Intelligence will be created with biology, not with computer science

There is really no reason to think otherwise, as we aren't there yet.

>>8453420

That's retarded. There is nothing special about meat that makes it a better substrate for AI.

>>8453420

im sure even if that were to happen it would still rely heavily on computer science and math. neuroscience is becoming increasingly involved with mathematical modelling and machine learning and stuff like that

>>8453429
However, it isn't computer science that will make the softwares, scientist are learning more and more computer science

>>8453446

- why does it matter the job title of the person who does it, its still maths, its still computer science, its still math.

>>8453475
Oh yeah nevermind that post

File: burningTerminatorHead[1].jpg (39KB, 452x318px) Image search: [Google]

39KB, 452x318px

But muh Skynet fantasies :c

when do you plan on biologists artificially spawning sentient creatures, given that they can't even replicate single cell organisms?

>>8453504
This has to be bait

>>8453507
I'm sorry, has an artificial lifeform been created that I'm not aware of?

>>8453541

its probably much harder to create intelligence by growing an organism since im sure we are far less competent at making biological than mechanical hardware.

>>8453549
oh, I thought OP was talking about creating an entirely new creature, is he just talking about genetic modification?

>>8453563
youve confused me but i dont know. i assumed new creatures.

File: 1478134952719.png (169KB, 600x700px) Image search: [Google]

169KB, 600x700px

>>8453563
>>8453549
Genetic modification if that comes first yea.


It could be similar for both areas :

For computer science,
Intelligence comes first then robots

For biology,
Intelligence comes first then creatures
My prediction is a combination of both like >>8453429 said

>>8453446
> It isn't computer science that will make the softwares
> scientist are learning more and more computer science

>>8453712
you what?

>>8453541
I can literally make one with my dick
>but that's not artificial
It was made my man how isn't it artificial

>>8453426
Is there something special about computers that make them better for AI?

>>8453733

when i mean artificial i mean one which has a physiological structure which is purposively designed by a man, not one developed over millions if not billions of years of evolution

>>8453748

have you ever seen a man-made biological computer?

we are far behind in terms of engineering biological hardware.

>>8453759
but we don't have to do it from scratch, just copy what's already there from humans or something

there's a biological computer in all of us

>>8453748

Yes, they're much easier to build and control with formal languages.

>>8453767

Copying a biological entity isn't contributing to AI any more than having a child the normal way would be.

>>8453771
well you could copy and then modify obviously

>>8453420
phd in Neuro(((science))) is literally worth as much as phd. in Woman's Studies

>>8453774

Sampling isn't songwriting.

>>8453790
okay say you sample an ai

it creates an ai

you just songwrote indirectly

>>8453793

No, the AI songwrote. You didn't contribute shit, you just copied something that already existed and that copy invented something itself.

>>8453774
i dont see how you could copy anything other than a human and if you were to modify it effectively, youd probably need to understand it enough that you could make a simulation of it on a computer anyway.

>>8453787

if youre going to do research in neuroscience which is where this is coming from... then a phd in neuroscience surely is not worthless?

>>8453798
you could just randomly poke at things and hope it changes?

>>8453796
the "you" was "biology"
so: biology created the 2nd ai

>>8453798
>youd probably need to understand it enough that you could make a simulation of it on a computer anyway.

Not everything we understand can be simulated. If we could, we would be simulating a child with autism and an autistic consciousness would exist inside a computer

>>8453807

You can only take as much credit in that case as you would take credit for the work of a child you had the normal way. Jimi Hendrix's dad isn't someone I'd credit with being a great guitarist.

>>8453834

yes and we are talking about something that hasnt happened yet in this conversation too. but im saying if we could just copy and modify organisms to be super smart, we would have to have a very high understanding to do it right. probably easier just to do ai stuff the non-biological way.

>>8453834

and just because you can simulate autism inside a computer, doesnt mean it will have consciousness.

>>8453801
>surely is not worthless?
if neuroscientists were to reverse engineer even a simple adder not to mention branch predictor they would come to into conclusion that
>geewheez it surely work using electricity
>number of "zeros" depends on the power draw
>"ALU" uses unknown system to store data in the registers
>invent some bullshit theory about clock signal and reset lines
>clock gating, f/V scaling would be considered "black magic"

>>8453859

wtf are you talking about

Everyone in academia and ai research is too afraid to admit that statistics is the most important major if we want to acheive singularity. It will be stats majors or math/stats double majors that will solve all the relevant problems in the world. After all, nothing in the universe is discrete, it is all distributions of probability.

