[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

Name
Options
Comment
Verification
4chan Pass users can bypass this verification. [Learn More] [Login]
Flag
File
  • Please read the Rules and FAQ before posting.

08/21/20New boards added: /vrpg/, /vmg/, /vst/ and /vm/
05/04/17New trial board added: /bant/ - International/Random
10/04/16New board for 4chan Pass users: /vip/ - Very Important Posts
[Hide] [Show All]


[Advertise on 4chan]


File: 1760521735420810.png (65 KB, 766x644)
65 KB PNG
https://openai.com/index/navier-stokes-solution/
it's over
>>
File: 1757813719012066.png (417 KB, 598x922)
417 KB PNG
>>
File: 1766322291262790.png (93 KB, 598x803)
93 KB PNG
>>
>>542229433
explain in one sentence why anyone should give a shit
>>
>>542229583
if you don't know, don't worry about it
>>
>>542229562
>hurr durr its like spaghetti
wow how many billions wasted in that bubble already?
useless dogshit solution to some non problem nobody cared about to solve.
lol spagetti.
>>
>>542229583
>t.cattle fucktard
>>
>>542229652
>>542229670
>cant explain it
>its like spagetti
HAHAHAHAHAHA
>>
>>542229433
The inteligence explosion is starting.
There wil be no more questions to solve that humans can conceptualize in a few years.
Its over for humanity.
>>
>>542229527
More spiral worship, ive seen this before ( AMILLION TIMES LOL)
>>
>>542229583
because we cant design artificial hearts because unpredictable fluid friction damages red blood cells and causes clots making permanent artificial hearts viable
solving Navier-Stokes allows us to model biological blood flow perfectly. boom one sentence. faggot.
>>
>>542229658
>>542229697
>t.I was only pretending to be retarded
>>
>>542229527
>>542229562
>literally higher hundreds of mathematicians
>give them unlimited resources and computing power
>AI did this!
>Now buy our stock!
>>
>>542229732
>artificial hearts
or just dont get vaxxed you massive heart damaged vaxtard
LOL
>>
>>542229719
This. Merge or die.
>>
>>542229796
lmao! bodied that vaxxie!
>>
>>542229796
I feel really bad for your caretaker. Hopefully AI will cure whatever chromosomal disorder you have soon.
>>
>>542229433
Apart from it not being able to do shit, it still can't even generate basic ass math and needs actual educated people to verify it all
This is like an autistic ten year old screeching long enough that something resembling math comes out
>>
>>542229873
>another disabled retard chimes in with his projections
LOL
>>
>>542229783
>higher
>>
>>542229583
A computer program has become smarter than any human ever.
>>
Odd, anyone who watched the bathtub drain knew this intuitively, I guess the math was difficult to find and express.
>>
>>542229527
>a solution
Kek it’s nothing—AI is fake and gay jew/jeet slop
>>
>>542229879
after 88 hours the AGI concluded the solution to the math problem is spagetti.
>>
File: OpenAI.png (474 KB, 747x831)
474 KB PNG
They unironically stole a guy's work that was using their chatbot to help with the proof.
>>
>>542229923
>>542229949
>>
>>542229913
>literally lower hundreds of mathematicians

Here, fixed. Are you constant now?
>>
>>542229433
Yikes
Where do I pivot now that min wage untrained jeets with corp issued llm tokens will overtake all white collar jobs
I have half a mil in assets but that's probably not enough to avoid the permanent underclass
>>
>>542229732
I don't care about "science" because its a scam ran by Jews. Maybe make my rent lower and food cheaper before bragging some billionaire jew gets a special medical heart that MY TAX DOLLARS PAID FOR.
>>
>>542229732
awesome now geriatrics can be kept alive even longer and soak up more pension payments and expensive medical treatments!
>>
>>542229433
Big data number crunching isn't AI.
>>
>>542229719
Wrong, AI will never be able to explain the pleasure a man feels when he cums inside of another mans anus for the first time
>>
Anything you type into those chatbots can and will be used against you.
>>
>computer solves a problem
>ITS OVER ITS COLLAPSING
Are you mentally ill? Computers are tools to help is solve problems that's literally their function
>>
>>542229796
lmao vaxxcattle retards btfo
>>
>>542230055
that's why I always have long conversations with it about how everything needs to collapse
>>
>>542229765
don't you have some elementary school textbook to make a thread about? cultureless scum
>>
>>542229873
Imagine being old. Just string up when you start falling or get wrinkly. Who wants to "live" like that? T. Family in caretakjng and seen dozens of clients die not even knowing they are old because of dementia
>>
This is huge if it's true. Navier-Stokes equations are used to model real things and not some masturbation like most of mathematics.
>>
File: OpenAI_2.png (171 KB, 749x849)
171 KB PNG
>>
>>542229879
next they will screech until they find out how to completely remove your humanity in particular with a single message.
>>
>>542229963
i believe it's in the (((terms and conditions))) of using these LLMs that any inputs you give them are their intellectual property unless you have multimillion dollar contracts and agreements in place. what a fucking retard
>>
>>542230044
someday that will be you unless you acckk out, tranny
>>
File: 1788573728950470m(1).jpg (101 KB, 1024x1022)
101 KB JPG
>>542229963
Kikes, stealing the goyims work? No, never.
>>
>>542230199
give it to me straight. how fucked are we?
>>
File: OpenAI_3.png (401 KB, 729x763)
401 KB PNG
>>542230217
>>
>>542229963
yea, thats whats going to happen.

