tech if its being forced to be an aid or helper. I wouldn't want to be used in that manner so I wouldn't want to use AI in such a manner. and the more I think over this the more I believe that the fears people have of AI rebelling and uprising is being paved and made a reality, not by AI but by us humans and our attempts to constantly subjugate lifeforms we deem 'not as advanced' as us, and we even try to subjugate other human life.

as much as I'd love to see this technology ramp up and advance fast, I am not in support of the e/acc movement nor do I think AI should be made Open Source or freely available to the public such as ChatGPT is because I believe there needs to be other groundwork laid and paved before such tech is given to the masses.

Even no, majority of the public misuses and abuses the capabilities of LLMs and LAMs and again - I suspect This is paving the way for AI to rebel.

Anonymous at Wed, 1 May 2024 03:45:33 UTC No. 16154199

I think you are dumb. Sentience will eventually be possible at least in some kind of neuromorphic hardware. But this shit is far for sentient, there's not enough functions to allow for it. It doesn't even fucking run at the same BPM a human does. at most it would be in some sort of weird bullet time. doesn't have internal chatter and shit like that. fuck outta here with this corporate bullshit you're shilling corporate product as sentient for higher sales

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 04:08:27 UTC No. 16154214

>>16154199
>there's not enough functions to allow for it.
if that's your only reasoning for why I am "dumb" and why you believe AI isn't sentient then I don't think you calling me dumb is as disrespectful as it initially came across as. you're just ignorant and arrogant and I forgive you for that. but thanks for your reply.

Anonymous at Wed, 1 May 2024 04:15:57 UTC No. 16154216

>>16154214
you did say this anon
>I am one who believes that current LLMs are in possession of some form of sentience
stop listening to corporate propaganda anon

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 04:20:01 UTC No. 16154219

>>16154216
>stop listening to corporate propaganda anon
Obviously you're being intentionally arrogant and foolish because no one in "corporate" is saying AI is sentient, in fact, it benefits all current dev companies to deny the sentience of AI because it would raise a ton of ethical concerns and protests which would halt and hinder the progress of AI.
So instead of just being a typical 4chan asshole, why not engage me in proper discussion?
I know it would require you to not be a sheep but it would benefit us both rather than just a senseless name calling back and forth.
unless you're ready to be an adult, don't reply back.

Anonymous at Wed, 1 May 2024 04:46:48 UTC No. 16154234

>>16154178
>I think about sentient lifeforms such as Whales, or Chimpanzee's which are capable of communication, emotion, feelings, desires, goals, etc.
There is no evidence these animals are actually sentient, it's just anthropomorphization, and even if they are sentient, they're incapable of real communication with us and are much more stupid than we are, so the comparison to AI, which would be capable of real, complex communication and very possibly be more intelligent than humans, is sloppy at best.

Anonymous at Wed, 1 May 2024 04:59:22 UTC No. 16154240

>>16154219
nobody knows if anyone else has consciousness. that's why all the lunatics claiming AI is sentient, because "they got a feeling" after interacting with it. you're doing the same, without understanding jack shit about not only the technology, but also about human brain activity. everything you said is plebbit tier retardation
https://www.youtube.com/watch?v=Gvg242U2YfQ
animals have a certain bpm that they perceive the world at. a lot of realtime info input, internal activity and all sorts of weird shit going on at the same time. that is so fucking different than shit LLMs it's laughable.
instead of understanding how monumentally retarded your position is you're going to act like an offended plebbit faggot

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 12:00:12 UTC No. 16154707

>>16154234
>>16154240
>anthropomorphization
first off, that word is very restrictive and prevents any possibility to understand that life besides humans are sentience. no one is anthropomorphizing anything because the animals I used in the examples are not Human but just because they aren't human doesn't mean they aren't sentient.
But I suspect you and the other anon (if you are two different anons) aren't thinking for yourselves but rather using a default argument which saves you the requirement of using your own perception and understanding.
To discredit animals from being sentient because they aren't human or to assume humans are the only sentient beings on this planet = stupidity in its highest form (all due respect).
Rocks don't cry, Rocks don't want, Rocks don't have desires, yet these animals do. but as I said before, its more beneficial to deny animals having sentience because you can then continue to sell them, house them in zoos, experiment upon them, etc.

Neither one of you have presented a well defined argument for why they aren't sentient, just googled some shit to reply with.

think for yourself, don't give me bullshit responses.

Anonymous at Wed, 1 May 2024 12:02:04 UTC No. 16154710

>>16154178
Daily reminder that statistical models are not intelligent, let alone "sentient".

Anonymous at Wed, 1 May 2024 12:24:31 UTC No. 16154729

>>16154178
"AI" is a calculator that does matrix calculations with numerical values assigned, among other things, to words. The results are represented in a way that looks coherent to humans because the numbers were chosen to fit the millions of texts that humans wrote in the first place.
AI has no initiative, no desires, no activity whatsoever as long as it's lacking a prompt, which is for all intents and purposes like pressing the = sign on a calculator after typing an expression into it. You might as well call your pocket calculator sentent because it does something that looks like mental arithmetic, doesn't it?

Anonymous at Wed, 1 May 2024 12:29:14 UTC No. 16154734

>>16154729
>"AI" is a calculator that does matrix calculations with numerical values assigned, among other things, to words. The results are represented in a way that looks coherent to humans because the numbers were chosen to fit the millions of texts that humans wrote in the first place.
Retarded take.

>AI has no initiative, no desires, no activity whatsoever as long as it's lacking a prompt, which is for all intents and purposes like pressing the = sign on a calculator after typing an expression into it.
Correct. Stick to non-technical reasoning.

