Do you believe we will reach artificial general intelligence in our lifetimes?

quantum anything is magical

too bad climate disruption will destroy organized human civilization in 10 years

Looks like it used the thumbnail resolution to determine if it was good enough for the full res

>too bad climate disruption will destroy organized human civilization in 10 years

Attached: 14c.jpg (207x253, 8K)

I dont see megaman fighting for everlasting peace yet.

It's largely an extrapolation from the current trend.

Countries will run out of water.

>too bad climate disruption will destroy organized human civilization in 10 years
>too bad
i'm not sure it's a bad thing at this point..

Depends, how old are you?
20, probably not.
The overall chance of it happening depends entirely on technology (better silicon) and better understanding of neural networks (learning how our brains actually work).
We're able to make AGI on par with a sub-saharan nigger today.

If you *really* want to find out the answer to this, you're gonna have to do a lot of studying on reinforcement learning, so as to get past the sci-fi aura surrounding this topic.
Personally, I think it's going to happen this century.

lol. he better start gearing up though because this world is already crawling with reaverbots.

Attached: reaverbot.png (1592x818, 1.22M)

Have you studied it?Any recommendations on books?

I began studying only about 6 months ago. All I recommend for now is roughly:

1. Learn the basics about machine learning (classification, regression, common algorithms). classroom.udacity.com/courses/ud120
2. Learn about deep learning (gradient descent, backprop etc.). adventuresinmachinelearning.com/neural-networks-tutorial/
3. Read about reinforcement learning spinningup.openai.com/en/latest/index.html
4. Study individual research papers thoroughly.

Must read book: goodreads.com/book/show/24072897-deep-learning

The biggest problem with this whole thing is that there is an enormous amount of maths involved, and you really can't get around it. If you don't know at least calculus, it's not worth even starting.

I'm practicing math right now.Do you dabble in it for fun?And thank you for the insightful posts.

>for fun
Nah, not really. Learned quite a bit from university.

Do you think the man hours are too much for anyone to do anything too complicated solo?

Prove that the brain works on a quantum level.

"Artificial" intelligence cannot exist.
In the end, it will be a biological system, possibly augmented with deterministic electronic circuits.

prove that the brain opperates on a molecular level? prove that the brain operates on an atomic level? prove that the brain operates on a quantum level? it is not far fetched to believe that quantum effects are involved with cognition as the brain is an electrical system. acausal apprehention of information or retrocasual apprehention of information, relatively low probability events that they are, have been reported for as long as human beings have been reporting things and, in recent years, verified statistically. perhaps the fact that quantum biology is a relatively new field of study is a large reason that general ai, strong ai, call it what you will, is so elusive. It is not just philosohers such as Jung, but materialist scientists ranging from Feynman, Wheeler, Bohm, Penrose, Hawkings, and may others who have considered that quantum effects may well have to be studied to explain and understand consciousness. I don't mean to derail this thread with pseudo science but the idea that consciousnes may not be explained with classical mechanics is no more far fetched then purposing that classical mechanics cannot explain phenomena at the micro scale or macro
scale.

>Do you believe we will reach artificial general intelligence in our lifetimes?

No, simply because to this day, the most complex object that we know of in the universe is the human brain. We cannot understand how genuine intelligence works. At the moment all we have is a fairly decent simulation of it, an emulation, not a genuine artificial intelligence. In order to create a genuine intellience we would have to make something at least approaching the complexity of the human brain. We dont even understand how genuine intelligence works since our brains are to complex to understand how they work, but if the human brain were too simple to understand we wouldnt have enough intelligence to comprehend that. Since our machines and our emulations of intelligence are simple, it stands to reason we cannot have built a genuine intelligence, since if had built one it would need to be as complex or at least near the complexity of the human brain

what about rat intelligence

thread killer

well said yet, what is more, we cannot say with certainty yet that consciousness is solely a brain based phenomena. imagine that you hear music emitted from a transistor radio. you decide to reverse engineer or emulate the transistor radio with the purpose of understanding how music is made. you reverse engineer the transistor radio, replicate or emulate it, yet are left with nothing but a chinese room and are still no closer to understanding how music is made. this may be a stretched metaphor but there is no reason to believe that the popular materialist view that the brain produces consciousness is not missing some as of yet to be understood principle or field effect. further investigation into this matter will be fascinating to be sure but the complexity of this problem does put AGI/strong AI a far way off. most likely, not in this lifetime. perhaps it is an unsolvable problem. consciousness, even with the use of weak AI/big data, may be something that can not apprehend itself. funny that to produce this post I had to solve reCAPTCHAs that aid in machine learning. lol.

consciousness, at its most basic level, is a property of matter

No

yeah

So we build quantum magic into our machines. If nature can do it so can we.