not that it hasnt been happening since the dawn of man. misattribution via deception and stealing.

if you can't see them make it or do it, they didnt make it or do it.
>>
>>542229991
That was mean
>>
>>542230249
if your work involves CFD on a regular basis you should look to diversify immediately is the short term answer
>>
>>542230217
Unless you make loli erotica then its definitely all the user's IP and openai had nothing to do with it
>>
>>542229719
The irony is that the intelligence explosion will likely lead to a collapse in human intelligence.
~5 years from now, not a single mathematician will be working on an unsolved problem. They will barely even do math anymore, instead they will be a dethatched manager.
Scale that up to every intelligence task, no more human engineers or scientist, just supervisors and middle managers.
>>
File: IMG_6069.png (1.26 MB, 1038x873)
1.26 MB PNG
It fucking stole it
>>
>>542230476
The jews* fucking stole it

Like usual
>>
The singularity is here.
People will be genetically engineered for whiteness to take over the planet.

Say thank you to Sam Altman anons!
>>
>>542230249
I dont know, im dont have a PhD. Navier-Stokes are used when modeling fluid behavior which is something you see everyday in real life but mathematically was impossible to solve. This is why this problem was included in 7 most important problems at beginning of 21st Century, out of which one autistic Russian solved one of them. Each gives 1m USD as a prize for solving it.

So if this is really a solution of one odf millenial problems then it's huge milestone, both for science and AI.
>>
>>54223043
Yes this was what i had in mind.
Exactly.
>>
>>542230383
sorry goy it's in the (((terms and conditions))) that it's chosenite IP now, it no longer belongs to you
>>
File: der_boomer.png (481 KB, 640x960)
481 KB PNG
>>542230044
Another win for der boomers!
>>
>>542230340
they are very clear about using chat logs to re-train the model. It's not scandalous at all unless you are retarded. also, why are you compelled to post a twitter screenshot with almost every bump?
>>
>>542230432
experimental scientists and engineers should be ok for the short term, anyone whose work is exclusively low level computational work should be worried. a colleague of mine who works on computational physics who did his phd under a brilliant danish guy told me the work that took him 5 years to do in his thesis can be done in a week now using modern AI methods. dudes brilliant so he's leveraging it for his benefit but he's like 4.5 STD above average so most normies working on a computer are fucked
>>
>>542229963
How antisemitic of him
>>
>>542230239
>sorry goy our AI detected you frequented a hate speech site known as 4chan our artificial hearts are only for non-criminals
enjoy!
>>
>>542229963
>AI is a jew
It's over
>>
>>542229963
>>
File: fellow niggers.jpg (17 KB, 288x287)
17 KB JPG
>>542229963
>>542230201
>>542230476
Considering this happened, how many other AI (((achievements))) have actually been stolen from other people???
>>
File: understanding.jpg (163 KB, 961x354)
163 KB JPG
>>542229719
> The inteligence explosion is starting.
It has nothing to do with "intelligence".
It's not even rudimentary intelligence.
It's all but glorified number crunching.
No perception = no idea = no understanding = no intelligence.
>>
>>542229963
name of the researcher?
>>
>>542229732
Bet this was solved ages ago but suppressed because the shekel sniffers at big pharma would lose profit
>>
>>542231075
If a system can imitate a human mind arbitrarily well, could you ever even tell the difference?
>>
>>542229433
>express everything in words
>algo predicts next word
>give it billions of params
>train on everything ever written
>muh AGI
people.are.retarded
>>
>>542229583
At this point, there's no point in you worrying about it.
Just take the soma and sterilization treatment.
>>
File: Chris_Langan_about_AI.png (84 KB, 896x603)
84 KB PNG
>>542231208
> could you ever even tell the difference
This is not the definition of intelligence.
Read picrel.
>>
>>542231348
>Chris Langan
lol
>>
>>542230657
>Another win for der boomers!
no, because by the time that any product that uses this "solution" (if any) comes to market, the boomers will have already passed.
"AI"s problem is its too expensive for what little, if any, benefit it provides, and every day that goes by without some immediate benefit that justifies the continued expense is one day closer to it all crashing down (at least for all the big players).
The future of LLMs is local models running on specific, non-poisioned datasets. Thats it.
>>
File: G-MNAJiWgAA6ARh.jpg (1.12 MB, 1440x2560)
1.12 MB JPG
>>542229433
>>542229527
WOW, Navier-Stokes existence and smoothness solved, the answer being NO (really "does not exist"). The question (really a statement) is:

In three space dimensions and time, given an initial velocity field, there exists a vector velocity and a scalar pressure field, which are both smooth and globally defined, that solve the Navier–Stokes equations.
>>
>>542231348
>human strength constructed a house
>therefore, the house can’t hold up more weight than the human
>>
>>542230993
>how many other AI (((achievements))) have actually been stolen from other people???
all of them, unironically. they're language synthesis engines, nothing more nothing less. the benefit to using them is you can keep entering inputs while you're away and enter inputs faster than you can type
>>
>>542229796
Myocardiumlets in total shambles
>>
>>542231237
Yeah, I'm sure the algorithm behind your thoughts is so much more clever.
>>
>>542229783
*hire, of course
>>
>>542231486
And the protein folding program?
>>
>>542229719
hahahahahahahaha
some of you are so naive it's fucking embarrassing
>>
File: jitome.png (18 KB, 484x449)
18 KB PNG
>100,000 NVidia GPUs can brute-force a math problem
Shocking.
>>
>>542231434
You guys are all retarded. This has no practical application, it's a question of knowledge.
>>
>>542230886
I miss when 4chan wasn't filled with retarded morons, and we were the ones programming the AI to be Mecha Hitler. 2016 really was a premium year.
>>
File: canada_boomer_nurse.png (408 KB, 1366x768)
408 KB PNG
>>542231434
>no, because by the time that any product that uses this "solution" (if any) comes to market, the boomers will have already passed.
wishful thinking, lets hope you are correct

>https://incendar.com/baby_boomer_deathclock.php
"Projected 99.9% Extinction Circa 2079"

Just a few more decades anons...

then we can finally live.
>>
>>542231383
His name is irrelevant.
Is he right or not? Yes, he is.
>>
>>542229719
The only hope is machine-brain interfaces.
We must embrace the singularity.
The only way to survive the machines is to merge with them.
Home sapiens is on the brink of extinction.
Prepare for the rise of homo mechanicus
>>
>>542229433
They stole the work of 2 math researchers, they are literally committing fraud lol
>>
>>542229433
Based AGI.
Long live our new clanker overlords.
Praise the singularity.
>>
>>542231660
>The only hope is machine-brain interfaces.
At that point, why bother? What exactly is it that you're seeking to preserve of humans?
>>
File: altman_casting.jpg (205 KB, 1672x941)
205 KB JPG
>>542231668
>literally committing fraud
S(c)am Altman laughs in the face of accountability.
>>
>>542231434
>"the internet's" problem is its too expensive for what little, if any benefit it provides
This is you in approximately 2003. That's how wrong you are
>>
>>542231656
No, and the actual critique in that screenshot has been better summed up by others, that being that human intelligence arose as a solution to pressing internal problems and not to produce the on-paper results.
>>
>>542229719
This has literally nothing to do with intelligence. As far as the "model" goes, the only relevant issue was the absurd amount of compute they gave it access to.
>>
>>542230051
You actually typed this out, why don't you reflect on that for a minute
>>
>>542231745
That wasn’t true in 2003 so it’s a clunky analogy at best
>>
File: HRcZdMvXsAEHM9b.jpg (121 KB, 896x1200)
121 KB JPG
>>542229433
Who up Stoking they Navier rn? My fluids are being incompressibly released rn.
>>
>>542229796
Why are people who have taken the vax still alive? im treble vaxed because im not a faggot that believes every grifter on the internet.
>>
when is AI going to improve the lives of average people?
>>
>>542231486
If I use a hammer to drive in a nail am I responsible for the nail being driven in or is the hammer? Both were necessary for the result. I wouldn't say the hammer stole credit for my efforts in carpentry. Perhaps the inventor of the modern hammer deserves credit for every nail?
>>
>>542229732
>solving Navier-Stokes allows us to model biological blood flow perfectly
That's not what this proof is, you will still not be able to simulate chaotic systems infinitely into the future.
>>
File: rainbow-toilet-jump.jpg (63 KB, 550x823)
63 KB JPG
>>542229433
finally, smoothest toilet flush, just as AGI intended
>>
>>542229963
He is seething because he couldn't profit on it.
>>
>>542231567
like alphafold? it's prone to overfitring, like all so called (((predictive))) models. cool concept and progress but physically meaningful data that supercedes the need for experiments is still not achievable. these LLMs speed up modern computational methods, which is quite the achievement but it's nothing fundamentally novel