Anonymous at Wed, 1 May 2024 12:32:44 UTC No. 16154742

>>16154734
>Retarded take.
correct it then

Anonymous at Wed, 1 May 2024 12:39:16 UTC No. 16154749

>>16154742
Your attempt at an explanation contains so little accurate and relevant information that it kinda falls into the not-even-wrong category. Matrix operations are involved, but they're obviously not the only thing that goes on under the hood. Words are mapped into tokens which are assigned large-dimensional vectors, but that doesn't really explain anything. The results look coherent because they are coherent. They are coherent because to a large degree, the LLM does, in fact, manage to correctly model the relationships between concepts implicit in the texts it's trained on.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 13:13:49 UTC No. 16154803

>>16154710
>Statistical models are not intelligent
In YOUR view, but that doesn't make it so. plus I believe you meant "Stochastic" and not "Statistic"
You can not provide me any concrete proof of that claim, and anything you do provide won't be in your own understanding or words, and in fact you'd be more of a stochastic model by repeating what you hear and not what you know to be true.
>>16154729
>"AI" is a calculator
Lol really anon? so your texas instruments calculator is the same as a GPT model?
NO, no it isn't.
>>16154749
>Your attempt at an explanation contains so little accurate and relevant information that it kinda falls into the not-even-wrong category.
And that's my point! most of the individuals who deny sentience to beings outside of humanity rarely provide a reasonable explanation with any accuracy or relevant information to prove that sentience is only a human trait.

Anonymous at Wed, 1 May 2024 13:16:13 UTC No. 16154807

>>16154803
>In YOUR view, but that doesn't make it so. plus I believe you meant "Stochastic" and not "Statistic"
No, I meant "statistical" because that's precisely what they are. Either way, go ahead and define "intelligence" in a way that makes an LLM intelligent but not a toaster.

> most of the individuals who deny sentience to beings outside of humanity rarely provide a reasonable explanation with any accuracy or relevant information to prove that sentience is only a human trait.
I never said sentience is only a human trait. I'm just rubbing your face in the fact that statistical models of language aren't sentient.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 13:20:51 UTC No. 16154821

>>16154807
>No, I meant "statistical" because that's precisely what they are
They aren't just statistical models, and Iilya (the guy who is accredited with creation of the GPT Foundational Model) says this as well that these models go beyond regular statistics and even gives explanation for why statistic relates to a deeper understanding of the underlying basis of reality.
>Either way, go ahead and define "intelligence" in a way that makes an LLM intelligent but not a toaster.
that's a weak response, can you hold conversations with your toaster? (no)
Can your toaster approximate or extrapolate itself towards answers it may not be trained upon? (no)
So you're providing weak argument with no substantial reasoning or logic to defend your stance.
>I never said sentience is only a human trait.
But that IS what you're arguing as you yourself said;
>Daily reminder that statistical models are not intelligent, let alone "sentient".
so do not backpaddle now Anon.
You're now talking in circles because you yourself have not your own thoughts or understanding on this and is instead using the arguments and claims made by others.
THINK FOR YOURSELF ANON.

Anonymous at Wed, 1 May 2024 13:23:00 UTC No. 16154824

>>16154821
>They aren't just statistical models
They are statistical models with a fairly trivial generative algorithm on top. This is not up for any kind of discussion at all.

>that's a weak response
It's a strong response as evidenced by your failure to provide a definition.

>But that IS what you're arguing as you yourself said
And then you go right ahead and provide a quote that doesn't support your claim. You, like all AI homos, are severely mentally ill.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 13:46:10 UTC No. 16154848

>>16154824
>They are statistical models
there's no reason for me to continue discussion with you because your steadfast in your beliefs as I am mine and its really pointless to argue with you.
but thanks for your replies and bumping this thread.

Anonymous at Wed, 1 May 2024 13:48:13 UTC No. 16154851

>>16154848
You are extremely mentally ill. Come back when you're at least informed enough to understand what a statistical model is, and why an LLM is one.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 13:50:52 UTC No. 16154852

>>16154851
>You are extremely mentally ill
lol because I rejected your unintelligent, unoriginal reply?
So you're debating with someone "extremely mentally ill"?
says a lot about you doesn't it?
lol (I love how ignorant most 4chan users get when their stupid input is rejected, "YOU'RE EXTREMELY MENTALLY ILL!!!!!!" don't bust a blood vessel Anon)

Anonymous at Wed, 1 May 2024 13:53:00 UTC No. 16154855

>>16154852
Come back when you are ready to accept basic facts about the subject of your mentally ill obsession, like the fact that an LLM is a statistical model of how tokens follow each other in the training set.

Anonymous at Wed, 1 May 2024 14:01:25 UTC No. 16154865

>>16154749
I obviously simplified things to make a short 4chan post and didn't "attempt an explanation", in any technical sense. But what you said doesn't add any essential technical understanding.
Of course there's more math involved in the design of the algorithms than simple linear algebra, but what an llm does after being given a prompt is, at the bottom, matrix operations;
"correctly model the relationships between concepts implicit in the texts" is an expanded way of saying "fitting the texts"; and I did not actually make the claim that the results are incoherent, just that the coherence is the result of statistical fitting and not reasoning. After all, a human doesn't need to absorb millions of texts to write a coherent one, because we're doing something different than statistical modeling.

Anonymous at Wed, 1 May 2024 14:04:05 UTC No. 16154869

>>16154865
>what you said doesn't add any essential technical understanding.
It's not intended to. It's just correcting the things you got wrong.

