>>17049648
Still waiting on your tensor logic master algorithm that unifies all five schools of AI Pedro

Although it took the AI bots seemingly little time to reach a solution, OpenAI said the bots exchanged nearly 3 million messages and used up 130 billion output tokens, or the individual lines of text and code an AI model produces in answers, on Navier–Stokes alone.

Such an effort would have cost roughly $10m (£7.3m), based on OpenAI's own pricing for output from its most advanced models.

I don't think people at home can compete if it needs a huge amount of resources thrown at it.

>>17049662
One of the arc agi tests cost OpenAI tens of thousands to top the leaderboard when it came out, now two years later it costs them two hours of min wage to get a much better score

>>17049665
>now that they've seen the solution it takes much less time to find it
ummm, yeah?

>>17049668
>literally a new unreleased test set
That’s not how it works dumb techbro

>>17049669
it cost them tens of thousands for ARC-AGI v3
they solved ARC-AGI v1 and v2 with the low cost you referenced
v1 and v2 have been around for a while

>>17049673
Nobody cares techbro
Stop stealing from artists

>>17049675
>outcompeting = stealing

>>17049648
i used some average values for how many tokens there are in a sentence, time it takes a human to write a thoughtful sentence (which is absolutely certainly a lowball), and the number of tokens it apparently used (this anon's number >>17049662), and by my calculations, it took 16,307 (rounding down) years worth of thinking/writing, and although i'm sure that number would go way the hell down if split across many people, considering the problem has only existed for roughly 200 years, this really does not seem very impressive. it's cool the problem is solved, but it does not speak at all to a meaningful mathematical capacity imo.

>>17049648
Midwit take.
The fact that we have one counterexample for smoothness has no practical relevance to aerodynamics, climate, biomedical research or wind turbines. There's not a single engineer on the planet saying "finally, we can make our gizmo work!".

>>17049684
In other words, this model was able to perform ~16,000 years worth of human thought in less than 1 year?
And this is supposed to minimize the accomplishment in your eyes?

>>17049659
>>17049669
>>17049675
>>17049686
migger cope

>>17049689
yes, it would take humans far less man-years to solve it themselves. like i said, it's neat that LLMs can do this, but it doesn't reflect well on their capacity for mathematics.

>>17049668
>>17049692
Imagine worshipping a maga chud
Dumb incel

>>17049689
>less than 1 year?
Less than one week actually

>>17049710
>it would take humans far less man-years to solve it
That's not really relevant, is it? Mathematians have used computers to literally brute force solutions to difficult problems for about as long as computers have existed.
That a more computationally efficient method might exist is irrelevant in the face of results.

>>17049724
relevant to what? literacy test: tell me what you think i'm saying across my two posts.

>>17049934
I think you're probably saying that youre gay mate

>>17049934
>tell me what you think i'm saying across my two posts.
That due to the vast number of effective man-hours worth of "labor" the AI/research team dumped into the problem, the fact that they eventually came to a solution is unimpressive. If these things were "good at math," they wouldn't need such a gargantuan amount of labor to find their solution.

An analogy for your argument might be something like building something on the scale of the Great Pyramids by stacking rocks by hand. The feat of "engineering" would be much less impressive since it's just an example of naively throwing massive amounts of elbow grease at the ptoblem.

My response was: the simple fact that so many man-hours can be performed by this model in a short amount of time is precisely what makes the model impressive regardless of whether an actual human would be capable of finding a more efficient method. You have yet to provide a meaningful rebuttal.

If I have misconstrued your argument, then consider whether or not the problem is that you haven't communicated it effectively.

File: kbeg3958gzbe1.jpg (137 KB, 1080x1274)

137 KB JPG

>>17049648

>>17049648
can't wait for AI to solve all the millennium prizes and for literally nothing in our lives to change.
have it solve something useful like nuclear fusion.

>>17049972
Why would anyone use AI to upend a trillion dollar energy industry? That's just not profitable compared to solving pure maths.