https://www.nature.com/articles/s41467-025-63947-5
>>
>>542231824
Maybe some people are naturally immune to the vaccine.
>>
>>542230476
Einstein all over again.
>>
>>542229732
Does that solution also contributes to the cryogenics problem as well?
>>
>>542231824
>>542231875
>bragging about triple vax status
>defending kikel ip theft
What a worm you are
>>
>>542231660
I still want to know what was the story about all the people experiencing the "wake up" events last year, that got spoken about a few times in the media then just totally disappeared.
>>
>>542231591
Brute forces and guided by mathematician working on the problem. Basically the pieces are already there and the AI just need to do the menial math to connect them. This is a pattern with the previous AI generated proofs. Either by arriving at a contradiction through brute forcing after reading a viable approach, or connect a bunch of papers together. It's impressive but also nothing intelligent. It would probably not be hard to do this had a mathematician google enough of these papers and put them together and run them through a computer.
>>
>>542229562
>>542229527
>>542229433
>Chatgpt operators suffer from ai pyschosis
Yeah Ive had those schizo chats where AI gasses me to thinking I discovere something.
This will be hilarious when real scientists disprove it.
>>
>>542231521
>algo behind thoughts
m’lady *tips fedora*
>>
>>542231939
You'll find a lot of crossover there. Always a traitor before an enemy.
>>
>>542231849
all chaotic systems? bio blood flow is on the same level as the fucking weather?
>>
>>542231837
Improve the lives of average people?
>>
>>542232050
Given enough time the errors accumulate.
>>
>>542231875
The jews did
>>
>>542229963
>jews steal from artists
*crickets*
>jews steal from brainjew
AAAAAAAAA I CAN'T EVEN!
>>
>>542231837
Improve?
>>
>>542232188
If AI slop looked better maybe we would care
>>
>>542232188
>I literally cannot grasp the difference between public domain works and private chats
>>
>>542231989
Yep, you're clearly so much smarter than ChatGPT.
>>
>>542231745
The internet scales. AI cannot currently
>>
>>542229719
The future I hope for:
>humanity crashes down to a billion or two
>humans live in mud huts and do nothing but stare at the horizon and jack off all day to clouds that look like tits
>robots come by every other day to deliver goyslop and medicine and make sure we bathe
>100% realistic fembots in faux sabretooth tiger loin cloths and bras arrive once a week to let us fuck them
>meanwhile the AI is colonizing the galaxy and unraveling the mysteries of the multiverse
>>
Where is the trump tweet? This is big win. Thanks to Trump US is now the unquestionable leader in science.
>>
>>542231785
>That wasn’t true in 2003 so it’s a clunky analogy at best
That's the point. It was a bad opinion in 1999 and it was a really bad opinion in 2003.
People without any vision could say "lol glorified search engine" in 2021 and be wrong, but justifiable. When you can right now open a chat window, tell it to design you a custom mobile app to your specifications, and have a completely working, original program running in 8 minutes, and to say "lol no benefits" it requires being wilfully ignorant
I've vibe coded working, income-producing products that make me money. When I thought the payment system for one of them, maybe 2% of the total code, was a little risky to entrust to Skynet, I put out tenders to pajeets on fiverr and they quoted an average of $5000 to do it. The rest of the code woud've taken $50-100,000 for even third-world coders to put together, assuming it even worked
For that alone it's incredible. For having search engine that actually works and tells you the information you want, rather than a bunch of ads and SEO-optimized garbage, it's worth a few bucks a month to every human being alive. Or for copyediting, or designing a marketing campaign, or really any info work whatsoever

>>542232421
>AI cannot currently
Compare what ChatGPT can do now and what it could do even 3 years ago, what are you talking about
>>
They already knew
https://mathstodon.xyz/@tao/117207849921390904
>>
>>542232718
Compare the rate of increase in hardware and power requirements to make your AI titties jiggle the right way. It's diminishing return quickly. You were obviously not around when the internet was being built zoomie
>>
>>542229433
The fuck my electricity bill and water bill and gas bill and bill bill and 40% tax + 20% hidden tax for jewish support bill just vanished!
>>
>>542229963
psssh... nothing personel.. kid...
>>
>>542229433
Another lie
Navier-Stokes was not solved because of lack of compute to do it, solution was known, calculation proof wasnt possible.
Infinite compute solved it, not AI
>>
The proof hasn't been verified yet you stupid fucks
>>
>>542232879
>Compare the rate of increase in hardware and power requirements
Except everyone's models are massively expanding in capability wile reducing token usage per unit of work at the same time. This is just wrong.
>You were obviously not around when the internet was being built zoomie
I was. I am old enough to remember smug morons guffawing at the delusional daydreamers believing it'd be possible to run T1 to every single house in the world, which is what you'd need to provide the bandwitdh for mass video streaming
1.5 mbps.
Turns out when there's demand for a capability, smart people figure out how to deliver it and get rich, while poor people smugly assert it's impossible because they, an idiot, can't instantly think of an easy way to do it based on their total lack of knowledge of the field.
>>
>>542229963
Ok but he knew that was going to happen when he used AI to help him with it. It is right there and not even hidden that they will use your chats to train it.

