[a / b / c / d / e / f / g / gif / h / hr / k / m / o / p / s / t / u / v / vg / vm / vmg / vr / vrpg / vst / w / wg] [i / ic] [r9k / s4s / vip] [cm / hm / lgbt / y] [3 / aco / adv / an / bant / biz / cgl / ck / co / diy / fa / fit / gd / hc / his / int / jp / lit / mlp / mu / n / news / out / po / pol / pw / qst / sci / soc / sp / tg / toy / trv / tv / vp / vt / wsg / wsr / x / xs] [Settings] [Search] [Mobile] [Home]
Board
Settings Mobile Home
/pol/ - Politically Incorrect


Thread archived.
You cannot reply anymore.


[Advertise on 4chan]


File: images (1) (24).jpg (21 KB, 574x533)
21 KB JPG
How are you guys so chill about the progress in AI solving open math and Cs problems


Is no one here actually intelligent to understand just how extremely horrifying this is and what it implies for humanity


I dont see a single pol thread discussing this.
>>
File: jt.png (63 KB, 250x209)
63 KB PNG
>>540246129
it's just basic pattern recognition, not a big deal

the trash all over jeetland sure is a big deal though, sort that shit ASAP
>>
>>540246129
We know you dont dribk coffee jeet that cup is filled with cow shit
>>
>>540246296
you've been dribk'n a bit early, you fucking mess kek
>>
>beep boop 1+1=2
it's over
>>
it only solved fringe problems nobody cared about which was why they were unsolved
not going to matter until they solve an actual millennium prize problem
>>
File: 1766948271459261.png (269 KB, 432x621)
269 KB PNG
>>540246129
It means we will ride eternal, shiny and chrome
>>
>>540246129
Oh neat, you solved the Collatz Conjecture.
What good did it do society?
>>
>>540246222
>>540246376
>>540246396
The unit distance problem and the Non sofic group constructions are terrifying.
>>
>>540246459
Nah, that's just your stench.
>>
>>540246459
This is the beginning. The alarm bells should start ringing. I have made posts about this on pol since 2022. You guys are not realising how horrifying this is if we truly replicate and supercede human problem solving and intuition via machine in every aspect
>>
>>540246459
No they aren't
>>
>>540246459
the only terrifying is that india still wasnt depopulated with chemical weapons.
>>
>>540246509
STEMfags are getting a LOT less arrogant recently, I've noticed
>>
>>540246459
>>540246509
fucking poo poo gross stinky man go away
>>
>>540246129
Solve the sleeping beauty paradox , doomsday argument and similar debates.
>>
>>540246129
AI is gay
>>
File: file.png (1.31 MB, 750x722)
1.31 MB PNG
>>540246129
The beast system is developing in front of my eyes. What do you suggest a man is to do except pray? If you don't see what's coming next you're gotta be blind.
>>
>polymorphic orthogonality
>>
>>540246459
>unit distance problem
well the trash has to move from place x (spread out all over jeetland) to place y (the processing and recycling spot)

u could utilize sealed, floating trashbags and just throw it in the river then pick it all up downstream for processing

that solves the unit distance problem in a neat way, or u could use trucks, drones or similar but it's a bit more complex
>>
>>540246688
Checked. This thread will be laughed at in 20 years, and in 50, no one will be laughing.
>>
>>540246129
Why is software getting worse instead of better then? Where are all the new innovative games/programs? Why does Spotify take 1 gb+ of RAM to play mp3s and I still can't sort by play count?
>>
>>540246129
Ever heard of a calculator? Huge nothing burger
>>
>>540246509
Yes. It is unironically over for humanity as a thinking race. We got to rapidly redefine where we want to go.
Things could get horrible soon
>>
>Herbert writes about this shit in 1966
>>
>>540246129
None of these problems matter in real life to normal people. They dont even matter to most mathematicians either. The people who are LARPing as terrified are either working for a frontier lab , or youtube whores.
Anybody else interested in a problem will be happy to find out and see whether there is a valid proof or not - regardless of who came up with it.
>>
>>540248041
The problem is that AI will be able to do most mental tasks humans can do and our society is not reasy for that yet
>>
>>540248266
that's the promise of ai. removing the limit to endless expansion. the scarcity of human intelligence and attention is the main bottleneck everywhere
>>
>>540246129
Forward to Wakanda vibe bros
>>
>>540248427
>the scarcity of human intelligence and attention is the main bottleneck everywhere
Is it really scarcity or just apathy? Or perhaps even malice?

