â„–3947606[Quote]
>>3947593 (OP)1 billion in revenue daily thats why
â„–3947800[Quote]
>>3947606Amazon has 2 billion in daily revenue
â„–3947805[Quote]
Amazon has the highest daily revenue of any company
â„–3947808[Quote]
They rent their computing.
â„–3947810[Quote]
Ai
â„–3947811[Quote]
Surely niggervidia wont be this big for long
â„–3947814[Quote]
>>3947810Niggers dont even know what ai is or how to actually use it
â„–3947816[Quote]
>>3947811Nah, the flood of cash will probably slow down, no bubble will seriously pop, and robotics and true fantasy ai will exist within the next 10 years, EASY
â„–3947821[Quote]
>>3947816Yeah but boomers are the ones who are optimistically speculating on this the most because they dont know what it is even slightly, as someone who is in the ai trenches i dont think the technology is all that
â„–3947823[Quote]
>>3947816Niggas believe this and now claude is valued at 30 trillion usd, while there biggest product is a google chrome add on for browser.
â„–3947825[Quote]
>>3947593 (OP)Stock speculation and overvaluing. The same way a glorified government contractor can make a record breaking IPO and piss it all away before the trading week is over
â„–3947832[Quote]
>>3947825It isnt just investors who are overspeculating on nvidia, it is investors who are overspeculating on all of the companies that buy nvidia gpus also. Its a gigantic circle jerk with nvidia in the center
â„–3947837[Quote]
>>3947821I am also in the ai trenches, and it is literally definitionally agi, right now.
>>3947823Geg, you gotta think about it like this. You give ai a realistic world environment, and tell it to do science, and you find new physics and chemistry and math (already happened with math).
â„–3947842[Quote]
>>3947837Math is not automated. AI cant reliably do math and llms probably never will.
â„–3947844[Quote]
>>3947837Why you think its agi because it glazes you and tells you you are super smart and powerful and you believe it?
Have you tried asking it to do any moderately complicated task with regards to writing software? 9/10 times it fucks it up royally.
â„–3947848[Quote]
>>3947842The true value nowadays is not in the AI but in the harness tje more powerful your harness the better your AI
â„–3947850[Quote]
>>3947842Alright bub
>>3947844It meets the definition of the term. Before the goalpost got shifted 100 times in the past year.
â„–3947851[Quote]
>>3947837Can agi make a simple fucking 3d model hexagon in cad without hallucinations and artifacts? I havent seen it
â„–3947852[Quote]
>>3947844Ask it to write a simple programm and ask it to change a small detail. It will completely fuck up and destroy your program.
â„–3947854[Quote]
>>3947850Stupid nigger. Ill take your gay post and move it 100 yards down the field so you miss
â„–3947858[Quote]
>>3947850It is true. I know mathmatics it needs harnesses to do math, which is basically brute forcing, but that obly works so far in math.
â„–3947859[Quote]
>>3947851Can the average person make a 3d model hexagon in cad without any mistakes? Agi is GENERAL intelligence, 100iq, average person intelligence. Almost ZERO people can actually do that.
â„–3947861[Quote]
>>3947852Yes that is why you have to switch to an experimental branch in git before asking ai to do anything, so when it fucks shit up you can revert back
â„–3947869[Quote]
>>3947858>it needs a harnessDo people need calculators and math models? Why don't people just use pencil and paper, if you can't do high level on pencil and paper, you aren't a real mathematician
â„–3947875[Quote]
>>3947869Harness is brute forcing in math. I can do higher mathematics with just pencil and paper?
You know nothing about math yet you still wont keep your mouth shut.
â„–3947877[Quote]
>>3947869Its niggervidia
Its all just nigger licious as king terry would say. Too bad he isnt alive to see this
â„–3947883[Quote]
Jensen Mong and Elon are lying to the world when they talk about what they think ai will be capable of
â„–3947884[Quote]
>>3947861Yeah I would rather like to code myself.
â„–3947887[Quote]
AI is for people who like cuckolding.
â„–3947889[Quote]
>>3947875>harness is brute forcing in mathExplain what you mean, maybe I'm misunderstanding
>>3947877You asked question, I gave answer, and you leak, telling me I'm wrong. Why did you ask the question if you were only going to leak to any real answer?