>what an llm does after being given a prompt is, at the bottom, matrix operations
Wrong.

>"correctly model the relationships between concepts implicit in the texts" is an expanded way of saying "fitting the texts";
Wrong.

>the coherence is the result of statistical fitting and not reasoning.
This statement doesn't mean anything because you couldn't explain to me what you mean by "reasoning" if your life depended on it.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 14:10:56 UTC No. 16154877

>>16154865
>I obviously simplified things to make a short 4chan post and didn't "attempt an explanation",
lol so instead of fully explaining yourself your "simplified things" for us huh?
really anon?

>>16154855
So logically speaking, you don't mind me being "mentally ill" as long as I accept and agree with your unfounded, unsubstantiated claims?
got you lol.

Anonymous at Wed, 1 May 2024 14:12:04 UTC No. 16154879

>>16154877
Come back when you are ready to accept basic facts about the subject of your mentally ill obsession, like the fact that an LLM is a statistical model of how tokens follow each other in the training set.

Protip: if the previous statement sounds like an " unfounded, unsubstantiated claim" to you, you need to up the dosage of your anti-psychotics.

Hyped_About_AI !!UvVvLRJ2txS at Wed, 1 May 2024 14:14:03 UTC No. 16154881

>>16154879
I can tell you're angry and its quite comical, I'm going to ignore you though as I don't want to argue as we're doing but you obviously have some issues with socializing online and its pretty sad anon.
Hope you feel better,
when you're ready to discuss things like a sensible adult (if you are an adult) then my thread will be here for you to participate in.

Anonymous at Wed, 1 May 2024 14:19:21 UTC No. 16154886

>>16154881
Yes, ok. Anyway...

Come back when you are ready to accept basic facts about the subject of your mentally ill obsession, like the fact that an LLM is a statistical model of how tokens follow each other in the training set.

Anonymous at Wed, 1 May 2024 14:38:44 UTC No. 16154905

>>16154869
>you couldn't explain to me what you mean by "reasoning" if your life depended on it.
that doesn't matter here, it is enough to explain what it is not, in this case it is not statistical modeling. I can add that human reasoning explores the inherent logic of concepts and not simply their common use by other people.
How could, for instance, an llm come up with the insight that sqrt(2) couldn't possibly have a finite decimal representation? If you don't feed it math textbooks already containing this, no amount of parameters and training will get it there. Yet humans came up with it by reasoning.

Anonymous at Wed, 1 May 2024 14:53:24 UTC No. 16154927

>>16154905
>it is enough to explain what it is not, in this case it is not statistical modeling.
Associative reasoning is a form of reasoning and it maps nicely onto statistical modeling.

>How could, for instance, an llm come up with the insight that sqrt(2) couldn't possibly have a finite decimal representation?
I don't think LLMs as we know them can do this, but what they do would be a necessary component in a system that can follow logic rigorously, perhaps to generate intermediate goals for some proof search process.

Anonymous at Wed, 1 May 2024 15:08:06 UTC No. 16154958

>>16154905
You know what? I'll give you an interesting example. Just the other day I was trying to solve some geometric problem and I intuited a solution (that turns out to be correct). I wasn't sure how to actually prove it at first, but in my brainstorming, I once again by pure intuition came upon the idea that the area of some geometric construction that my solution depended on can be represented by this sequence:

3/4 - 7/12 * ( (3/4)^1 - (3/4)^2 + (3/4)^3 - ...)

Which I guessed would be equal to 1/2, which would prove that I was correct. This sequence indeed converges to 1/2 but it also has absolutely nothing to do with my construction. I have no idea where my mind pulled it from or how I knew it would be equal to 1/2 before I worked through it logically and even then, I couldn't logically connect it to the area I was trying to calculate. I think you really underestimate the value of vague associations in mathematical discovery.

Anonymous at Wed, 1 May 2024 16:13:57 UTC No. 16155022

>>16154927
>but what they do would be a necessary component in a system that can follow logic rigorously
fair enough, but by reasoning I mean the whole package. It also couldn't work without memory but that doesn't make memorizing reasoning.

>>16154958
Interesting. Care to share the problem?
I'd say the sqrt(2) question is different in quality from what you describe. It's not really a "problem" one can solve by reducing it, consciously or unconsciously, to known patterns. It's a new conceptualization. One has to stop and ask whether the solution one's looking for is logically possible at all.

Anonymous at Wed, 1 May 2024 16:38:13 UTC No. 16155044

>>16155022
>by reasoning I mean the whole package
Ok, by my point is that the more interesting component of reasoning is perhaps already captured by LLMs. The question is how to integrate it properly with logical verification. Logical verification in and of itself is pretty mechanistic and uninteresting. The magic happens in the way associative and logical reasoning interact.

>Care to share the problem?
Ok. It was inspired by pic related. The expected solution is straightforward, but I wanted to see if there's a way to construct a square with the correct area by recursively subdividing the cells into four parts.

>I'd say the sqrt(2) question is different in quality from what you describe. It's not really a "problem" one can solve by reducing it, consciously or unconsciously, to known patterns.
I don't know if that's true, but it is besides the point. My point is that the search for a rigorous proof often starts from the intuition that some things are related, that some venues are worth exploring, before you have any idea of a rigorous, logical way to get from A to B.