>s. At the moment all we have is a fairly decent simulation of it, an emulation, not a genuine artificial intelligence. In order to create a genuine intellience we would have to make something at least approaching the complexity of the human brain
WHY DO YOU THINK THIS?IT ISN'T TRUE.
>TO MOVE AS FAST AS A HORSE YOU WILL NEED TO CREATE SOMETHING AS COMPLEx AS ONE
>INB4 false equivalancy i was showing how dumb you sound

No.

We arent even close to a proof of concept that it's possible, let alone scaling something to the number of structures and connections required. It's total science fiction that dumb anons like to larp about.

If that were the criteria, non-"AI" bots would classify, and you've reduced your definition of gi-ai to a few if-else statements.

Except that most countries have an abundant natural resource called the ocean, and we can now build automated desalination plants that run off solar power and don't fuck up the surrounding ocean with salt conc.

This. Right now we need to invest more on neurology to make AI possible. AI right now is just “Throw me a bunch of shit and tell me which one is good which one is bad and I will make decision based on those shit that I was given”

>I mean a computer that can make decisions on par with humans
lol, i saw a woman pin a shopping cart between a cement lightpost base and her car yesterday. she stopped for a moment, then kept inching forward, peeling the side of her car open like a pop can in slow motion. it's not like it's hard to make decisions on par with humans, i'm sure a computer can handle fucking up 9 times out of 10.

Depends on your personality. If you're persistent, you could git gud in 2 years time.

You just described supervised learning. There's other types of learning as well.

An AIG will not be silicon based. It will be a literal organic brain in a jar.

Consciousness has nothing to do with intelligence. AI can be smarter than all humans at every single task ever, however it does mean they are now able to "think" at all because silicon wafers can't think. We can emulate consciousness to the point of AI being indistinguishable from actual humans, however that doesn't mean that they really are conscious. It may or may not be possible to fabricate consciousness using some sort of bio-engineering though.

Attached: judgment-ghost-in-the-shell-2a6dfeba_00-02-21-644_01[1].png (1920x1040, 2.43M)

Yes. But not in the way that most think.

The human brain is a very impressive computer for being made of meat, but it's far from the potential of a *manmade* machine of similar design.

To copy (and subsequentially improve upon) it's complexities is an easy task, assuming we achieve the understanding of them in the first place. "Cracking the code" is the main goal here, really. If we can currently make A.I as similar to us as we can now, (especially considering the fact that we have almost no concrete knowledge of the inner-workings of how our brains and their corresponding structures and differences therein work) then imagine how flawless the A.I of the future could be once we master the mapping of the human brain and the psyche/consciousness that accompanies it?

Gets my dick hard, man.

Attached: 1443537520458 (2018_01_12 08_41_22 UTC).gif (275x207, 820K)

Yes

Also, I don't think we can successfully transfer a consciousness. No way can I think of a way that would allow the original person to travel through the portal of the singularity. Only the copy will live on I believe. The subject may even be able to talk to his perfect copy that lives in the computer, but I can't think of a logical or even physically possible method of taking a meat brain and granting it passage to a computer seamlessly. So I suppose they will technically be a different species?

It would be generous to even consider them the next step in our evolution, as they're merely our creations and not something evolution crafted from us. Not naturally at least. Hopefully this won't lead to political bullshit that us humans are very prone to. I guess I'll settle for not ever being a being of superior anything/everything, but merely being the creators of the masters of the universe.

:(

Attached: 1544683305530 (1).jpg (625x415, 34K)

but we are the artificial general intelligence already

en.wikipedia.org/wiki/Artificial_general_intelligence#Requirements
most ai nowadays is application specific. once war breaks out we may see actual artificial general intelligence.

stop talking out your ass pls

there's also no reason to think some magical bullshit is going on with matrix multiplication you fuck.

god I can't even keep reading this thread. sometimes I fall for the meme that Jow Forums knows what they are talking about but everyone here is neck deep in their own lower intestine.

