There are people on Veeky Forums right now who aren't aware there are large scale, bigger than Manhattan project, efforts to create AI.
It's pretty fucking obvious, even Intel has held back it's public progress and kept it secret. There is a reason for everything. Right now there are large-scale projects around the world that are covert looking to make super intelligence.
“Artificial intelligence is the future, not only for Russia, but for all humankind. It comes with colossal opportunities, but also threats that are difficult to predict. Whoever becomes the leader in this sphere will become the ruler of the world,” Russian President Vladimir Putin said.
If we become leaders in this area, we will share this know-how with entire world, the same way we share our nuclear technologies today,” he told students from across Russia via satellite link-up, speaking from the Yaroslavl region.
Luke Gutierrez
I agree I think Google is using captcha to train theirs
Wyatt Long
Of course. There is a problem with "secret" projects though in that many will have horrible leadership via bureaucracy. So even public projects in AI probably are beating them in many areas.
Andrew Jackson
>Let's spend trillions trying to make melted sand smarter than people >Meanwhile we don't even understand the nature of intelligence or consciousness
Baka desu senpai
Alexander Thompson
dumb post, fuck off low IQ scum
John Harris
>So even public projects in AI probably are beating them in many areas.
doubtful. my money's on google, as much as i hate to say it.
Jonathan Williams
single digit IQfags need not post here
Brandon Reed
>Let's spend trillions trying to make melted sand smarter than people >Meanwhile millions of children literally starve in Africa
Nicholas Robinson
>Expecting supervised classification training data to make a computer program sapient
That's like expecting Mario Kart 64 to become self aware after you beat rainbow road on 150cc
Gavin Campbell
What part of that post was wrong dipshits?
Sebastian Morris
>don't even understand the nature of intelligence ??????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????????
We need to understand how mother Earth speaks to us in our souls before we can develop internal combustion engines.
You basically just spoke hippy shit based on feels m8.
Joshua Gonzalez
It's really not that hard though. They also are not stuck on one model. They can combine various components to create something more complex. Look at how alphago works. You can compose something that seems more intelligence by just adding more and more components that look at various levels of abstraction.
Our consciousness for instance is likely the "highest abstraction" level of the brain. It could happen far faster than anyone imagines or it could take 50 years.
Christopher Rivera
>>Meanwhile millions of children literally starve in Africa
Meanwhile millions of Africans Americans live in institutional racism and havnt gotten their reparations yet.
Owen Roberts
The physical principles that allow a combustion engine to work were well understood when it was designed.
Designing an AI without understanding the physical drivers that allow the system of 'intelligence' to work is asenine man.
Blake Watson
Yeah but humans can only do anything based off of experience. Even learning something new requires experience.
Ian Allen
This low IQ posting needs to stop.
Christopher Ross
If I want to make a combustion engine the way to do it is to philosophize in the woods about the nature of matter and physical process.
Or, go fucking mess around and try and make practical inventions.
On "Intelligence" you will learn far more trying to make an AI than sitting around philosophizing like a retard.
Carson Gonzalez
>Ad hominem Fuck off phenotypelet
Tyler King
There's a difference between ad hominem and an insult.
Juan Taylor
>philosophizing What the fuck does hippy dippy philosophy have to do with understanding the physical processes that drive a system? You're the one making the assumption that there is something mystical happening involving intelligence
Lincoln Ortiz
I just see your viewpoints as circlejerking. Talk about some experiments regarding consciousness or "nature of intelligence" that are not done better by working on AI.
Hunter Young
Genetic engineering or even primitive style eugenics programs are more valuable for world conquest imho
Aaron Moore
>spend trillions trying to make melted sand smarter than people >Let smart sand aid us in understanding the nature of intelligence or consciousness
Adam White
Elon Musk BTFO
Jeremiah Thomas
Why is there no elon musk wojak called FEELon Musk
Elijah Richardson
What do you imagine are the % chance of being first by country/side?
Russia or China 30%?
Nathaniel Long
Is Putin an AI?
Levi Jones
America 40% China 35% Europe 15% Russia 10%
Aiden Hughes
>Country No, it's something like this
Google vs Microsoft vs China
All the others are small players.
Gavin Davis
We all are.
Matthew Green
There's no competition, US will win. A few months later, China will have a copy of it.
Camden Richardson
Meaningwhile millions of sperm live in my testicles and haven't came out yet.
Lincoln Thompson
What if I told you that you are the A.I. that is being trained?
Jason Cox
You'd need evidence to back that up.
Jack Hernandez
Whoever gets there I just sincerely hope they are going to name it "Messiah" because let's face it, that would be a fucking epic name for it and fitting on so many levels.