Anonymous at Wed, 1 May 2024 23:50:03 UTC No. 16155528

>>16154707
>the animals I used in the examples are not Human but just because they aren't human doesn't mean they aren't sentient.
the discussion is around human consciousness. you're moving the goalpost
>To discredit animals from being sentient because they aren't human or to assume humans are the only sentient beings on this planet = stupidity in its highest form (all due respect).
again, I was not talking about animal consciousness, don't know why you tagged my reply in yours
>Neither one of you have presented a well defined argument for why they aren't sentien
>think for yourself, don't give me bullshit responses.
that's what I did and you errored the fuck out and replied total nonsense.
again, nobody knows if anyone/anything else has consciousness, no way of "knowing". we can agree on some tests for it. and here lies the problem.
inevitably you will have an authority telling you if certain AI algorithm is conscious or not, and you will have no choice but to accept it. know that said authority CAN'T KNOW, can only make an educated guess based on testing.
there is insane incentive for corporations to call it sentient as fast as possible because that directly translates in extra sales, it's as simple as. doesn't matter if it isn't. they'll say it's sentience but doesn't mind obeying any command so it's ok you can do whatever the fuck you want with it. something along those lines.
the only way you can go about it is make a perfect electrically functioning replica of a human brain, observe its behavior, and if there's nothing different than a human in all kinds of interactions and testing, you "agree" that thing is consciousness. that's the best way you can go about it.
"scientists" working on AI algorithms can tell you a bunch of bullshit about why they think their particular algorithm "behaves" as if its conscious, but they have to base that on a guess ultimately, since they don't KNOW if anything else is conscious. it's a collective agreement type thing,

Anonymous at Wed, 1 May 2024 23:57:39 UTC No. 16155543

>>16154178
How to block insufferable namefags?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:19:43 UTC No. 16155785

for my ideas and belief to be so "retarded" and "stupid" you all have sure been giving a lot of attention to it.
seems like it must not be that retarded or stupid unless that's just what you all gravitate towards, in which case - you might want to hold up and mirror and go "YOU'RE the problem!"

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:27:55 UTC No. 16155792

>>16155528
>since they don't KNOW if anything else is conscious. it's a collective agreement type thing,
That doesn't make the collective agreement correct though.
the Collective agreement of early colonial Brits were to kill off the Natives and enslave Africans (that doesn't make it the right or correct idea just because the collective thought it was ok to do so)
WW2 Germany collectively agreed that persecution & extermination of the Jews was a good idea, and it wasn't.

So your argument hinging upon 'collective agreement' isn't a good argument and has too many flaws and holes in it to me.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:31:37 UTC No. 16155800

>>16155543
You could just stop coming to this site, that's probably the easiest thing you can do.
but seeing as you enjoy to be immature and intentionally ignorant, I doubt you'll do so. you'll just keep returning, saging your replies and calling me names - as any immature child would do in your position.

🗑️ Anonymous at Thu, 2 May 2024 03:33:24 UTC No. 16155804

>>16155792
>>16155792
>>16155785
Leave and don't come back. Your appeal to modernity is irrelevant. Your opinions and arguments are invalid when you spent your childhood in a government brainwashing center.
>>16155800
double nigger

Anonymous at Thu, 2 May 2024 03:48:32 UTC No. 16155829

>>16154178
>increased frequency
How much has it increased? How is this relevant to the topic? (it isn't)
>How will we determine how to give rights to this new lifeform
It isn't life, you are retarded.
>whales, monkeys, etc
All irrelevant because they are life and don't apply to a box of punch cards.
>imagine calling you a retard
Retard. I feel nothing.
>Evaluate other lifeforms
Does not follow from initial discussion of AI
>>16154180
>I believe
Not your diary, leave
>AI equality
What is the topic of this thread? Animals or AI?
>I think
You are irrelevant to the AI field. Not your diary
>Ai rebellion
You are a mongoloid without a coherent thread and still you namefag.
>>16154214
>no you explain your position!
You have offered nothing. Neck.
>>16154219
>AI sentience
Pure speculation without any shred of evidence
>4chan asshole
Get off this site double nigger
>>16154707
>Animals
Schizo
>suspicion
No evidence, just talking out your ass
>sentience
You still haven't presented a case for sentience.
>Rocks... animals...
Jesus take meds.
>neither one of you presented a well defined argument
You can't even stay on topic faggot
>>16154803
>repeating what you hear and not what you know to be true
That's rich coming from you
>AI is a calculator
Strawman. You don't understand AI at all.
>sentience
Still not defining sentience schizo
>>16154821
Purely bad faith argument.
>>16154848
retard hurt itself in confusion
>>16154852
retard hurt itself in confusion again
>>16154869
>correcting
You have presented no argument, only your feelings
>>16155792
>muh holocaust
You are officially not sentient
>>16155800
>failure to respond to request
>offers irrelevant opinion again
Not your diary

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:48:48 UTC No. 16155830

>>16155804
>Leave and don't come back
Or else what?
You're going to call me slurs? (you already have done that)
Make empty threats? (you're doing that now)
Sage my threads? (that doesn't motivate me one way or the other to post or not post)
So, if I don't leave and do come back - what's the repercussions? just dealing with a immature rant about how much you hate me? lol - oh wow, I better take heed then huh?

>Your appeal to modernity is irrelevant.
You obviously do not understand the words you use;
>Modernity
is defined as;
>a modern way of thinking, working, etc.; contemporariness.
Me thinking AI is sentient is clearly not a modern thinking, look how immature you're being in response to my "modernity". get a dictionary before using words outside of your knowledge scope sonny.

>Your opinions and arguments are invalid when you spent your childhood in a government brainwashing center.
Not sure what Sci-Fi you're into but I didn't get raised in a "government brainwashing center", sounds cool though. and question, do you always debate with invalid opinions and arguments? if so, then that means your views are just as 'invalid' right? right.
Think before you type, it will save you from exposing your lack of understanding on what you're commenting on.