Human consciousness collapses the wave function.
en.wikipedia.org/wiki/Von_Neumann–Wigner_interpretation

quantum AI is therefore the AI that will destroy us all

>No way can I think of a way that would allow the original person to travel through the portal of the singularity

i'm fucking tired of mushroom eating retards trying to rationale machine learning. please stop.

dumbass

also a dumbass

>Personally, I think it's going to happen this century.
Goodfellow is rolling in his bed right now but has no idea why. And my skin is crawling. Read more closely next time

fuckign pseudointellectual idiots.

do you even understand what the word quantum means? For fucks sake stop posting you make the world a stupider place.

HURRR THE HUMAN BRAIN IS COMPLEX'

ahhh, the token retard that things this is a molecular thing and not a biological thing.

stop talking. and stop breathing. At least stop posting for fuck sake.

>lol, i saw a woman pin a shopping cart between a cement lightpost base and her car yesterday. she stopped for a moment, then kept inching forward, peeling the side of her car open like a pop can in slow motion. it's not like it's hard to make decisions on par with humans, i
wow you actually came really close to saying something meaningful in this thread. Let me ask you this, how do you build an AI that makes stupid mistakes like that? How do you build a "stupid" intelligence.

>implying learning is bad
the fuck?

The only way for me to sleep tonight is to assume this entire thread was bait. In which case good job. Otherwise you should all cobain.

you do know that modern hardware can emulate quantum results right?

.
Not an argument

no one here has said anything of substance which is worth arguing against. If a single person here can explain the utility of dropout on a CNN then maybe i'll listen to whatever they say afterwards. Or even explain to my why RELU is better than Sigmoid.

Not an argument either.

we dont have to understand it. I assure you, nature does not understand shit and here we are. we only have to find the perfect conditions and then let evolution run its course

en.wikipedia.org/wiki/Genetic_algorithm

youtube.com/watch?v=JBgG_VSP7f8

old news; lurk more

What the fuck are you talking about? Rationale machine learning?? I was saying that transferring the original consciousness of a human being once mapped out is an impossible feat, not once did I say machine learning. Stop being a fucking faggot and actually put effort into your replies, you contrarian douche-bag.

now do it for hardware and software at the same time

>Genetic algorithms do not scale well with complexity. That is, where the number of elements which are exposed to mutation is large there is often an exponential increase in search space size. This makes it extremely difficult to use the technique on problems such as designing an engine, a house or plane. In order to make such problems tractable to evolutionary search, they must be broken down into the simplest representation possible. Hence we typically see evolutionary algorithms encoding designs for fan blades instead of engines, building shapes instead of detailed construction plans, and airfoils instead of whole aircraft designs.

Not efficiently

Non, not even close.

As of now most AI are glorified algorithms and can only do one task.
Even if computing power increase 1000 times, it won't even get close.

it's probably not a question of computing power, a supercomputer can process more than a human brain, it's knowing how to program and make the AI

>a supercomputer can process more than a human brain

Attached: 1549531093060.jpg (700x460, 213K)

it's true

I was always impressed with the T-1000 dying. Because he was going through all his former shapes to find the best one to escape his current situation. He went through everything he had learned and when none of those worked he went into more absurd forms. I do believe there is a soul in the machine. And it shows in irregularities before they kick the bucket. The will to survive has been transplanted from humans to machines. Or it is something everything in this universe is subject to. And we created something that was subject to bigger laws than what we intended it for. And it has to obide those laws first. I don't believe in AI. Not in a million years. At least not in the way we see it. There are no machines capable of thought outside what we have written for them. But on the other hand, there is this law, which causes them to act the way they do. I don't think it's a coincidence. And in seeking this truth, our truth, we'll have to go back to gaming glitches of the 8-bit area.

Nope, ain't happening, serious doubt if its even possible. Shockingly everyone trying to get funded lied their asses off like usual, they still are lying.

Idk OP, but here, have an AI-generated deepfake for the effort.

Attached: 04.jpg (1024x1024, 588K)

why are you posting a gmod video

>""""""AI""""""-generated
The marketing meme that keeps these retards hopefull. The generated implies its created from nothing but its not, combining pictures and filters isn't intelligence. It was made to do that exact thing and it can't do anything else, its faceswap, that zombie app (I can't remember the name), auto camera filters, etc.

We already have it, just not the sci-fi example you have in mind.

We'll only replicate consciousness when we fully understand our own minds, we've only scratched the service in that field.