I personally wish i did more cs electives in my time at university, ive jumped into stats post grad and its a fucking grind trying to go back and start from the start for a lot of the software because i dont have the background.

>>8453860
this is what i'm exactly talking about you're just fucking glorified matlab technicians pretending to do meaningful work

>>8453864

>Everyone in academia and ai research is too afraid to admit that statistics is the most important major if we want to acheive singularity.

Nobody's afraid to admit statistical approaches work much better than rules based approaches for AI. All machine learning is done through probabilistic methods.

>>8453860
oh and i'm >>8453876 fucking disgusted with the cons like the one of you

>>8453864

very true. And the more precise or advanced a field gets, the more math it will use. already happening for the most promising framework in neuroscience of which some big contributors are math and computerscientists

>>8453864
bet you'll have a fit when your job is depreciated by AI

>>8453876

i dont know what youre talking about. obviously a neuroscientist isnt a computer scientist if thats what youre talking about.

>>8453426
The brain is the most sophisticated computer that we know of, and we have no idea of the details of how it works and how it leads to consciousness.

File: 1449335284045.jpg (167KB, 500x385px) Image search: [Google]

167KB, 500x385px

>>8453834
>>8453843
>>8453850
Kek why autism

>>8453928

>we have no idea of the details of how it works

That's not true. There are a lot of good ideas of how it does the different specific things it does.

>consciousness

Meme word. Doesn't mean anything.

>>8453928
I think his point is that we're more likely to succeed with computers than with organic matter

File: wolves2.jpg (132KB, 800x600px) Image search: [Google]

132KB, 800x600px

Could you imagine making, say, a wolf very intelligent? What would it look like?

File: 1454567127784.gif (2MB, 388x218px) Image search: [Google]

2MB, 388x218px

>>8453942
No he meant an intelligent robot would be better than a living being.

Like, it could go into space.


However a geneticly engineered living being could breath under water and be more likable
I like this thread

>>8453939
Of course we know a few things, but we are missing the key pieces. Otherwise this wouldn't be a thing: https://en.wikipedia.org/wiki/BRAIN_Initiative

>>8453942
Organic matter is able to compute with very limited power and do better than our current supercomputers. Even on an algorithmic perspective, there must be some tricks that the brain is doing.

>>8453947
It wouldn't be much different at all. Sorry 'bout thuh bummer eh. They can't really vocalize human speech, and don't really have the means to manipulate objects precisely at all. So there isn't a way for us humans to witness their intelligence in a credible way.

>>8453959

>Organic matter is able to compute with very limited power and do better than our current supercomputers.

Do better at what specifically?

>>8453963
I think they could adapt. They can still manipulate objects to an extent, it would be interesting to see creation coming from them, especially if they are more intelligent.

>>8453963
>Sorry 'bout thuh bummer eh

What did he mean by this?

>>8453947
do they have enough different muscles in their mouth and tongue to talk like a person?

>>8453964
Logical deductions and inference. Pattern recognition. Learning from one's environment. What we commonly define as intelligence. We still don't have hard AI. While if you give a human some food for a few years, you have it.

>>8453975

>Logical deductions and inference. Pattern recognition. Learning from one's environment. What we commonly define as intelligence.

I don't think any of that is something you can objectively test a machine against a person for. You would need a more specific task to make a fair claim about comparative ability between the two.

>>8453973
It doesn't matter for communication. Crows and dolphins have complex ways of communicating for exemple

>>8453959
>>8453975
yes, brains work much better than our current computers, great logical deduction. This doesn't mean shit about which method of science is more likely to work, it simply means that mother nature/god/whatever is much better at programming AI than we are up to this point

>>8453959
>with very limited power
Electricity/power is not an issue for organic matter. A sufficient input of a list of molecules of the right concentrations is.

>>8453977
Whatever dude. Just have a computer reproduce the behavior of a human. It's not happening.

>>8453983
It'd be arrogant to think we have nothing to learn from those biological processes. Sure we can continue pursuing statistics based ML, but it's clearly still unable to reproduce the efficiency of a human brain. Understanding the brain would at least make us able to reproduce it.

Arguably, the full understanding of how the brain works remains the biggest unsolved mystery.

>>8453994
You need energy for any kind of computation. How do you think those concentrations are achieved?

>>8454017

>Just have a computer reproduce the behavior of a human.