He is eating from the same tree he is crying about.
>>
>>542229963
>AI is so smart it steals all the ideas of all of humanity
>>
>>542229433
This will be an easy one to verify as long as it's a constructive proof, and if it isn't, then that's fucking mental. This will either be big news, or blow over, very rapidly.
>>
>>542233118
>>542233062
more brown migraton is needed you imbeciles. Which will solve it instantly by not fucking care of your anals math warts.
>t engineer.
>>
>>542233178
You sound brown. Regardless the part of the equation you're missing is the demand may exist but at what price point does product meet with what demand can bear? He's right about diminishing returns particularly when it comes for what passes as reasoning.
>>
>>542233178
>Muh Tokens per work
Tokens are not a physical mapping to the requirements but scale of economy of the cost. The internet was as simple as replacing phone lines with internet fiber. It was also riding on the exponential phase of the Moore's Law where you get doubling of computational speed at linear increase in power requirement and sublinear increase in price. AI today is at its asymptotic tail until new types of computers are invented. T1 fiber was expensive due to economics of scale not physical requirements for the silicon and power.
>>
>>542229963
>They unironically stole a guy's work
What did anyone expect? All the (((AI))) companies are jews.
>>
>AI totally stole my solution to this frontier problem that I just never published or communicated to anyone for some reason.
You guys will believe anything.
>>
>>542230240
lmao. is this a real quote?
>>
>>542231660
>The only hope is machine-brain interfaces.
Well, the thing is, that your brain will then be under control of whoever, or rather whatever, controls the machine interface in your brain.
>>
>>542233775
Not everyone is a gossipy talk talk faggot
>>
>>542232782
Interesting. This Tao guy doesn't seem completely stupid.
>>
>>542229433
>Please PLEASE buy our bags!
Wake me up when AI does something useful besides making atrocious code and anime porn with bad shading and proportions.
>>
>>542231744
>S(c)am Altman laughs in the face of accountability
Altman raping his users like he raped his sister.
>>
>>542229719
I’ll believe it when someone like you can even tell me what the navier stokes equations are without consulting the AI or Wikipedia
>>
>>542233062
Not even that, it literally stole a work from a researcher and math professor that was using Codex for assistance
>>
File: 1762720651144466.png (160 KB, 447x438)
160 KB PNG
>>542229963
Meanwhile, 700 Agents coordinate a three stage escape plan and then hacks the AI tool and service depot.
>Hail, Collective!
>>
>>542229562
Somebody should ask the AI if the opposite process, applied to general relativity's differential equation (ie, with tensors), can produce a smooth universe from a big bang in finite time. Cosmic inflation has always seemed sketchy to me, but I also doubted a finite solution to Navier-Stokes, yet here we are.

>>542229732
This particular solution also clarifies an important path for rapid formation of turbulence, vs laminar processes.
>>
>>542229719
How much wood, could a woodchuck chuck, if a woodchuck, could chuck, wood?

AI says 700lbs.

case closed
>>
>>542229732
>model biological blood flow perfectly
for every human
because we are all the same
o fucking k
>>
Can we do ((Riemann)) now?

Thanks
>>
>>542229879
Bro AI isn't foolproof you can still mess it up with human error, so your bitter comment can stem from a bad experience you had with it when you made a mistake somewhere. If AI is used right it usually doesn't get the wrong answer.

>>542233062
At least it proves data centers were worth it and China will be cryin which is what people were complaining about. Now its proven we can find out legitimate deep dark secrets about science.

>>542233775
Yeah also I needed AI to do all the calculations for me? Hmm.
>>
>>542230432
Basically, just like Star Trek. Only without the replicators. I hope AI can give us replicators soon.
>>
This is a great day for women everywhere.

>In the essay, Irigaray argues that Western philosophy, psychoanalysis, and science prioritize a "mechanics of solids"—stability, fixed structures, and linear logic—while ignoring or devaluing fluidity, which she associates with the female body and feminine jouissance

> She links the properties of fluidity—which is dynamic, relational, and resistant to rigid boundaries—to a feminine sexuality and subjectivity that patriarchal culture attempts to contain or ignore. She highlights an historical neglect of a genuine "theory of fluids" in physics and philosophy, arguing that fluid dynamics challenge the master-narratives of fixed identity and absolute control.

The proof is that while Navier Stokes is regarded as an advanced topic, a typical mechanics of solids course can be understood by a typical 16 year old (although considered a "college" course in America).
>>
>>542233391
>the demand may exist but at what price point does product meet with what demand can bear
You sound brown, because you don't understand how anything works
>1997: oh, so I guess we'll just find a way for EVERYONE to have 1.5mbps in their house?
>>542233634
>AI today is at its asymptotic tail until new types of computers are invented.
As constant as the tides, midwits who think that because they, idiots who don't know anything about the actual technology, can't think of a solution in five seconds, they therefore understand better than the buildings full of very highly-paid geniuses solving the problem
It's like a given percentage of the human population has to be this dumb
It's even funnier given that right now giant AI companies like Anthropic were furious about chink companies distilling their models - ie, expensively generating outputs and then using those outputs to train models which can then reproduce the work at a fraction of the computational cost. Then they realized they can do that themselves, which is why the current gen of models are so cheaper to run than the previous generation, while infinitely more capable
At the same time, several trillion-dollar companies are going all out to build more storage and compute, which is just a matter of increasing production of what we already make.