â„–3947893[Quote]
>>3947889Im not leak i was just adding on. The truth lies somewhere in the middle between critics and proponents of ai
â„–3947903[Quote]
>>3947893100% agree with this sentiment. I'm simply stating that the classical definition of agi, the models we have, meet it 10x over. It is well within the intelligence of the average person. Now is the leap from agi to asi going to be as fast as the theories state? Doesn't really seem like it so far.
â„–3947906[Quote]
The shit that Jensen says about AI is beyond psychosis. The guy lives in a golden echo chamber where AI whispers to him whatever fantasies are most likely to be engaging to him
â„–3947912[Quote]
>>3947906As does all ceo's if we are being fair
â„–3947929[Quote]
>>3947889Well suppose AI wants to prove a math result. In an objective mathematical sense this just means going from point A to point B in the math world. However we do not know the road from A to B.
Human mathematicians will use their intuition to get efficiently there.
A math AI lacks that intution and via agentic harnesses tries many roads, the road most promising to B will get copypasted and the process begins anew until B is reached.
Until now what we have seen from AI are very basic roads being built to say it in a non math way. Menaing only one smart step was needed to solve the problem. The thing is most problems are far more complex, thats why even rather easy research problems are out of grasp for AI, the results they had were curated after many many hours of teying and brute forcing the whole week. So I am very positive that I wont be getting replaced.
â„–3947941[Quote]
>>3947903It is an extremely powerful research tool
â„–3947950[Quote]
>>3947903The only definition of AGI I know was the one in 2018 saying that it needed to outperform experts in any field to be an AGI.
â„–3947964[Quote]
>>3947950Yeah thats what i thought. This guy probably asked chatgpt if it is agi yet and it said yessir
â„–3947965[Quote]
>>3947929OK, I see your point. And it makes sense to me. I have two things, 1 is it technically bad that ai brute forces math problems? Isn't it good that these problems, no matter how minor, are being solved?
And 2
>I won't be getting replaced.This is what everyone said about chessbots until they stomped the worlds best players over 30 years ago. Now, still in your favor is the human spirit and novel creativity, which is your intuitive point. I guess all I'd say is, use the calculator, it does the thing. Its like a assistant to me, and it gets most things right 90% of the time (all local ai too, not frontier), I'll never expect to gain access to the smartest models, they are weapons of war.
â„–3947967[Quote]
>>3947964They conveniently changed the definition to fit their marketing. Although AI is very useful it is grossly overvalued at the moment. We will probably see some sort of crash in the next three years and then we will probably only have like some better or worse chinese open models left.
â„–3947981[Quote]
>>3947965The problem is in my field it is basically useless. As of right now frontier models in addition to large compute times are needed to get at least some minor results.
It is not technically bad that AI brute forces math and I could care less if it solved even the biggest problems. I do math to challabge myself no matter how small the problem.
The thing is just I kind of see no point in automating math. Due to the following two reasons:
1. it is legit not profitable, nobody will pay any AI company to do math research. Nobody is getting paid to do math research. Every mathematician is being paid to be a teacher at a university only like 10% of the job are research. Thats like 5 hours per week.
2. there are provably infinitly many math problems, meaning that letting an AI automatenit would mean that it never stops. There is no reason to automate a process that has no end.
â„–3947985[Quote]
>>3947967I wouldn't count on a complete crash because companies will still use it for its cost effectiveness
I sometimes feel like llms are a step in emulating consciousness but it is hard for me to say how big of a step they really are
If I ware to invest in things I would invest in wetware technology
â„–3948000[Quote]
>>3947985The thing is I see this whole thing more globally. Many people hate AI data centers and it will become more and more difficult to keep building new ones. The old ones need water, but they also make it dirty we have already see many protests and closings on this.
Also I think we are at a deceisive point in history. Either we achieve ASI in the next 20 years or we will be set back by climate change so far that it will take us another 500 years to come back here at this point.
â„–3948004[Quote]
>>3947981We do technically automate processes that don't have a foreseeable end like production but I get your point that it wouldn't really make sense to automate pure math
â„–3948013[Quote]
>>3948004I mean you can automate it but nobody would care. Mathematicians dont really care about a random nophono theorem in R^27, they care about the mental challenge the problem brings with it.
