Thread 107147791

AGI can't exist because Alan Turing and Godel proved that a machine that runs binary code can't encode every concept that exist IRL.

Poop and niggers dying..

>>107147791 (OP)
Exactly. You can't reach AGI with mathematics, but retards are gonna retard.

>>107147791 (OP)
you don't need to encode every concept. it's enough to just cover concepts that will make money and take your job.

>>107147791 (OP)
>can't encode every concept that exist IRL
Neither can you

>>107148103
Are you crying lil broeycuh? This nigga crying. Ayo Alty go butt fuck your sister again asf nigga rn

>>107147791 (OP)
>can't encode every concept that exist IRL
what cant it encode?

>>107148190
contradictory information.

>>107148211
what do you mean? you can store two contradictory datapoints in the same database

>>107148054
Sodomites tend to be dumb as fuck, and that's why they give each other aids like it's some business card and die before 35, but this fellow was rather brilliant. Such a shame that fudgepacking forever marred his otherwise impeccable reputation in science.

Hope he got saved at some point.

>>107148248
how would you code a statement that is both false and true at the same time like division by 0.

>>107148190
e.g. a program that takes in code and outputs whether it terminates with 100% accuracy

>>107148257
You can emulate a quantum processor using binary, and that in theory would solve this very issue.

>>107148257
you can't have actual, literal, physical manifestations of paradoxes and contradictions in real life either because this existance itself doesnt allow it so what are you talking about?
>>107148263
is there ever a true "100%" accuracy in real life? no. and that doesnt matter at all pragmatically anyway

>>107148275
>is there ever a true "100%" accuracy in real life
the speed of light in a vacuum

>>107148272
>>107148275
Some stuff like the that paradox that particles are waves at the same time, would be quite dificult to implement.

Or P != NP.

Some math problems can't be done by chat gpt because nobody has a clue how to solve them.

>>107148299
>the speed of light in a vacuum
and how do you know that the speed of light in a vacuum exists at all and isnt just a figment of your hallucination because you are just a brain in a vat somewhere? is there 0% chance that this is the case? no? oh.
>>107148301
limitations of human knowledge isnt the same as something being literally impossible to do

>>107148301
Ultimately that's an issue of brain power rather than a hard limitation of binary.

>>107148315
>>107148317
some math shit is impossible to do and has been proven to not have a solution.

We can't emulate by example all the molecules of a gas.

>>107148325
>some math shit is impossible to do and has been proven to not have a solution.
if its possible for something to exist within this reality, its possible to encode it in a machine, which is part of this reality.

you talking about paradoxes, literal contradictions etc, things that are literally impossible to actually manifest in this reality in any way shape or form is a separate and irrelevant category of information pragmatically.

>>107148348
>paradoxes
kinda.

There's a bunch of paradoxes that you can make a PC crash because they enter into an infinite loop and causes a halting problem.

>>107148348
Encode my dick up your ass nigga

>>107148413
cant encode nonexistant matter

>>107148190
I don't know what specifically OP is referring to, but I suspect he's talking about the "halting problem". To put it simply, you cannot write a program A that can always determine whether a provided program B will halt. There are also things such as incomputable sequences (see Busy Beaver numbers) and "Godel's incompleteness theorem" which proposes that there will be true but unprovable conjectures for any system of mathematics.

Having said all that, I don't think any of this has much to do with AGI and >>107148091 is more or less correct. A more relevant topic would be "Nothing is real and we all live in a simulation", since that requires contending with computability.

>>107147791 (OP)
>>107147890
So you can't emulate a brain on a computer by simulating the physical behavior of every molecule? Or do you think that you could but the brain would give out different outputs? Explain.

>>107148437
>halting problem

>>107148348
>you talking about paradoxes, literal contradictions etc, things that are literally impossible to actually manifest in this reality in any way shape or form is a separate and irrelevant category of information pragmatically.
>>107148315
>and how do you know that the speed of light in a vacuum exists at all and isnt just a figment of your hallucination because you are just a brain in a vat somewhere? is there 0% chance that this is the case? no? oh.

>>107148437
>>107148453
I'll explain it this way.

Let's say you have the CS math to build AGI.
Pretty sure you will run out of resources like money, energy, silicon chips, for an AGI to basically be so complex it can sort of predict and simulate all humans or at least a couple billion humans.

We don't have the current tech to do that simulation.

That's what I mean by the halting problem.
Our current tech will run out of resources before reaching the point it can control all humans.

Also, such AGI will require basically such masssive datacenter that it will become a very easy military target to put a misile if the AI decided to be funny.

We have now run out of data, and AI slop can't be used as data, that's a core limitation to the AGI dream.

>>107147791 (OP)
Oh, you know words?
Name every word

>>107148257
boolean = false; is true that it's false

>>107147791 (OP)
You are a retarded nigger monkey that does not know what he's talking about. Your whole knowledge of the topic comes from popsci articles and jewtube videos which essentially peddle misinformation to midwits who then shit out retarded nonsense like this.
Neither Turing's Halting Theorem nor Gödel's incompleteness theorem have anything to do with encoding information. Rather, Turing showed that certain problems cannot be solved by computation and Gödel showed the existence of true but unprovable statements. Ironically, Gödel's proof relies on being able to encode such statements, perhaps not with binary code but rather with prime numbers. Either way, binary code can encode basically anything.

>>107148257
"division by 0" is not a statement. If there is a statement that as you say is both true and false, you can encode it however you want; technically, even the English language coded in ASCII is a valid encoding so the fact that you can even write the contradiction down on an anonymous Mongolian basket weaving forum is proof that you are retarded.