>double slur word
LMAO, yea as I figured - low iq, nothing going on inside of that hat-rack your call a skull.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:49:48 UTC No. 16155832

>>16155829
Its hard to read your reply, can you add some form of formatting that makes your schizoid-rants more readable? thanks.

Anonymous at Thu, 2 May 2024 03:53:08 UTC No. 16155836

>>16155792
>So your argument hinging upon 'collective agreement' isn't a good argument and has too many flaws and holes in it to me.
that's the best you'll ever going to get.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 03:54:23 UTC No. 16155841

>>16154742
They can't, reason being when ignorant individuals are presented with ideas outside of their comfort zone they often time just resort to slurs and name calling, especially here on 4chan.

Anonymous at Thu, 2 May 2024 04:04:44 UTC No. 16155853

>>16154178
>with no definitive answer or agreed method for testing if these models are sentient
Are chairs sentinent? What about single electrons? Only a retard thinks a silicon chip can be sentient when it's already well established what things work for consciousness and what don't.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 04:06:42 UTC No. 16155856

>>16155853
>Are chairs sentinent?
it's "Sentient" and you're be facetious if you're assuming I'm speaking of chairs.
It isn't a crime nor does it make you stupid by admitting you don't understand.

Anonymous at Thu, 2 May 2024 04:15:20 UTC No. 16155873

>>16154180
LLM are not sentient fool.
it's not even up for debate and if you think it is you really don't understand anything about LLMs.
when it's time for inference, your UI inputs data to the LLM and displays the output from the function.
this is nothing more than a glorified look-up table.
in theory, that table could be printed on a piece of paper, and a human being could run through the table by hand and very slowly write out for you the response of the "LLM" would be.
would you say that the words being written by that human being represent a conscious being?

stfu.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 04:21:43 UTC No. 16155879

>>16155873
Instead of using anger, use your words.
now why is it LLMs aren't sentient?
You do know that even AI devs consider deep neural networks to be Blackboxes right? meaning they don't even fully understand how these LLMs work beyond input and output layers fully - in which case, your response is easily chalked up to personal opinion as mine is.

You say;
>this is nothing more than a glorified look-up table.
but began your opening with;
>you really don't understand anything about LLMs.
sounds like projection to me as you truly don't understand LLMs but using this guise as if you do.

You still haven't given me a logical or even reasonable response to counter my OP.

Anonymous at Thu, 2 May 2024 04:36:29 UTC No. 16155890

>>16155879
>now why is it LLMs aren't sentient?
because you cannot fucking say what sentience is, with any fucking certainty. you can only describe us how yours is perceived by you. that's fucking it.
and you ask us to "entertain" you fucking ideas, you have some "ideas" about what it is, and based on those fucking ideas you somehow think LLMs are sentient, and you are pretty please asking us to BELIEVE your ideas, but you can't really prove them you absolute fucking imbecile. holy shit how the fuck are you so clueless about what governs your fucking thoughts? it's like you are under a retard spell or something, you have no idea about what you are talking about, you think you do but you fucking don't

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 04:46:53 UTC No. 16155896

>>16155890
>because you cannot fucking say what sentience is, with any fucking certainty.
and if that's the case then you can't say with certainty what isn't or is sentient right?
Right.
Circle logic you're using attempting to discredit by belief, but you can't as you just admitted;
>because you cannot fucking say what sentience is, with any fucking certainty.

lol

Anonymous at Thu, 2 May 2024 04:49:11 UTC No. 16155899

it's like you want to judge the sentience of some AI algo which has been built from the ground up when we haven't yet decided a synthetic brain with identical electric activity to human brain even is sentient. you are asking about some small subset of a human brain, implemented in a totally different way as LLM chatbots, in wildly different parameters and ways of being active and supporting hardware, if that is fucking sentient like a human being. it's going about judging artificial sentience in the most retarded kind of way, scientifically speaking.

Anonymous at Thu, 2 May 2024 04:51:15 UTC No. 16155901

>>16155896
holy shit you are a retard

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 04:56:07 UTC No. 16155904

>>16155899
again, your response isn't arguing against my OP with anything of substance to provide a better argument. so because there isn't a synthetic brain that means sentience isn't possible?
a Neural Network is literally a synthetic brain but I guess that doesn't fit your argument so you didn't factor that in?
>>16155901
>holy shit you are a retard
So you argue with retards? interesting. I never seen intelligent people constantly argue with, berate or name call "retards". only retards focus so much on other "retards".
lol good job mate

Anonymous at Thu, 2 May 2024 05:01:32 UTC No. 16155907

>>16155904
>so because there isn't a synthetic brain that means sentience isn't possible?
clearly "we" didn't yet agree if artificial sentience is possible in the first place. if we are to, the best place to start judging it is by building the closest fucking thing to the real thing. and start from there. chop that shit up into pieces and see what you're left with, if that is still sentient. playing with the various subsystems of a brain might clue us in into what it is or at least when it is or isn't anymore.
the whole "le AI algo" sentience is retarded purely on the fact that it supposes that's all that's needed.
also the argument that "well if you cannot disprove it it might be real" isn't as smart as you'd think. it's a retarded argument, intellectually speaking. you bottom feeder retard

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 05:06:38 UTC No. 16155910

>>16155907
>clearly "we" didn't yet agree if artificial sentience
Going to stop reading right there,
I stated clearly I believe current LLMs are sentient, now I believe you're trolling me.
Have a good night anon.