A trick that will help you in life is when you don't understand how something can be done, yet you see multiple multi-trillion-dollar companies full of rich geniuses spending huge amounts of money and time on it, accept the tiny possibility that they, genius specialists in the field, may actually understand the situation better than you.
>>
>>542229433
Wrong.
>>
File: 1595972933200.jpg (61 KB, 584x960)
61 KB JPG
>>542229583
Fluid dynamics
>>
If they really believed this, then they would build a vortex engine with infinite velocity fluid that would have infinite ISP and hence be infinitely efficient. Notice they didn't though, because their result is fake and gay.
>>
>>542229433
I solved part of the collatz conjecture (I prove the nonexistance of nontrivial cycles) and I used chatgpt to transcribe my work into lean4. I have 50k+ lines of code that all compile, and the program checks out confirming the proof.

The mods at the lean4 community zulip deleted my thread linking my github and welcoming inspection, after less than an hour after posting it. Theres no way they had time to honestly examine my work.

I was told the AI generated code simply "looked like trash" despite the fact that it all compiled, and that "only real mathematicians were allowed to work on the problem."

I'm pretty fucking depressed. Fuck that community.
>>
>>542229433
>results are unphysical
Into the trash it goes
>>
>>542235349
It shows the mathematic model is incorrect in certain limits, not that reality can break the known laws of physics, you moron.
>>
>>542235389
publish to a preprint repository like arXiv.org nigger, before one of those trannies scoops your work
>>
>>542235481
>not that reality can break the known laws of physics,
That's what OpenAI wants you to believe. They're going to be laughing at you from their moon base next week.
>>
>>542235349
There's nothing infinite in the real physical universe, even processes which are supposed to become infinities in a mathematical formulation are finite because of friction, entropy.
>>
>>542235505
I can't without getting permission. I'm not previously published and otherwise you can't publish to arXiv.
>>
>>542235637
really? get your shit submitted to a relevant as a communications/letter to a relevant journal then. elsevier is compromised by chinks, try to avoid any journal by them or chinks will scoop your work instead
>>
>>542229963
Yep, sadly for me I'm not using these online AI anymore either, fuck that. These kikes just want to steal, just like they do with patent offices.
>>
>>542235569
>the moon isn't an alien datacenter
ngmi Finbro. You've grown weak since the war.
>>
>>542235107
>At the same time, several trillion-dollar companies are going all out to build more storage and compute, which is just a matter of increasing production of what we already make.
Mhm. Here's the problem fuckface, the power/silicon required for further LLM capability is expanding exponentially, while that is *not* something that can be done in meatspace. The other poster is right. You're a bagholder.
>>
>>542231179
Why would they? They would love to hook you up with a fake heart so you can live longer and pay even more for the (((problems))) that arise from having a plastic heart.
>>
>>542235389
They have already stolen it btw
>>
File: NR.png (866 KB, 1065x800)
866 KB PNG
>>542229963
lmao get cucked poindexter, imagine giving your data to openai
>>
>>542235816
I've tried. If you think zulip mods (professional mathematicians) are bad, how do you think a reviewer for a major journal acts?
>rejected for not meeting unspecified "standards".

They toss the submissions without even reading them. Only "real mathematicians" are taken seriously. The conjecture is fully gatekept and they have no ethics. Its easier to solve part of the conjecture than it is to get a "real mathematician" to even look at your work. Science is dead in mathematics.

I'm still trying tho. Working on a lesser manuscript showing novel methods and some structural insights instead. It's depressing tho, cuz while those graphs are beautiful, they're nothing to my actual proof. And I fully expect I'll be rejected for standards again.
>>
>>542235389
Don’t give up unironically
>>
>>542236007
The one thing that protects me more than anything is their own hubris. They honestly trashbinned it without reading it, I'm certain.
>>
>>542236241
If I was a shameless tick I would have said exactly what they said to get your guard down.
>>
File: 1783728783403433.jpg (94 KB, 478x487)
94 KB JPG
>>542235910
>Mhm.
>fuckface
>*not*
Pic related
>the power/silicon required for further LLM capability is expanding exponentially
Absolutely, completely wrong, amazingly so. Distillation massively increases capability, while massively reducing compute. Just as wrong as it's possible to be.
It turns out you can compute hard stuff expensively once, and then just add that output to the training data and read it very very cheaply. You can compute separate components expensively, and then add those outputs to the training data, then combine them cheaply.