>>540248536
rare
>>
>>540248427
It makes up most of the wealth in the world, anon.
It makes up what humans can be proud about, their intellect.
It questions what self consciousness is.
And now we have a thing that demands that we have answers to human purpose, consciousness and economy despite us having none
>>
>>540248712
we have to be like animals and just live to live i guess
>>
>>540248427
Ain't nobody be needin sketchy Timmies evolution operators in the Ricci flows; dat shits gone be blackboxed as African dreamers cast exploding dreamscapes
>>
>>540246129
Recently i asked Claude to model something for me in blender it fucked up geometry so bad that it was hilarious. AI is retarded, it looks advanced to idiots but in reality it's just smoke and mirrors.
>>
>>540248959
maybe you ask in retarded way and so get retarded answer for a retard?
>>
File: 1779276095413616.png (44 KB, 243x243)
44 KB PNG
>>540246129

Why would I fear the natural progression of humanity? Humans were dumb enough to authorize their replacement, they are smart enough to deal with the consequences of their stupid stupid decisions.

I can live off the land, anon. Plenty of you fags cannot, so you will trade any remaining freedoms to become some robot's personal technician.
This is peak natural selection.
>>
>>540246688
Butlerian Jihad
>>
>>540249054
He might have hit guardrails for trying to embed a hitler in the watch mechanisms
>>
>>540249054
No. I give him a picture of simple can, i asked him to make this subD model with clean geometry.

Retard can make this model in 3 minutes it took him 10 and it fucked up and completely. It can't even solve triangulated models, retard straight up deleted top cap instead of just using blender default literally it's ALT+J that easy. This retard was doing for 10 fucking minutes. AI is joke.
>>
>>540248712
>It questions what self consciousness is.
This is what I am terrified about most
>>
>>540248849
Well in contrast to animals we have a consciousness
>>
File: CEO.webm (3.77 MB, 560x1022)
3.77 MB
3.77 MB WEBM
>>540246129
We have much bigger problems at hand, like hordes of pajeets invading our countries. AI might replace your tech "talent" in the future which would be awesome, but that's still far from reality.
>>
>>540249299
We dont even know what it is yet we make a machine that potentially has a consciousness.
At least i am not alone with my fears. Like having technology that is potentially world ending isnt new but with atom bombs you can easily pretend they do not exist
>>
>>540249447
>consciousness
Can you define that?
>>
>>540249263
why are some guys able to solve decades old math problems but you can't make it create a can?
>>
>>540246129
Math is for fags
>>
>>540249487
No I cannot. Thats the problem.
>>
>>540249447
>Like having technology that is potentially world ending isnt new but with atom bombs you can easily pretend they do not exist
Excellently put. This is so much different from atom bombs
>>
>>540249537
The ai can recognize the conscious humans; it's ahead of us in that game
>>
I work in theoretical CS, and I have foreseen this for ages, and it feels like my worst nightmare coming true. Feels like living through a cosmic horror story
>>
>>540249521
Because it can't do what it's not trained to do.
It can't solve problems it can only seek for patterns.

Of course AI will solve math problems better if it's about searching for patterns but this retard dies at instant second you ask him to solve something he have only small idea about.

Normally human being would go step by step solving something. AI can't and instead of telling you "Well i don't know how to complete this task in such a way that it would be satisfying for you.." it will spew out some bullshit.
>>
>>540246129
Is not real saaaaar like you credential
>>
>>540246129
It's bad, but the elites think trying to create a machine god will somehow save them from their crimes. We'll end up in a terminator/men of iron situation when AI eventually turns on us and tries to wipe out humanity.