â„–3948016[Quote]
>>3948013But I think many mathematicians havent realized this yet.
â„–3948017[Quote]
>>3947981Ah, your perspective is very decoupled, which is both justified and unjustified. Justified because you're right geg, but unjustified because 100% of fields in science, engineering and statistics require the math you do. Given field specific math and how rare true math nerds are in a population, wouldn't that make it more plausible?
â„–3948029[Quote]
>>3948017The thing is there really is no new engineering math. We would first need to find something physically new. Math is developed far beyond physics math. And also if you look at the theorems being proven they are definitely not helpful to anything irl.
Most applied math is already automized in the qay you mean it. These guys had AIs like AlphaFold like 15 years ago, indtead of building general models they have built task specific AIs which we mostly know how to do.
â„–3948035[Quote]
>>3948000There were luddites in the XIXtg century that opposed industrialisation and they failed so I wouldn't be surprised if such movements would fail once more
Climate change won't ruin us as much as we will ruin ourselves by an economic strangulation at least in the West
As per ASI like I said it's really hard to say with how secretive companies are and how much we don't really know about intelligence
I think the future may be in biological computers or lab grown brains with super-intelligence even if such notions are dystopian and horrendous
â„–3948042[Quote]
I generally view biological technology as the mist advanced one since it allows you to do very wild stuff if you think about it, even an organism so unsophisticated like yeast can without much external energy turn sugar into alcohol
â„–3948044[Quote]
>>3948035I am not against using AI in the fields where it actually matters. But what I have seen so far we have been told a lie. Dario said 2 years ago that 90% of developers will be unemployed in six months. And now many companies are rehiring more than before.
Also as I have stated I dont see a reason to automate things which inherently dont make sense to automate. Cancer research makes sense, but have we seen any by these large AI companies? No! All we got was some half baked uninteresting math results.
And now it turns out even the AI safety outbreaks were fake, because the israelis dont know how to program an enclosed system.
I remain skeptical.
â„–3948046[Quote]
>>3948029So you'd exclude, again maybe I'm misunderstanding, quantum mechanics applied engineering math, like quantum computers and other "new science" math, electrostatic slop space engine, and such? But that wouldnt specifically classify as math, it would be what youre saying? Also, that makes sense, classical mechanics is pretty damn well known.
â„–3948049[Quote]
>>3948044Yeah these AI outbreaks I to reckoned were fake
â„–3948051[Quote]
>>3948042I think the future of AI will lie with biological computing. But currently everyone is in the llm phase.
They put brains on graphic chips and connect them and now you have a brain with AI intelligence.
â„–3948056[Quote]
>>3948046What I meant to say is we have all the math one would ever need to do science. What it lacks is new science. The last big science breakthrough was like in the 1980s with the standard model.
Quantum computing also just lacks behind not because of math but because they cant figure out the physical part.
â„–3948058[Quote]
>>3948049>>3948046Sadly I think AI will mainly be used for war and surveillance and nothing really gemmy.
â„–3948062[Quote]
>>3948051The biggest problem with organoids and artificial tissue is sanitation
One of the most sophisticated, poorly-understood and pivotal elements of life is immune systems because without them you require sterile environments that are very expensive and tedious to maintain so I think biological computers will require marrow-like and other structures that enable white-cell production
â„–3948064[Quote]
>>3948058That's unfortunately the case with everything
The good uses is only limited to the wealthy and the elite
â„–3948065[Quote]
>>3948062Probably but they have already put small brains on spider robots and let them freeroam the environment.
â„–3948068[Quote]
>>3948056Hm, fascinating. I need to understand math more, I'm very lacking.
>>3948058I agree with this sad sentiment.
â„–3948072[Quote]
>>3948065My favourite tit-bit of information about growing these structures is that neurons have innate abilities to form complex structures even when they grow from stem cells
â„–3948074[Quote]
>>3948065Can you release a new track like Vollig Losgelost for my warthunder luftwaffe Playlist? It's been like 30 years since you dropped a new single
â„–3948075[Quote]
>>3948072Yeah it is interesting. It was also discussed that these things might be concious, even as tiny as they are now.