First define in clear terms what behavior you're testing. Computers can do many things people do, but you have to be specific or else you will never have a fair benchmark for assessing performance.

>>8454022
I'm not here to argue over semantics in a 4chan board. Strong/weak AI are concepts that you can look up yourself.

>>8454017
I think it's almost certain that AI will be successfully created by mimicking the human mind, but it's still going to be in digital form, organic AI is yet to even become a field of study

>>8454032

It's not semantics. You're setting yourself up to never be satisfied by any AI because you're not even beginning to think about what it is you're trying to see reproduced.

It's much more difficult to reverse engineer it. That much is clear. We already have intelligence. In us.

>>8454035
I think so too. I'm not OP. I still think that neuroscience has a better chance of achieving AI than CS with their statistical methods. Thinking that those are fundamental seems misguided.

>>8454042
>62 posts in
>I'm not OP
all right

>>8453420
Neither of them will. Statistics will.

File: 1476969275767.jpg (243KB, 1620x1079px) Image search: [Google]

243KB, 1620x1079px

>>8454053
Believe what you want he isn't.

I didn't think my thread would be so serious, I meant it to be fun. Like just people shitposting random thoughts.

What about sending intelligent cats to space to colonize mars?

>>8453928
That's a pretty idiotic statement to make. The study of the brain is one of today's scientific frontiers. We already know and are in the process of learning a lot about it.

The "failure" of AI happened because as soon as you understand the underlying process it suddenly ceases to be "intelligence" and becomes something else. When people figured out the neuro-ophtalmological process of edge detection in perceived images, it became just that -- no longer part of our "intelligence".

>endless consciousness debate

How many of you actually studied any of this?

>>8454289
>How many of you actually studied any of this?

Did you?

>>8454321
Yes.

>>8453842
you can take a more than that but who the heck cares who the credit goes to anyways?

it'll still be a biological ai

>>8454376
Ok so how cool would it be if>>8454125
>sending intelligent cats to space to colonize mars?
happens, genius?

>>8453973
animals can't talk because they have no cheeks.

>>8454871
It'd be totally not cool at all. They'd die as soon as they leave their automated ships.

>>8454042

the behaviour of neurons and populations of neurons or the whole brain wont be understandable without statistical and mathematical methods either.

>>8453420
Learning is just an algorithm. Inevitably, it will be deciphered and meaning will finally be encapsulated into, and manipulatable by the machine. The great thing about intelligence is it's a level playing field. No one is ahead of anyone else, at least not by any reasonably significant margin that couldn't be eclipsed by a sneeze. If you have a brain, you're equally capable of understanding what it is to learn.

>>8455433
>Learning is just an algorithm.
If only it was so simple

>>8455677
trapdoor function

>>8454125
Will the fur help on mars? You know, cause it's cold.

>>8455677
The universe could be thought of as just a complex algorithm. :^)

>>8456672
you look like a complex algorithm

File: it's bullshit.jpg (19KB, 480x189px) Image search: [Google]

19KB, 480x189px

>>8453420
>Biology with genetic engineering and neurology will be able to create intelligent creatures and steal the geek's spotlight
Okay, tell me when you've created something that can actually surpass what is already being done with computer science to create thinking machines, and then we'll see how well you can define it's behavior and control it. Because the current advancement with computer science is how to make machines that do what we want them to do, and billions of years of biological evolution has advanced in such a way to make the individual better at surviving and passing on its own genes, not serving.

File: MODOK.jpg (292KB, 600x458px) Image search: [Google]

292KB, 600x458px

>>8453420
Don't try it, man.

>>8458164
Wtf dude if we create intelligent living beings it is not to enslave them, it's just for the fun of it. Maybe create a new elite in society

>>8458500
>Maybe create a new elite in society
That's dumb. Why would we create a new elite with a genome we made up (ie, not or children, like in the movie Gattaca) to rule over us? That makes us the servants who are reliant on the elites to consider us still useful to survive. The point of AI is to improve human life, not start juggling fire for the fun of it, and see what happens.

>>8458515
the fuck are you talking 'bout

the point of ai is ai and that's that

>>8459559
>the point of ai is ai and that's that
You're absolutely and unequivocally wrong.

File: ASIMO is designed to help people.png (360KB, 1242x340px) Image search: [Google]

360KB, 1242x340px

>>8459559
>>8459576
Some evidence

>>8459578
0/10

>>8459578
Kek