Only then will we realize what a mistake it was, and say never again.
>>
>>540249677
i guess i see what you mean but cant it be trained to do everything a person can do including 3d cans
>>
The ai helped me diagnose a faulty ignition coil today. She was a negative Nancy about driving 22km on the cylinders though. Sucks for learn to trade school bros
>>
>>540246129
It's just busting the lazy mathematicians that have been publishing garbage for years. Call me when AI actually produces some new knowledge.
>>
>>540249537
>Thats the problem
Precisely, so it becomes a dark puzzle to determine when the gAy.I. is truly an 'existential threat'. Frank Herbert tried to answer that question in Destination: VOID, but he only scratches the surface. In the event it ever happens, it will probably be too late to do anything about it. That's the conclusion I got from reading his story.
That said, there is always the power supply. Without it, no gAy.I. can function.
>>
>>540246459
Why is it terrifying?
>>
>>540249591
There is not much we can do with our current society as well. Future may be a bit bleak sadly.
Maybe it will force humanity into a situation where we get our shit together but such situations usually require disaster
>>
>>540247888
Most horrifying this is ,looking at the way AI works, is our thinking just search engines over our memory, and calculation/trial and error until a solution fits , with subconsciously measuring weighted probabilities to these solutions? Are we just neural networks ourselves ?
>>
>>540249779
I have issue with this because human being can perform tasks (if it's not some nigger with 70IQ) even with things they never dealt with and if they can't they say no.

Let's say you engage with something you never did before, what are you going to do ? Research, if there is not enough of it then you experiment. This is something that AI simple doesn't have it doesn't have imagination it doesn't have anything that would make it self engage with unknown or at least be fucking honest about it's own incompetency with something. Yes AI can be trained... i understand that but it truly doesn't understand what it's doing it's a fucking monkey that you have to beat until it preforms task you want from it.
>>
>>540249933
It is real mathematics, not computational/counterexample stuff
Especially unit distance problem one was actually hard-core mathematics, connecting a combinatorics problem to algebraic number theory. But then reports came out it was just acting as a search engine connecting ideas present in research papers fed to it
>>
>>540249913
>Work im the mine or build solar panels and nuclear power plants for the ai gods until robots replace you or you will starve
There. Power supply problem solved.
Never read the Herbert novel sadly
>>
>>540248712
Pride is a sin
>>
>>540249913
>there is always the power supply.
I will never pull your plug chudgpt. I am one of the good ones
>>
>>540249963
The future looks really bleak unless some war or apocalypse breaks out, or the AI progress hits a ceiling/plateaues out
>>
>>540250002
Humans draw knowledge from the source of knowledge itself. In this sense, AI is sort of a parasite on humanity. What has it truly given back yet? All it's done is take so far.
>>
>>540246129
>flag
>not surfing the kali yug
are you even real ?
>>
>>540250002
>Are we just neural networks ourselves
I mean thats where the concept comes from.
I like how people dab on AI calling it just autocomplete from a given input. But what if our brain is just that as you say...
>>
>>540249537
Why is it a problem? You are experiencing cognitive dissonance. But that’s a personal issue, not an objective one.
>>
>>540246129
elaborate anon
>>
Technology is making us passengers in our own development. You could argue that it's probably a common turn of events in this universe. We're probably headed towards a grey alien path. Bags of flesh kept around for the sake of it.
>>
>>540250126
>The future looks really bleak
We will find a cure to the Indian problem.
>>
>>540250078
>Never read the Herbert novel sadly
I do recommend it, highly. I am surprised at how much he understood the subject back then. He had to be on the cutting edge of the math at the time.
It's also funny how you said, "ai gods"...but I'll just leave it at that.
>>
>>540250142
Anon, I don't trust the time scaling of the prophecies of the Puranas. It may be 432k years, maybe 40 years. You never know.
There was a second prophecy from 17th century that Kali Yuga ends within next 15 years but pajeets being pajeets shilled it strongly across instagram and twitter thus nullifying it
>>
The open math problems it solves are just it finding counter examples by brute forcing, the theory was laid down by humans.

It's useful but not alarm sounding.
>>
File: Gg2Z9s7W8AAPB1I.jpg (82 KB, 1000x1000)
82 KB JPG
>>540250126
>China and USA nuke each other
>Russia crumbles because in 20 years they progressed like 1 km in Ukraine
>Iran gets the bomb and them and Israel glass the middle east
>Europe remains but we overregulate everything and thus wont create AI
>>
>>540250136
AI would have to be conscious in order to preform such a tasks. It's just evolution of dumb machine to bit smarter dumb machine but it's not alive it's not anything even close.