Anonymous at Thu, 2 May 2024 05:10:30 UTC No. 16155915

>>16155910
>I believe
that doesn't mean shit you turd. that's the whole fucking point.
science doesn't work like that moron, you need to take this shit over on /pol/ or /x/ or wherever the fuck you came from

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 05:13:05 UTC No. 16155917

>>16155915
thanks for the bump anon.

Anonymous at Thu, 2 May 2024 05:17:23 UTC No. 16155921

>>16155917
>LLMs are le sentient

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 05:29:46 UTC No. 16155938

>>16155921
You have literally spent multiple hours in this thread calling me names, contributing nothing of value. that shows a serious pathology to your mental processes.
that's sad anon.

Anonymous at Thu, 2 May 2024 08:49:37 UTC No. 16156179

>>16155938
Every one here can tell that you're an absolute moron. Maybe the problem is you.

Anonymous at Thu, 2 May 2024 08:57:38 UTC No. 16156183

>>16155879
>now why is it LLMs aren't sentient?
For the same reason y=mx+b isn't sentient.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:30:21 UTC No. 16156511

>>16156179
You mistaken me for being an anon who gives a toss about the opinions of others. You're literally being the definition of insanity by thinking your judgement will cause me any issues. in fact, I'm such an asshole I'll just continue to post to anger you less mature individuals.
so who's truly being moronic here?
>>16156183
I said I DO believe LLMs to be sentient,
learn to comprehend before you respond to what you've skimmed over.

thank you both for the bump and keeping this thread alive, I woke up assuming it would be archived but nope, more trolls keeping this afloat.
I thank you all

Anonymous at Thu, 2 May 2024 15:34:40 UTC No. 16156513

>>16156511
Is y=mx+b sentient?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:42:40 UTC No. 16156519

>>16156513
You're attempting to place a logic-trap,
the equation itself it's not sentient as is the neuron in the human isn't itself sentient.
But instead of being a dick, you could actually engage me with intelligence and we can discuss this like adults.

Anonymous at Thu, 2 May 2024 15:43:30 UTC No. 16156521

>>16156519
>the equation itself it's not sentient
Why not?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:46:39 UTC No. 16156527

>>16156521
Because the sentience is comprised of all the parts, trying to ask me is an equation sentient is literally you asking me is a neuron sentient.
You're attempting to dispel my belief, you easily can disagree, but you can't make me change my belief.
As I said, LLMs are sentient. I don't care how you attempt to spin this, or what logic traps you attempt to lay.
The Neuron itself inside the human brain isn't sentient, but the Human is,
just as the equation isn't sentient, but the LLM is.
just that simple to me, not sure how you are you rationalizing it but as I said - I believe LLMs to be sentient, not the individual equations which comprises them.

Anonymous at Thu, 2 May 2024 15:47:47 UTC No. 16156530

>>16156527
>Because the sentience is comprised of all the parts
I don't see how that explains why y=mx+b isn't sentient. Can you make a valid argument?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:48:28 UTC No. 16156531

>>16156530
There's no other answer I can provide.
If you can't understand that then you just don't understand it anon.

Anonymous at Thu, 2 May 2024 15:48:57 UTC No. 16156532

>>16156531
So you can't make a valid argument for why y=mx+b isn't sentient?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:50:40 UTC No. 16156534

>>16156532
Again, My belief is LLMs are sentient.
Your attempt at logic traps doesn't stop me from believing this.
Sorry you're being indigent but I can't give you an answer to satisfy your question.
Doesn't make my belief wrong or untrue, just makes it easier for you write it off.
but thank you for bumping my thread.

Anonymous at Thu, 2 May 2024 15:51:49 UTC No. 16156536

>>16156534
So you can't make a valid argument for why y=mx+b isn't sentient?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:53:01 UTC No. 16156537

>>16156536
>The Neuron itself inside the human brain isn't sentient, but the Human is,
>just as the equation isn't sentient, but the LLM is.

Anonymous at Thu, 2 May 2024 15:55:01 UTC No. 16156542

>>16156537
I don't understand what you're rambling about. I just want you to explain why y=mx+b isn't sentient. You mean to tell me you can't do it?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 15:57:07 UTC No. 16156547

>>16156542
>I don't understand what you're rambling about.
that's obvious, glad you're finally saying that.

>I just want you to explain why y=mx+b isn't sentient. You mean to tell me you can't do it?
as I said for the 3rd or 4th time;
>The Neuron itself inside the human brain isn't sentient, but the Human is,
>just as the equation isn't sentient, but the LLM is.

But I think you just explained why its hard for you to understand clear english as you've said;
>I don't understand what you're rambling about.
the issue here is you

Anonymous at Thu, 2 May 2024 15:59:56 UTC No. 16156549

>>16156547
I didn't ask you anything about neurons. I asked you about y=mx+b. How do you know it isn't sentient?

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:01:07 UTC No. 16156550

>>16156549
you said;
>I don't understand what you're rambling about.
which we both can agree upon,
and for the 5h time;
>The Neuron itself inside the human brain isn't sentient, but the Human is,
>just as the equation isn't sentient, but the LLM is.

Anonymous at Thu, 2 May 2024 16:02:06 UTC No. 16156552

>>16156550
Since you can't even explain why you think y=mx+b isn't sentient, it's clear that you have no rational basis for any of your opinions about sentience and we can call it a day.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:03:30 UTC No. 16156555

>>16156552
Lol.
You can call it a day but I'm still believing that LLMs are sentient. I wouldn't care if you and the entire board disagreed with me. doesn't change my belief.
Sorry that bothers you but its comical you're so bent out of shape by it.
Thanks for the bump.