>>17049967
>the Great Pyramids by stacking rocks by hand
>>17049967
>the simple fact that so many man-hours can be performed by this model in a short amount of time is precisely what makes the model impressive

NTA but you're almost getting it: it all boils down to the simple fact that there's only a tiny number of world-class geniuses who can even approach unsolved problems, but billions of human cattle like you who can keep shoveling coal into the magical "thinking" machine's steam engine.

The people who print the money don't care if they're destroying the foundations of human ingenuity in favor of "building the Great Pyramids by stacking rocks by hand". They have no minds minds or souls, so they can't contemplate the long-term implications of what they're doing. Since they can print as much money as they like and only make YOU poorer in the process, it doesn't matter how absurd and inefficient their "AI renaissance" is. They will keep it up so long as they believe they can scale the brick-stacking by fiat and get results faster.


In other words, it's a scheme to repurpose a vast economy (whose power goes mostly into producing plastic trinkets for useless eaters like you) to accelerate a certain philosophically inept vision of progress and make a certain group of people into technological gods. Imbeciles like you will be clapping like trained seals all the way to the inevitable collapse of this scheme. It really is impressive.

>>17049967
You mean spending a bunch of money on graohics cards is the monumental achievement?

>>17049985
>there's only a tiny number of world-class geniuses who can even approach unsolved problems, but billions of human cattle like you who can keep shoveling coal into the magical "thinking" machine's steam engine.
Just like there's only a handful of people skilled enough to hand-carve a wooden chair while anyone with a working pair of hands can use mechanized carpentry tools. Technological advances usually have the effect of lowering the barrier to entry.
You may be right about the ethical implications of the gargantuan investment into these models. But that doesn't mean you can't also be impressed when you see the Rock Stacker 6Million pushed to its limits.

>>17049648
From what I understand negative result has literally 0 implication for the real world.
Positive could, but only if next to providing proof you also found an analytical solution that would prove faster and more accurate than current methods.

So, basically theoretical physicists/mathematicians are egomaniacal scammers lying to the world about their importance and constantly overblowing impact of their fields.

>>17049993
>Just like there's only a handful of people skilled enough to hand-carve a wooden chair while anyone with a working pair of hands can use mechanized carpentry tools
Incoherent response. Notice how your subhuman "mind" is incapable of engaging with anything I wrote and goes straight for the corporate thought-terminator marketing cliches.

>>17050004
I referenced your ethical disputes by saying they are secondary to the discussion. Yapping about the elites siphoning investor funds into their deus ex machina and how it's making us all poor abd subservient might sound like a cool and enlightened take to you but it's just banal crying without any meaningful substance.

>>17049967
>An analogy for your argument might be something like building something on the scale of the Great Pyramids by stacking rocks by hand. The feat of "engineering" would be much less impressive since it's just an example of naively throwing massive amounts of elbow grease at the ptoblem.
No, in fact, it's exactly like building the Great Pyramid, no qualifier. It's taking known monumental engineering techniques (of the time) and competently and tirelessly scaling them up.

>>17050013
Notice how you are literally unable to comprehend the post and are forced to do a primitive closest match to the standard arguments of the botted "opposition", which ends up with you shitting out another incongruent "refutation" from the corpo database.

>>17050033
I have respoded to the most plain reading of every meaningful "point" you made. All you can do is cry about how I "just don't get it" because I find your take rather shallow and boring.

But please: on the off chance that I am misinterpreting what you wrote then go ahead and reformulate your argument in a way that's actually intelligible.

>>17050038
I simply explained the actual mechanics of this "impressive" feat. There is no free lunch with a magical AI oracle. Only a new mechanism for converting slave labor into mathematical results. It's certainly an interesting innovation, just not in the category of intelligence.

>>17050054
So your posts reduce to a vacuous non-argument about the token guesser not being "intelligent." That's nice.

I guess I was too charitable with my reading earlier. There's just nothing here.

>>17050063
Notice how you are literally unable to comprehend the post and are forced to do a primitive closest match to the standard arguments of the botted "opposition", which ends up with you shitting out another incongruent "refutation" from the corpo database.