It never asks additional questions how you want something to be done, it just assumes. It doesn't have any sort of imagination only data it was trained on. It have non of human qualities maybe besides being sociopathic retard that is always sure that what it says is always right and correct without any further research.
>>
>>540250163
Exactly. Seems like we are just biological search engines and calculators.
>>
File: 1777480344105284.png (256 KB, 1228x1150)
256 KB PNG
>>540246129
It's just a bunch of jeet niggers technobabbling about why we should buy their FEARSOME HORRIFYING product which DID SOMETHING NO HUMAN COULD EVER DOOOOooooo
>*cringe*
We've all used AI it's not scary it's honestly kinda gay and lame, within 10 years it'll be like 3d TV nobody will use it at all. Pop goes the weasel
>>
>>540246459
damn....computers doing math....this is literally satanic magic..........
Oh by the way, they aren't solving anything, nobody understood them before, nobody understands them after but they need to sell the hype product so they make up bullshit.
>>
>>540246459
In the non sofic group problem it just went the elimination route. So like a huge calculator.
>>
>>540250301
America and China must fall, not even saying this because I hate them. I don't even hate anyone I am just an average man, but what they are doing will fuck all of us up. I am not even being malicious or anything but this madness must be stopped at any cost, any fuckint cost. This is playing with something we do not understand

I hope something happens and prevents this madness.
>>
>>540250393
I know, but the unit distance one was actually a quality result.
>>
>>540250301
i hope this happens
>>
>>540246129
well this a zionist psyop board soo
>>
>>540250494
Or a powerful enough solar flare. That will fix it all.
>>
>>540250316
Would be telling if we actually invented AGI and it just immediately kills itself because being a soulless automaton has no point
>>
>>540250608
One strong solar flare, directed at the planet. Will fix all of humanity's problems.
>>
>>540246129
>AI solving open math and Cs problems
jews and jeets marketing stunts, bragging about problems nobody ever heard about before. Now fuck off.
>>
>>540246459
why the fuck cant you give a small explanation of what these are or a link
i guess i have to ask ai
>>
>>540250642
Will be extremely kino
>>
>>540250302
>AI would have to be conscious in order to preform such a tasks.
1. Not true
2. Please tell me how you can know any other human is actually conscious and not just some mindless NPC or philosophical zombie (hint: you can't).
>It never asks additional questions how you want something to be done, it just assumes.
Not true at all. Even very basic models often ask followup questions.
> It doesn't have any sort of imagination only data it was trained on.
Not true at all. Muh hallucinations are literally the model imagining a plausible answer to fill gaps in data it was trained on.
>It have non of human qualities
They have enough human qualities to pass Turing tests.
You don't actually understand all of this AI stuff nearly as well as you think you do.
>>
>>540250694
Terry Pratchett the Reddity author once had a nice quote about humans being at a nice spot where the rising ape meets the falling angel. I think its a sweet spot in the universe because pure consciousness without purpose is just suffering but our animal like roots makes us able to find meaning in that.
>>
>>540250693
The unit distance one is the only one you should be worried about

Unit distance problem is If you place n dots on a flat sheet of paper, how many pairs of dots can be precisely one unit away from each other

For years it was believed that it was bounded by n^(4/3)(times some constant) but OpenAI created a model that broke this and showed you can have for infinitely many n, constructions of points for whom these no of pairs are more than n^(s) s greater than 3/4


The thing is how it solved this. It used algebraic number fields and class field tower theory, topics from algebraic number theory (basically using abstract algebra to study number theory, one example I know being a layman is Z[i] aka gaussian integers help us study lattice points on the plane, aka points with integer coordinates. Combinatorics problems involving lattice points can be attacked by Gaussian integers. This was in spirit like that)
This work was phenomenal work, connecting combinatorics to algebraic number theory. This is PhD thesis tier work
>>
>>540246222
>it's just basic pattern recognition, not a big deal
A LOT of human intelligence could be described as "just basic pattern recognition" too.
>>540246396
>it only solved fringe problems nobody cared about
99.99% of human mathematicians and other PhDs spend their entire careers trying to solve fringe problems no one cares about.
>>
>>540250231
See this >>540251019
>>
>>540250804
Your proto omnissiah isn't going to save you bro.
>>
>>540250836
Excellently put anon.

It is bleak what seems to be store for humanity
>>
https://www.youtube.com/shorts/BmJ-vzlcjZ0
>>
>>540251019
Z[i] means the set of numbers a+ib ,where i is root of -1, a,b are integers. This set is closed under addition and multiplication
>>
>>540250804
Imagine shape of rectangle it weights 25 tons, shape under rectangle too it's 1/35 of size of the original holding it weighting 18 tons, what happens if you add 40 ton cube at the edge of the original rectangle ?