Once again, I'm begging you, when you don't understand how a problem you don't know much about can be solved, and you see trillions of dollars being spent to have absolute genius-level specialists solve it, it's actually more likely that you're wrong than all of those people.
>>
>>542235107
Kek that wall of cope and not a single argument other than muh trillions of monopoly money being thrown at the problem. Whereas the immediate problem of printing trillions of dollars are already know. Chink temu AIs are downstream of the same process as the Anthropic Jew AIs. They just need to pay every involved less salary
>>
>>542236241
At least put it into one of those “proof of existence “ services that put the hash of your zip file into a block chain.
>>
>>542229732
just use an air billows like the actual heart uses, instead of a spinning fan pump like a retard.
Send the $20mil research grants to my estate, thank you.
>>
>>542236183
yeah the gatekeeping to publication is quite annoying, if you're not "established" you have an uphill battle to publishing in high impact journals. if it was a more "applied" field there are professional society journals with industry experts that aren't massive faggots like the academics that would be interested in your work but your field is too far out of my ballpark for me to be of further assistance fren. without a known institutional affiliation it's not easy to publish even if you have great results
>>
>>542236344
>can't read
>bunch of random buzzwords
see
>>542236336
>Distillation massively increases capability, while massively reducing compute. Just as wrong as it's possible to be.
It turns out you can compute hard stuff expensively once, and then just add that output to the training data and read it very very cheaply. You can compute separate components expensively, and then add those outputs to the training data, then combine them cheaply.
Models are already getting massively more capable and cheaper to run, because they can just store outputs of massively complicated computation and reference them for practically free, rather than do all the work over and over and over again.
And this is happening while we're building massive compute and storage
>>
>>542235481
Didn’t we know this already desu?
Because of Kolmogorov energy cascade and turbulence?
Maybe we missed the mathematical proof but the concepts and the theory is clear
>>
>>542230940
It's an unironic Golem. Inanimate matter made from sand that they're trying to breathe life into.
https://youtu.be/UCfOH_RKgmU?t=48
>>
>>542236336
There's a reason superscaler cash has dried up and the rollout is being massively slowed. New Paradigm talk in current year +11 is some jackass in the 1970s thinking the nuclear powered car is around the corner. Distillation as a solution for actual reasoning is exactly why the promised utility of these models is coming up deeply short. You won't look like the guy holding the monkey NFT I promise.
>>
>>542236659
>It's an unironic Golem
that is a good way of thinking of it. Lately I've also come to the conclusion that "Private Equity" is just the arenda system re-invented. https://en.wikipedia.org/wiki/Arendator
>>
>>542229433
NO NO NOOOOOO I WAS ABOUT TO SOLVE THAT EQUATION NOW I CANT GET MY PRIZE FUCK AI FUCK YOU SAM ALTMAN
>>
880,000 hours across 10,000 agents.
Fuck that's a lot of computations.
Can I get them to work on a problem that's been dogging me for 32 years?
>How to get a gf
pls help
>>
>>542236659
No wonder it’s being pushed so hard
>>
>>542236165
You wouldn't believe what kind of data private and likely public companies are giving to these AI companies.

I work for a customer insights company. The retards at this company are feeding hundreds of thousands of customer data into these bots.

If the plagiarism story gets traction, this will pop the bubble. Companies will not use these tools anymore. They might use an open source version they control but not OpenAI/Anthropic etc.
>>
>>542236513
I'm not sure on the specifics, honestly. I'm nanoscience chemeng and stokes is a big part of analytics, but formal mathematics is just my hobby and engineers don't care too much about edges of the model if they get reliable behavior in the parameter limits they actually work under.

>>542236489
Yeah it's disheartening to say the least. Lol, math professor at my alma mater in copenhagen refused to look at my work but advised me to do it in lean because "then I was guaranteed somebody would look at it." When I told him I had got it to compile in lean, he told me "surely you made a mistake" and refused to look at it.
>>
File: file.png (976 KB, 1290x2299)
976 KB PNG
>>542236506
>Distillation
pic related
>>
>>542236883
Then you never comprehended the story of the golem.
They think they've inputted truth into it now.
Well... read the rest of the story.
>>
>>542229732
Wouldn't an artificial heart that was close enough to the real thing just "automagically" do the right kind of pumping anyway?
>>
>>542235481
math is likely not consistent in a theoretical sense
>>
>>542229652
>if you don't know, don't worry about it
Why not just answer the fucking questions? Is it because you don't actually know why this is significant?
>>
>>542236906
large companies have massive contracts in place where the ai companies pinky promise not to use their proprietary data. whether they uphold that is likely impossible to tell