>>107148263
That's not encoding a concept, that's just trying to create something that doesn't exist. It might surprise you to find that even though the word "unicorn" exists in the English language, they do not exist in real life. So you can encode the concept of such a program (see above) but not such a program itself, since no such program can exist. Note that this is not a restriction of binary per se but rather of the Turing machine (which can use an arbitrary but finite set of symbols) and seemingly also of our universe.

>>107147791 (OP)
>proved

>>107148453
the brain can't compute everything either.

concepts are derived empirically from experience anyway. Just a human invention. If you think, concepts are something divine that we are somehow able to "receive", you repeat the same mistake as second rate German idealists that tried to understand Kant.
a machine that shuffles around words without experience is about as close to god as an npc freshman trying to write his first essay.

>>107147791 (OP)
AI doesn't exist. It is simulated intelligence, not actual intelligence. None of these algorithms are truly intelligent since they can only remix, never create new.

theoretically you could simulate neurons with quantum computers, but not with current tech so it's a jewish scam and a bubble

>>107151611
>since they can only remix, never create new
And what makes you think people are any different?

>>107151611
>heavier than air flight will never exist

>>107147791 (OP)
He was wrong about the halting problem. https://github.com/OutfoxSemiLLC/HaltingProblemSolutionIdea

>>107147791 (OP)

The GitHub Description:

>>I solved the halting problem, it doesn't exist.
>>Just check every n + 1 clock cycles whether the previous savestate (a term from speedrunning) matches the current savestate exactly, byte for byte. You can discard the previous previous savestate.

>>n can be a arbitrary precision integer.

>>You add 1 to n every time you check.

>>Also, my algorithm detects all infinite loops, but some
infinite loops may be whitelisted like the OS kernel.

>>Now where's my million dollars for proving P=NP...

Waiting for my million dollars for proving P=NP…

vast majority of humans dont really embody high concepts either.

>>107147791 (OP)
ESL jeet brainlet thread

>>107148453
The TL;DR version would be is that you cannot compute conscience because it requires a higher mode of thinking or as they like to call it "out of the box". Mathematics is a closed system that cannot think outside the borders. See https://youtu.be/e9484gNpFF8?si=AVpfjw9POuhrfvKj and more specifically Kurt Goedel and his Incompleteness Theorem for more explanation. Ignore the interviewer, he's a retard faggot.

I think it's possible to create intelligence but not with mathematics alone, you would probably need some sort of synthetical biological element as well.

>>107148257
Using ascii/utf-8. Just like your post is.
Are you retarded or something?

>>107153867
>Strings are the same as the mathematical constructs that allow for branch controls of the flow of the code
>This is the level of "computer programming" expertise of the hackers knows as /g/

>>107153855
This is false and they never proved that. What more, Turing conjured that everything that can be calculated by a mathematician with pen and paper can be calculated by a Turing machine, making them equivalent.

>>107153891
What do you think Godel encoding is you absolute larping retard?

>>107153916
how do you code a branching instruction that allows at the same time two diferent outputs because the input statement is a logical contradiction?

>inb4 a return with a vector2 or vector3

AGI is not necessary. Once you have an autonomous robot that kills people from an opposing nation for you, you don't care if it can understand poetry or not.

>>107153947
AI tards literally think that an LLM trained on reddit and coded by indians is a civilizational threat.

>>107153916
>This is false and they never proved that.
Proved what? Incompleteness Theorem? That was proved in 1930.

Also, Turing never proved that. He only hypothesized the statement.

>>107153937
You use LaTeX.

>>107154275
>Proved what?
>>107153855 >that you cannot compute conscience because it requires a higher mode of thinking

>Also, Turing never proved that. He only hypothesized the statement.
What do you think a conjecture means?

>>107152189
>holy mother of all strawmen
I'm interested if anyone really ever said that, has that person heard of birds, and is he aware they are not lighter than air?

>>107154559
>What do you think a conjecture means?
I know it doesn't mean proof.

>that you cannot compute conscience because it requires a higher mode of thinking
Is there a math model of conscience?

>>107154982
>Is there a math model of conscience?
No. Which means Turing and Godel could never prove anything related to it.

>>107155003
Goedel's theorem is valid for any mathematical model.

Can a math model of something as complex as human conscience exist within those limitations?

>>107155077
>Can a math model of something as complex as human conscience exist within those limitations?
It's up to physicists to figure out, not mathematicians or computer scientists.

>>107147791 (OP)
>a machine that runs binary code
None of their proofs assume that
>can't encode every concept that exist IRL.
And most certainly not that. Both of these guys used formalism and formal but logically inconsistent infinities to prove nosense statements like "all concepts" which is distinct from "every concept that exists" let alone "every concept that exists IRL"

>>107148453
For starters, we don't know fully how brains work. If we did, what you're describing is basically building an artificial 1 to 1 brain. Which according to what Turing and Godel proved, would need to involve something more than running binary code.

So your question is a non sequitur.

>>107153916
>everything that can be calculated by a mathematician with pen and paper can be calculated by a Turing machine, making them equivalent
What do you zoomers get put in drinking water? Turing completeness has nothing to do with the problem at hand.

>>107156040
>nosense statements
So you just declared their work "nonsense" and used that insult as an argument to support your baseless position.
What's that called again...? Oh! A fallacy!

>>107156147
I don't need to provide you a sane definition of all concepts YOU need to show that it isn't paradoxical on it's face
Is it a set? certainly not since that would be paradoxical. So what is it?

>>107147791 (OP)
AGI can't exist because electrons moving through a wire fundamentally don't produce the same outputs as an unified biomechanism, it will forever be relegated to running rudimentary calculations and repeating back the lines you feed to it. That being said I really don't give a shit what Turing or Gödel have to say, one was a kid diddling faggot and the other a foreskin snipping jew.

>>107156398
Define biomechanism