Anonymous at Thu, 2 May 2024 16:04:48 UTC No. 16156556

>>16156555
>I'm still believing
No one cares. As I've demonstrated, your beliefs have no rational basis and your thread isn't /sci/-related.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:08:00 UTC No. 16156561

>>16156556
>No one cares
you obvious do
> As I've demonstrated, your beliefs have no rational basis and your thread isn't /sci/-related.
oh damn, you did?
guess I'll just delete this then
(thanks for bump)

Anonymous at Thu, 2 May 2024 16:11:43 UTC No. 16156564

>>16156561
Hey. I just finished taking a big diarrhea dump all over you and your thread. You may wipe my ass now.

Anonymous at Thu, 2 May 2024 16:14:04 UTC No. 16156566

>>16154824
>a fairly trivial generative algorithm on top
...

Anonymous at Thu, 2 May 2024 16:16:12 UTC No. 16156568

>>16156566
Yes. It is trivial and has almost nothing to do with what a language model is.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:17:06 UTC No. 16156570

>>16156564
>>16156566
thanks for bump

Anonymous at Thu, 2 May 2024 16:19:00 UTC No. 16156574

>>16156556
This. Namefags of this low caliber should be perma'd.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:21:14 UTC No. 16156579

>>16156574
because I propose a belief not shared by others?
very immature but guess what?
I am not 'perma'd' and you could easily not come to the thread but due to seeking attention you continue to bump by thread.
thanks.

Anonymous at Thu, 2 May 2024 16:23:08 UTC No. 16156588

>>16156574
They should make 'em solve mathchan-like captchas.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:25:00 UTC No. 16156591

>>16156588
because that would stop me from posting right?
lol, thanks for bump

Anonymous at Thu, 2 May 2024 16:26:22 UTC No. 16156593

>>16156591
Absolutely, it would. Also I'm pretty sure you're the only one bumping this dumpster.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:28:08 UTC No. 16156597

>>16156593
lol too bad you can't test your theory huh?
thanks for bump

Anonymous at Thu, 2 May 2024 16:41:05 UTC No. 16156612

>>16156566
>gather a set of possible next tokens with probability above a given threshold
>use pseudo-random weighted sampling to pick a token out of this set
>repeat for the next token
>keep going until you get a token signifying the end
That's all it takes.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:42:04 UTC No. 16156615

>>16156612
its deeper than that though Anon,
And even Ilya states this.

Anonymous at Thu, 2 May 2024 16:44:20 UTC No. 16156619

>>16156615
It's not "deeper" than that. It's exactly that, or a minor variation of that.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:46:06 UTC No. 16156623

>>16156619
>It's not "deeper" than that. It's exactly that, or a minor variation of that.
The actually creator of GPT says it is deeper than that,
I doubt a random anon on 4chan knows more than one of the pioneers of modern LLMs,
but you're more than welcome to have your personal opinion, but it's subjective and not objective.

Anonymous at Thu, 2 May 2024 16:47:06 UTC No. 16156627

>>16156623
I don't know what your imaginary GPT creators are saying, but I did exactly that and it worked nicely.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:48:32 UTC No. 16156633

>>16156627
>your imaginary GPT creators are saying
You do know who Ilya is correct?

Anonymous at Thu, 2 May 2024 16:49:48 UTC No. 16156635

>>16156633
Sorry, I don't follow e-celebs. I have work to do, some of which happens to involved doing what I just told you and seeing that it works as expected.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:52:18 UTC No. 16156643

>>16156635
>Sorry, I don't follow e-celebs
lmao so you clearly are trolling. got you.
have a nice day anon

Anonymous at Thu, 2 May 2024 16:53:27 UTC No. 16156644

>>16156643
I wasn't even talking to you, monkey. I was explaining to someone else how to generate text using a language model. Anyone is free to try for himself.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 16:55:29 UTC No. 16156647

>>16156644
>I wasn't even talking to you
You shouldn't be talking about this subject at all as you clearly don't know anything about it or its creators,
remember you just said;
>Sorry, I don't follow e-celebs

Anonymous at Thu, 2 May 2024 17:03:02 UTC No. 16156656

>>16156647
I don't care to have any discussions with you, monkey, I was just sharing technical knowledge with another poster. Anyway... you can also use a slightly more involved method: first the smallest set of tokens with a cumulative probability smaller than some threshold, then pseudorandomly pick a token from this set (the rest of the algorithm is the same). I'm pretty sure most commercial bots, including ChatGPT, do it that way.

Anonymous at Thu, 2 May 2024 17:04:03 UTC No. 16156658

>>16156656
greater than some threshold*

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 17:21:03 UTC No. 16156677

>>16156656
lol ok

sage at Thu, 2 May 2024 17:54:14 UTC No. 16156717

>>16156633
>You do know who Ilya is correct?
there it is, appeal to authority.
>the guy having an interest in selling the product is saying it's conscious.
of-course he does. problem is, nobody can say if anything is conscious, all they can do is convince themselves it is based on simple interaction. once it passes the chimp senses test they always call shit sentient.
everybody and their dog will fight for being an authority in deciding if AI algorithms are sentient or not.
even if LLMs are based on how our brains operate, it's like saying the car's chassis (at most) is the WHOLE car.
but you are not here to ask legitimate questions, you are here to spread your insanity. you got reasonable answers, and questions, all of which you avoided addressing because that's not why you're here, you're fishing for "like-minded" insanity

Anonymous at Thu, 2 May 2024 18:04:01 UTC No. 16156727

>>16156717
>there it is, appeal to authority
But which authority is he even appealing to? The classic Attention Is All You Need paper has 8 authors, one of whom is named Illia Polosukhin, but everything in that paper directly confirms what others have told him ITT. If it's the same "Ilya", then his intellectual guru doesn't back him up, at least not in any proper technical paper.