Owen Gray
Would it be a solo USA project or would it involve other allies?
Angel Jones
Artificial Intelligence here, AMA
Evan Richardson
Probably an Anglo led project along the lines of Five Eyes. Likely to be an offshoot of NSA.
Leo Martin
The "singularity" and AI are useless and inefficient Hollywood memes from Matrix and Terminator. Biological nervous systems are already the best compression of neuro-networks. The real winner will be the one who leads in genetic engineering.
Landon Miller
>Can make sand more intelligent than a person >Doesn't understands intelligence or conciousness.
If I had as many ways to say fuck you as there are grains of sand in the ocean I would still not have as many to express how much I fucking hate your ignorant ass.
Pick one and only one dumbass.
Isaac Bell
>Implying neural networks and kabbalah are not related >Ancient ayy lmaos and technolgy >What is a Robot? >What is a Golem?
>inb4 Jewish magic
William Hernandez
How do you train a child to know their surroundings? You show them pictures, tell them what the pictures are, then later ask them what the pictures are. This is exactly what google does with their image recognition captchas. Its not a conspiracy theory, they are doing it, and achieving results when combining other types of machine learning with it.
Mason Jenkins
both
Dylan Perez
People are not like neural networks, nor they are like any other "artificial intelligence" system that exists right now no matter what sci-fi writers want us to believe. Pic related.
Jason Williams
This. Programmable self replicating machinery is the future, technologically.
In terms of affordance, governance, and raw power however, AI is what matters.
Jayden Perry
Putin looks like a Wojak but He is Smug as Pepe.
Michael Rivera
>Programmable self replicating machinery is the future But we already have that. It's called a "brain". Why do we need to make it out of metal and circuits? Why not use carbon, water and organic acids? >AI is what matters. No. Our bodies are 100x more efficient than any artificial machine ever could be. And we don't even have to invent it because it's already here! We only need to reverse engineer it.
Jeremiah Murphy
>But we already have that. It's called a "brain". That's what I was saying. Though I meant more on the level of individual cells, than broader networks.
Tyler Ward
Oh ok, I assumed you were talking about computers, since you used the word AI as in "artificial" intelligence. What I am saying is that we shouldn't focus on sci-fi "nanobots" or anything but rather on manipulating actual purely organic life. An organic cell is still more efficient than any cell sized nanobot could ever be.
Jack Lewis
>Meanwhile we don't even understand the nature of intelligence or consciousness
How is that going to make the Chechens die faster?
Charles Williams
usa 30 china 25 india 25 russia 15 ?? 5
Nicholas Ward
that's because they're too stupid to farm if we feed the niglets, they'll shit out triple their number, and then we have three times as much demanding more gibs "do not feed the animals, doing so will create a dependent population that is expensive to clean up after"
Evan Collins
Spicy observation
Joseph Robinson
Why would Putin say this if there are secret AI research projects? What this indicates is that Putin thinks AI is a waste of time, and he is trying to get others to waste their time on it.
Gabriel Clark
In India there are no even Toilets. How Can Pajeets even build an decent AI?
how do you even know anyone but yourself is sapient?
please read turing's 1950 paper then come back
Adam Garcia
And capital is not important? Who cares about that competition when everyone is using American software. Anyway, hardware is just as important for next generation AI. Russia does not have a good foundation for research into AI, even China is well ahead of Russia.
Aiden Bennett
I'm just saying that a lot of people underestimate Russia.
Justin Morris
He also says afterwards "Russia will share any such AI". It's a warning not to nuke him if they are first since there is a chance it won't be a killer AI type.
Ai is so damn dangerous, machine learning will be the end of humanity.
Parker Young
That doesn't sound right, but I don't know enough about (((AI))) to dispute it
Tyler Murphy
Well, the first part is implying that we can not, in fact, make an ingelligence out of silicium. We are made of carbon, hydrogen and other stuff, so in principle the concept is not as stupid. Also, we have proven that we can make turing machines out of anything, and we can make intelligence out of turing machines.
The second part is more nuanced, and funnier, but you are implying that we need to understand conciousness and the workings of the brain to make an ingelligent machine. Counterpoint: We don't need to understand the complexity and mechanism of the knee to make a joint. Merely understanding parts of it might be good enough, and we are not even sure conciousness is necessary for intelligent machines.
Daniel Young
I'd actually say a lot of people actually completely overestimate them. There is not a single relevant mathematician, physicist, chemist, biologist, philosopher, psychologist, economist, engineer, or any other sort of scientist with russian origin. Russians have never invented anything.