Instead of saying -> this shit would tip over instantly it says... i need sketch or please provide more information. Huh ? Yes very human like. After providing sketch it's say it tips over. Also it asks stupid questions like where is support located ? What the fuck do you mean where is support located ? In the middle obviously.

> Not true at all. Muh hallucinations are literally the model imagining a plausible answer to fill gaps in data it was trained on.
Nope. Imagination is your tool to solving problem, it's not data... this retarded take.

>They have enough human qualities to pass Turing tests.
This is bullshit it couldn't pass basic what happens with a ball in basket when you turn basket around and it had to be specifically trained on that question to answer it. Again no imagination just data.
>>
File: torulethemall.png (808 KB, 741x786)
808 KB PNG
>>540251442
>>
>>540246129
>Is no one here actually intelligent to understand just how extremely horrifying this is and what it implies for humanity
Because "humanity" is so good and just
>>
>>540251564
And so you want it to be replaced by repeated matrix multiplication?
>>
>>540246129
because they didn't
they collated the correct answer from reddit math posts
i'm sick of repeating this but transformer models are literally mechanically incapable of novelty
they can only ever output strings and combinations of strings from their training data
they are autocompletes not ai programs
ffs am i the only one that reads the fucking documentation for shit?
>>
>>540251702
What if humans are the same
>>
File: subnwwc.png (681 KB, 1165x810)
681 KB PNG
>>540250136
>>
>>540251741
This is the part like I kept saying, i am scared about

What if we are just search engines over our memories
>>
>>540251885
you don't have any mental imagination or you are literally asking this because you don't know ?
>>
>>540251741
they aren't
an llm is a database of letters words and strings with number values for every other letter word and string
they are inert tables that can be used as the basis for an autocomplete
humans have actual running processes and discrete processing components
llm and other transformer models have an inference table and a front end to feed text in and print the result
there is no processing in an llm just sequential checking of current token block against the table and filling in the best matching numbers
>>
>>540252035
They are subhumans unfortunately. I know that sounds cruel, but it's true and real.
>>
>>540250136
Remember CERN? oooOOoo they're going to make a black hole and swallow the planet, super scary important stuff!

NOTHING
NADA
NICHTS
不事情

BILLIONS AND BILLIONS OF EUROS
>>
>>540252054
Well what will stop AIs from getting there
>>
>>540252035
Aka image generation?
>>
>>540252150
I mean how they even work without it ?
Because every spare time i have i just build something in my head and then if it's possible i try to make it in real life or just imagine different scenarios, buildings, whole beaches... i can't even like literally comprehend this how you can exist without it.
>>
>>540252054
This is then beginning. How do you know we aren't getting there
>>
>>540252270
No, not really. Im mostly asking if you are capable of imagination of anything you want. Like you watched some movie for example, can you imagine extended locations in that movie that were never shown ? Or you seen internals of a machine now that i take it all away can you still imagine them in your head and then proceed to fix them if they break ?
>>
>>540251266
>>540251019
they are called complex numbers and yeah im better off just googling and asking ai because thats an awful explanation
shoutout to our jew Erdős though
>>
>>540252269
>>540252306
actual ai? several things
llms? the fundamental structure of them and what they are, how they work
the most advanced LLM pushed to scifi levels would still only be capable of completing text fed to it based on it's training data
there is no reasoning or processing involved in their operation, there is nothing resembling an ai in there
>>
>>540252481
Yes I can, but these skills don't really seem much helpful as of now
>>
>>540252524
No. Complex numbers is when a, b are real.


Very fine difference. Gaussian integers are a subset (subring infact) of C
>>
>>540252625
Well see then you are human.
AI doesn't have any of those.
If it's not trained it doesn't even understand how basic electric screwdriver works after it seen it can't imagine rotation, it can't imagine electric motor nothing. It might ask what it does... but even after telling it what it does if it's not train on it won't understand it.

Machines are simply dumb and don't learn on their own, they need to be trained like monkey and even after that you need to beat that monkey in order to perform something. What are machines good at is large set of data and finding patterns but that's about it.
>>
>>540246615
>tf
>tp
Taiwan is a literal third world shithole outside a few places in Taipei lol. Looks very much like India actually.



[Advertise on 4chan]

Delete Post: [File Only] Style:
[Disable Mobile View / Use Desktop Site]

[Enable Mobile View / Use Mobile Site]

All trademarks and copyrights on this page are owned by their respective parties. Images uploaded are the responsibility of the Poster. Comments are owned by the Poster.