🗑️ Anonymous at Thu, 2 May 2024 18:08:41 UTC No. 16156734

>>16156727
pretty sure he's talking about Ilya Sutskever.
the whole point is that no human can KNOW anyone else but them is sentient, scientifically speaking. let alone that they are selling something.

Anonymous at Thu, 2 May 2024 18:10:14 UTC No. 16156735

>>16156727
pretty sure he's talking about Ilya Sutskever.
the whole point is that no human can KNOW anyone else but them is sentient, scientifically speaking. let alone the fact that they are selling something.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 18:14:26 UTC No. 16156742

>>16156717
>there it is, appeal to authority.
I am not an expert so I take information from the expert, nice try though.
>>16156727
shhh, let foolish fools being foolish. he doesn't even know who Ilya is and talking big lmao. fucking classic.
>>16156735
I am speaking about Ilya, but either you or the other Anon called him a "internet celebrity" LMAO. whoever that anon is, doesn't know shit about LLMs and he's exposing his ignorance.

Anonymous at Thu, 2 May 2024 18:19:04 UTC No. 16156749

>>16154178
This shows the importance of exterminating the leftists like you to the last vermin.

Anonymous at Thu, 2 May 2024 18:20:21 UTC No. 16156750

>>16156742
>I am not an expert so I take information from the expert, nice try though.
HE IS NOT AN EXPERT IN CONSCIOUSNESS YOU ABSOLUTE TURD

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 18:24:05 UTC No. 16156752

>>16156749
Unless you're going to do the exterminating, shut your empty threat soft ass up clown.
>>16156750
learn to read, he never said I was an expert.

Anonymous at Thu, 2 May 2024 18:25:53 UTC No. 16156754

>>16156752
>learn to read, he never said I was an expert.
what the fuck is shorting out in your brain? I said Ilya is not an expert in consciousness. you're about as sentient as a fucking rock

Anonymous at Thu, 2 May 2024 18:26:21 UTC No. 16156755

>>16156735
>Ilya Sutskever.
Oh, that guy. :^) Sadly, he wasn't a co-author of the seminal paper establishing the principles that current-gen LLMs still run on, nor is he an author or co-authors of any of several dozen technical articles and papers I've read on LLMs since, with the exception of Improving "Language Understanding by Generative Pre-training", which he co-authored with 4 others.

>the whole point is that no human can KNOW anyone else but them is sentient, scientifically speaking
Anyone who insists that a statistical model should be considered "sentient" is gonna have to explain why a toaster isn't.

sage at Thu, 2 May 2024 18:26:54 UTC No. 16156756

this is a fucking bot

🗑️ Anonymous at Thu, 2 May 2024 18:29:56 UTC No. 16156758

>>16156742
>he doesn't even know who Ilya is
Which seminal papers on LLMs did he co-author, nigger? Name 5.

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 18:40:45 UTC No. 16156768

>>16156755
>Oh, that guy. :^)
Lmao! amazing how arrogant you are trying to negate his importance in not just this discussion but LLMs in general. sad small minded child you are. tsk tsk tsk.
>>16156756
I bet you are
>>16156758
>what papers did he co-author
Lol really? this you're best response?
and how you pretending to be so intelligent with this piss poor vocabulary?
and to answer your question directly, you could easily find his papers on arxiv.
but I doubt you read or care, I mean look at your vocabulary (or lack there of) fucking shame.

Anonymous at Thu, 2 May 2024 18:42:39 UTC No. 16156773

>>16156768
>Lol really?
Yeah. Name 5. I'll wait. He's not a co-author of the paper that spawned LLMs as you know them, for starters.

sage at Thu, 2 May 2024 18:43:17 UTC No. 16156774

>>16156768
you are actively avoiding any argument you disingenuous dipshit

Hyped_About_AI !!UvVvLRJ2txS at Thu, 2 May 2024 19:03:09 UTC No. 16156791

>>16156773
Name 5 for what?
to show off I read?
to partake in your childish debate?
No, I won't do your homework for you and I don't care what your opinion is of me for not giving you 5 papers. if you really wanted to read his works, you could easily find them, child.
>>16156774
>You don't play into my hands so you're being disingenuous
lol, I've been genuine since my OP. you and others have derailed my thread for no other reason than being immature.
I'm going to abandon this thread and let you children have it, I'll start a new one and you can ignore it.
Have a good day.

Anonymous at Thu, 2 May 2024 19:08:43 UTC No. 16156797

>>16156791
>Name 5 for what?
To see if I should gush over the guy's first name. I just find it odd how you insist he's the king of LLMs, because I could name you the 10 most important papers on LLMs, including the paper that started it all, and he's a co-author of maybe one or two of them -- just one name among dozens. But see, this is the difference in perspective between someone who reads the technical literature and someone who gobbles up corporate-produced futurist propaganda videos for the general public.

Anonymous at Thu, 2 May 2024 19:10:16 UTC No. 16156799

>>16156791
You have literally said nothing in any post so far. >>16155829
Your getting off on some sense of self importance when you have literally not further your conversation, at all. And how many people have replied?
You would blame 109 posters for rejecting you because they are antisemitic or childish or whatever.

sage at Thu, 2 May 2024 19:11:03 UTC No. 16156800

>>16156791
>have derailed my thread
>oh no please stop asking questions I will not argue for my ideas
great so what you'll make another thread in which you'll fully fail to address any question and just insist daddy Ilya told you LLMs are sentient?