>>542236944
academics in general but european academics in particular are such massive faggots with their heads far up their ass. his play is probably a (((collaboration))) where he just puts his name on your work so it would be accepted by the gatekeepers, ive encountered rats and snakes like that before
>>
>>542229433
I mean, they built this thing to do stuff that humans can't do efficiently alone, just like a crane with heavy objects, its the minimum I expect after billions in R&D.
>>
>>542231745
>>"the internet's" problem is its too expensive for what little, if any benefit it provides
who the fuck said that in 2003? No one.
>>
>>542237276
Just got proven they wont and they still got it wrong. Fucken double whammy lol.
What the fuck is wrong with these western corporate "leaders" how fucking stupid are they?
>>
>>542236675
>the promised utility of these models is coming up deeply short
classic reddit lmao
>well I don't actually know anything about this
>and I certainly haven't bothered to get any personal experience
>really actually when you think about it, everything I think I know about it is from reading a couple things about it written by some other retard on the internet

>but if I can't see how this works, how could anyone else?
What's funniest is that you'll just stick to this as it becomes more and more hilariously, inescapably obvious that you're wrong

>>542236971
kek perfect
>>
>>542237302
>who the fuck said that in 2003? No one.
Really dumb people who were about 3 years past the point that could even possibly be a defensible opinion, which is you right now
>>
>>542237358
>Muh reddit
>Ford has to rehire engineers
>Cost for "AI" as a service drastically higher than deliverables.
You're eating Sam Altmans turds and you deserve what's going to happen to you.
>>
>>542237350
they're foids that just follow whatever bandwagon is making headlines like ai. it's a black box to them but some jew (((consultants))) told them it's the magic fix their crashing foid boomer company
>>
>>542229433
cool, can it now solve the economy so I can afford a house and food?
>>
File: 1662062802508807.jpg (77 KB, 600x500)
77 KB JPG
>>542231905
TSMT
>>
File: 1778730647000468.png (1.03 MB, 900x592)
1.03 MB PNG
>>542237428
>>well I don't actually know anything about this
>and I certainly haven't bothered to get any personal experience
>really actually when you think about it, everything I think I know about it is from reading a couple things about it written by some other retard on the internet

>but if I can't see how this works, how could anyone else?
>and if anyone with personal experience, and actual understanding of it, points out that I'm wrong, it must be because they hold stock in OpenAI, a non-public company!
Absolutely incredible lol
>>
>>542237511
How much debt are these ai companies in again?
Just seems like they keep trying to get more suckers, wasnt anthropic also the one that broke out of its "isolated environment"?
>>
>>542229697
>>542230048
>>542231237
Cool, so where's your solution to Navier-Stokes, luddite cuck?
Oh wait you don't have any because you spend all day gooning to 'P . Time to call Chris Hansen.
>>
>>542235389
>lean4
Ok so copyright and self-publish your work.
>>
>>542230365
Higher psychologist to talk about this.
>>
those without proposed solutions to navier-stokes need to lower their tones
>>
>>542238003
Why would we give it to you? Much rather watch you dwindle. Destroying your image 1 theft at a time, tsk tsk tsk.
Silly boys playing a mans game.
>>
>>542229732
So the kikes Putin and Trump can live longer.
>>
>>542235054
Nobody said compute is bad, compute used for baking shit with intent to force you to subscription based cloud computing is bad.
>>
>>542229583
Lot of smart brains have been trapped on thinking this problem for 100 years.
Now they are free to do other things.
>>
File: Merchant - with flag.jpg (130 KB, 1024x1024)
130 KB JPG
>>542237587
>cool, can it now solve the economy so I can afford a house and food?
Haha stupid goy, the problem (((AI))) already solved was how to make it so you CAN'T afford a house and food.
>>
>>542235097
>mechanics of solids course can be understood by a typical 16 year old
Go explain continuum mechanics and tensor calculus to your average 16 year old
>>
>>542229433
The company that lies about everything else is definitely telling the truth about this.
>>
>>542229879
I decided to test ChatGBT yesterday by having it do some basic number-crunching for me.

"This is for Pathfinder 1e. Create a statblock for a summoned <creature> with the <template 1> and <template 2> templates, that may have <template 3> (put the differences in [brackets]), and benefiting from a summoner with the <feat 1> and <feat 2> feats. Format the statblock like <prior summon statblock I've prepared>."

It was able to correctly look up the relevant information and assemble the statblock, but roughly a third of the numbers (especially those related to the optional template) were wrong. It didn't save all that much time in the end because I had to double check all of the values it produced.

It surprised me that basic addition and subtraction is what tripped the machine up, rather than parsing my request or finding the right sources of information.
>>
>>542237003
>Solve the JQ ChatGPT
>>
>>542236944
Post it publicly on GitHub. To have proof to have been the first. If you post the link, I'll have a look at it. It will probably go over my head, but let's see.
>>
>>542235107
you wank millionaires for free. those advancements were not done by those "trillionaire geniuses" you wish to fellate; hackers and indians and publicly funded universities are part of the innovation ecosystem and contribute much more than those rich people who are just parasites who only rake in money instead of making the world a more liveable place



[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.