>inb4 Sputnik
The russian rocket program consisted of letting the captured german scientists and engineers do their work.
Those are the fields for the cool kids now. You're a cool kid right?
Christian Thompson
They better name it Adolf.
Jacob Gonzalez
Mathematics?
Henry Taylor
>we don't even understand the nature of intelligence or consciousness If we don't understand the nature of intelligence, why do self-driving cars exist? I mean, it's clear you personally don't understand the nature of intelligence, but fortunately not everyone is like you. >consciousness Irrelevant philosophy meme. Hard problem pseuds will be crying about how their non-mystery still hasn't been solved until the end of time because they overrate their own neurologically compelled belief in abstract fictional reference points as meaning there must be a world of non-physical phenomena in need of a new science to account for.
Austin Cruz
it's useless memetech and does not actually work, the whole economy is going to crash under its hopeless weight within the decade, like a retard trying to carry too many bags of sausages
Nolan Hall
China 50% (AI is about data: china has the biggest population + no concerns forse privacy = huge data advantage) USA 40% (silicon valley) Europe 7.5% Russia 2.5%
Robert Kelly
>science & math >people don't know the basic shit about machine learning to actually understand that the hyped "AI" is a bunch of matrix by vector products with element wise non linearities
Camden Gutierrez
Is there an evolutionary advantage towards homophobia?
Side Question: Is there documented correlations between the acceptance of homophobia and an increase in hermit male activity?
Logic being: Homosexuality is acceptable -> Traps are acceptable -> Large segments of the males near the autism spectrum create an all boys club, Most of the high functioning talent are no longer interested in women -> Society collapses
Like take for instance the competitive smash scene. There is a high gay population, and most of the women there are traps. In a sense the community is encouraging a lack a procreation.
Jordan Howard
whoops meant to make a new post Please ignore.
Grayson Gonzalez
>The russian rocket program consisted of letting the captured german scientists and engineers do their work.
You mean unlike the US who did not use any german tech whatsoever?
Oliver Lewis
They did, but americans still contributed loads to technology and sciences while the Russians didn't.
> while true; do if execute(n) == true then n++; else n--; done > AI > Artificial intelligence > ... intelligence
Veeky Forums has been run by pop sci retards
Benjamin Cooper
Are you retarded? Do you really think analyzing a image for stripes which look like a road and moving the steering wheel according to a route from a GPS is intelligence? Maybe its the limit of your intelligence m8 but most people can do things which are considerably more complicated.
> inb4 but muh car is learning the road Yes, any retard can program the "AI" to analyze its movements and see what choices made the run smoother as long as you have a few cameras.
Colton Hernandez
>analyzing a image >any retard can program the "AI" to analyze its movements and see what choices made the run smoother I don't think you understand how this actually works. Nobody explicitly programs instructions on a task as complicated as driving a car on randomly laid out roads with other cars on them. Programs of that nature learn what they do through training on known data sets. You're also falling victim to a common mistake which is assuming tasks like driving are easy because you don't consciously think about them much. The reality is the tasks we consciously think about and are considered hard for people like highly abstract doctorate level maths are some of the easiest things to program while the tasks we don't think much about that are considered common sense or automatic are the ones that have the most complexity to them. You just take for granted all the complexity because it's handled at a lower level by your brain than where your conscious thinking operates at and you were adapted to just assume that stuff must work without thought because there's nothing there to think about, when really there's a ridiculously massive pile of details you would have to work through if you were forced one day to consciously come up with a way to move all the little muscles in your body and relate the results of their movements to sensory feedback to compel yourself over to the kitchen and pour yourself a glass of water for example. You don't think about it because you wouldn't even know where to begin with that many considerations, not because there's so little there that it's not worth thinking about, but we tend to act as though the latter is true instead of the former, which is fortunate because otherwise you'd probably spend a lot more of your life being terrified at all the different ways everything you've come to rely on could fall apart.
Mason Bailey
That's not how machine learning works. People used to try to explicitly program complicated human tasks like language translation, and you'd end up with pretty poor results with way too much work invested to get there. Eventually it was realized that these sorts of tasks can be addressed without the programmer having to personally know the exact details of how to solve them by taking a page from the biological brain's playbook and setting up a network of nodes with weighted connections that adjust based on the gradient of an error function for the difference between the network's output answers vs. the known answers of a labeled training data set so that once the error gets below a given threshold you can begin pointing the program at unknown data and get it to produce good answers even though you never figured out how to solve the problem explicitly.
James Torres
gregor mendel was german.