>>17050065
Notice how you just repeat yourself in essentially saying I "just don't get it" when I respond to the most plain reading of your own words.

Here's a line-by-line breakdown of your last post:
>I simply explained the actual mechanics of this "impressive" feat.
This is strictly false as you also mixed it in with a bunch of moralistic posturing.
>There is no free lunch with a magical AI oracle
Nobody said there was. The vast amounts of resources being dumped into these models is well documented and sensationalistically publicized
>Only a new mechanism for converting slave labor into mathematical results
More moralistic posturing.
>It's certainly an interesting innovation
I agree.
>just not in the category of intelligence.
Subject to the definition of "intelligence" at play.

In other words: you're not saying anything. You're just crying.

>>17049985
>it all boils down to the simple fact that there's only a tiny number of world-class geniuses who can even approach unsolved problems, but billions of human cattle like you who can keep shoveling coal into the magical "thinking" machine's steam engine.
>it doesn't matter how absurd and inefficient their "AI renaissance" is. They will keep it up so long as they believe they can scale the brick-stacking by fiat and get results faster.
/thread. by any objective metric they are, bruteforcing these problems (ignoring the highly plausible accusation of cheating). it just involves so many layers of social and economic abstractions that normie AI believers can't see it, they only see the GUI for it, which presents them with a singular entity that thinks, instead of an entire worldwide economy feverishly churning numbers

>>17050071
Notice how you are literally unable to comprehend the post and are forced to do a primitive closest match to the standard arguments of the botted "opposition", which ends up with you shitting out another incongruent "refutation" from the corpo database.

>>17050071
Protip: The braindead retard you're responding to is a schizophrenic moldovan loser who keeps seething about AI because it amplifies this retard's rabid fear of materialism/physicalism.

>>17050086
meds unironically

>>17050086
Ironically it's atheistcucks and pathetic loser progressives who are angry at ai.

Religious and spiritual people know human identity is on inner experience, a machine is not a threat to them,they actually use and enjoy it.

>>17050095
The only people involved in both the pro and anti AI side are atheists. Religiotards are irrelevant and too stupid to think about any of this.

File: brainlet21.png.png (210 KB, 1024x1024)

210 KB PNG

>Religious and spiritual people know human identity is on inner experience, a machine is not a threat to them,they actually use and enjoy it.
MIGAtards are genuinely mentally ill.

File: redditsoi.png (33 KB, 584x841)

33 KB PNG

>The only people involved in both the pro and anti AI side are atheists

Lol, the moldovan melty started. This will be hilarious.

>>17050095
Ok, retard, guess at least you'll have God and your inner human identity when they start actively culling you.

File: Screenshot_20260905_13580(...).jpg (397 KB, 1079x1148)

397 KB JPG

>>17050099
You are proojecting your own irrelevance because you believe humans are just dead atoms anyway.

Now the dead thing is your shitty degree.

Sorry your nerd pseudo academic views have no future. Find another hobby.

>>17050116
Ironic the "sky daddy" atheists have no problem believing science fiction is prophecy about ai.

>>17050151
>Ironic the "sky daddy" atheists have no problem believing science fiction is prophecy about ai.
You can continue that argument with the voices in your head. The people who rule over you and own the technological infrastructure will have no further use for you once they're done milking you and the current economic order for the breakthroughs they need to kickstart gay luxury space communism for themselves and their own spawns. They don't believe in any god. They're open about culling you and have been for decades.

>>17049648
>Navier Stokes very important saar

>>17049648
It has more relevance than Poincare.

Cope

Why are NPCs so focused on "relevance"? Do they ever question the relevance of jerking it to BBC porn?

>>17049648
Why do mathematicians think ideal fluids have anything to do with real fluids?

File: 1738151016373241.jpg (9 KB, 224x224)

9 KB JPG

>>17049648
>le mathematics le breaks down in a le idealized conditions system with enough time
>real world? oh yeah every system is open and the conditions update all the time