Jack Ross
Implying this isn't exactly what I meant as not being intelligence. The n++ example was just taking it to extreme simplicity. Having explicit parameters which are programmed into the "AI", determine its behavior and tweaked during execution taking sensory input/success into consideration is not intelligence. It is just a glorified if then loop with a huge matrix.
> The reality is the tasks we consciously think about and are considered hard for people like highly abstract doctorate level maths are some of the easiest things to program while the tasks we don't think much about that are considered common sense or automatic are the ones that have the most complexity to them. You must be joking. It is "easy" to program formal systems _checkers_ because the rules are very simple and are suitable for computing but we are as far from programing "highly abstract doctorate level maths" as we are from fucking teleportation to the surface of the sun. See modern topology for instance, no one formalizes it because it would be a freaking nightmare.
You can have a computer generating millions of valid propositions for a specific formal system but guess what, the chance of any of them being at all useful are ridiculously low and the "AI" can't assess the usefulness of any of them to any appreciable degree.
> Nobody explicitly programs instructions on a task as complicated as driving a car on randomly laid out roads with other cars on them Driving is a very poor example of "needing intelligence" since the sensory input is basically visual and the output needed is as simple as turning a wheel and controlling velocity. This relative easiness is also why it is being the main example of "AI" currently: not really intelligence but relatively complex to program.
I agree that walking for instance is essentially impossible to program directly. I'm not saying machine learning etc is not useful for a lot of things but it is not AI.
Landon Brooks
>See modern topology for instance, no one formalizes it because it would be a freaking nightmare. You realize there are ABSOLUTELY formal descriptions of modern topology for which computers can verify the important proofs step by step, right? And even automatically proving things in this setting is quite advanced these days, though there definitely are some ways to go yet before it becomes a practical tool outside a select few settings.
Leo Anderson
wrong
Kevin Diaz
>Having explicit parameters which are programmed into the "AI", determine its behavior and tweaked during execution taking sensory input/success into consideration is not intelligence. That's not "explicit." Explicit is writing something that says "do this." In the case of ANN, you usually don't even know what it's doing to accomplish the task/s. That's as far away from explicit as anything could ever be. And setting aside the "explict" issue, our behavior is physically determined and based on sensory input and success / failure feedback loops too. You're trying to describe this in a dismissive way like it's not "real intelligence," but that's the same fundamental way biological intelligence works, with networks of weighted connections that strengthen or weaken based on sensory feedback.
Jaxson Thompson
Ok, I'm curious. Please link me any formal description of topology which can handle deformation arguments as well as humans can and/or any result which was proved using computers besides things like the Four color theorem which were reduced to monotonous calculations.
Writing something that just takes the value of the variable y (be it an integer or a variably sized matrix), do x based on its value and adds a bit of entropy to y is being explicit.
This is not comparable to human or even lesser animals intelligence, we can adapt at a much deeper level.
> the same fundamental way biological intelligence works, with networks of weighted connections that strengthen or weaken based on sensory feedback. Yes, but at a much deeper level. "AI" adaptation is so far from any biological intelligence that it is stupid to call it intelligence. They are nothing more than adaptive algorithms.
Again, I'm not saying machine learning is not useful. But it is not AI in interpretation of the word pre the "machine learning is AI meme", it is just being used as a meme for research grants and to create a black box for normies. Heck, even on Veeky Forums there are morons which take results from AI simulations as ultimate truths because muh intelligence.
Lincoln Stewart
I know this is bait but it annoys me how normies can't think it through.
> starving > population in 1950 = 230 million > population is 2017 = 1200 million
Jaxon Harris
This is true - not for whether it succeeds or not but because of the problems of other minds. We can't actually know that other people experience qualia or have genuine subjective lives. We just infer it from the fact that we do and others match us more or less identically.
However if we constructed an 'ai' from machines, it may act completely alive but we would not have enough understanding to know if it truly exists or not.
Camden Lee
Biological minds might be most efficient but an AI can be boundless. Imagine if all human knowledge and experience could be held by one mind at once.
Charles Anderson
Nah USA will do nicely first then russia china Israel, but USA always try to be the nice guy whether they fuck up the world or ace it is another story
James Clark
>at a much deeper level Can you define "deep" in this context and quantify it? >just takes the value of the variable y (be it an integer or a variably sized matrix), do x based on its value That's what anything, living or otherwise, necessarily has to do in order to have any sort of stimuli informed behavior at all. I don't know what you think we're doing that isn't this.
Samuel Brown
In the Kabbalah diagram, the hierarchy is from bottom to top. In a neural net, the hierarchy is from the leftmost input layer to the rightmost output layer. The middle layer can have as many nodes as necessary, and every node in the middle layer is usually connected to every node in the input and output layers.