https://www.youtube.com/watch?v=Khuj4ASldmU
Artists are now on suicide watch. The machines will prevail again.
how far is too far...
>artsy filter that works sometimes
Wow, so impressive.
>>8066031
some point after mankind's extinction that's for sure
>>8066106
atleast art will move on after were dead
>>8066108
better than it did under us (even before pomos)
>>8066115
>/sci/ autist doesn't get pomo
what a surprise!
A fucking composite of a bunch of rembrandt portraits was on the front pages of national newspapers here a while ago with headlines like "computer makes new classic painting" fucking bullshit makes me mad af
>>8066025
Th-Thanks Sci...
Now you have succesfully taken Art Majors name job...
>>8066168
Jesus christ I know ML in the media makes me so fucking angry.
>>8066186
>art majors
>jobs
>mfw all artists who worked to animate the movie below could be replaced by a script now
https://www.youtube.com/watch?v=47h6pQ6StCk
>>8066025
Artists are always on suicide watch. Nothing new.
>>8066025
This technology will simply make artists more productive.
>>8066025
>implying artists don't feed on "negative" emotions and experience to make their work
Ironically this will only make them stronger in that it now gives them a reason to make two new movements one for art automation and one against automation.
>>8066482
This. Artfag here. Figuring out how to package this thing for Archlinux. My job is about to get a lot easier.
>>8066511
>Artfag who uses Arch instead of fucking macshit
You're one of the good ones Anon. Thank you for existing.
>>8066511
>artfag
post your stuff
>>8066482
Not really -- give it enough time and it will be an admin/programmer pushing a button for a robot artist to start creating stuff better than any other human
>>8066482
True, but now we will have a few very talented, highly paid artists instead of lots of mediocre artists. Automation turns the industry into winner-takes-all.
This isn't necessarily a bad thing, but it does change the change the dynamic
>>8066520
The thing is nothing the robot can make will be original. It will be based on an algorithm. So sure you can make the algorithm produce random results but still, those random results are based on more algorithms. Humans will still have art jobs but like this guy said >>8066525 but there will be less of them for sure.
>>8066533
>those random results are based on more algorithms
Not if we use background radiation signals as an input.
>>8066331
Hello darkness my old friend...
>>8066533
>So sure you can make the algorithm produce random results but still, those random results are based on more algorithms.
So just like humans?
>>8066525
>>True, but now we will have a few very talented, highly paid artists instead of lots of mediocre artists.
Doesn't really follow. It will make art more accessible to more people. Potentially talented people who didn't have the time or patience to master the technical skills will play ball as well, so we'll have even more talented artists. Technical skill is just a hindrance that keeps unskilled creative people back
>>8066511
>artfag
>Arch
What the fuck am I reading?
>>8066533
neural networks are programmed identically to neurological processes, so if you are saying they are based solely on algorithms then you are saying a human artist's creations are as well
>>8066598
>neural networks are programmed identically to neurological processes
rofl
>>8066601
nice rebuttal
>>8066603
not him but you're completely retarded. neural nets are FAR FAR FAR FUCKING FAR away from being IDENTICAL to their biological counterparts. they're an incredibly huge simplification. tard no one even knows how the biological counterparts are programmed
>>8066603
it's not worth it
read a book you retard
>>8066607
the only difference is the setup of the network, neuron by neuron they are identical
>>8066607
in what way are they simplified?
>>8066612
read a book
>>8066572
I bet you believe humans don't have free will. Please leave.
>>8066616
>(you)
>>8066025
>applying a filter
>""""""""""""""""""""""""""""""""art"""""""""""""""""""""""""""""""
I MAD, fucking media spreading misinfo again, holy shit how can they be wrong every single time?
>>8066618
> you're obviously more ignorant than I am
except im doing my master's in machine learning with a focus on neural computing lel. read hodgkin and huxley (1952)
>>8066025
Machines will always be great at copying. Finding one that "creates", now that's the real challenge
>>8066630
>butthurt artfag sperging out
>>8066639
every fucking nerd on the internet and their mother is doing a phd in machine learning these days, particularly when they're trying to win internet points
>>8066639
can't you at least give me a summary of the differences? it doesn't seem that paper proves anything other than a more direct underlying explanation, but it still seems to act the same as a single programmed neuron
>>8066331
kek'd
one which board are we again?
>>8066679
point is that hodgkin and huxley, izhikevich, and other models attempt to create a closer copy to the biological neural nets (not identical, because we don't understand them enough to be able to model them identically, so that's out of the question), but they've been largely unsuccessful for practical applications. on the other hand, simplified versions that don't take into account spike rate/timing, ion channels, and other biological details perform a lot better. ibm tried to push their hardware based on spiking neural nets (north something) and got 0 attention because they don't work. saying that neural nets "are programmed identically to neurological processes" is ridiculous
>>8066696
well I guess that's a little closer to what I was looking for, thanks
In terms of my original argument, programmed identically or not, the workings of the brain can be described "algorithmically" so to say art will never be created by an ANN is almost definitely not true
>>8066696
What's a good book to understand
a) how this roughly works
b) (possibly another book) what works and where it can be used
?
>>8066108
Until some faggot ayy lmao hipster comes along and declares the ruins of our civilization his readymade art piece.
>>8066704
haykin (2009) is often recommended as an introductory book (my undergrad intro to neural nets was based on this and beale and jackson (1990)) and for deep nets goodfellow et al released a book this year, but you can also try the online course provided by karpathy and fei fei cs231n (though it was recently removed but there are torrents floating around see https://www.reddit.com/r/MachineLearning/comments/4hqwza/andrej_karpathy_forced_to_take_down_stanford/ )
anyway gotta go back to studying
>>8066718
thx.
Found deep learning online, and also it's readable online here
http://www.deeplearningbook.org/
PS I'd argue the guy has a narrowly focused intelligence if he puts up unattractive pics of his on each outlet I encounter. People are shallow, you wanna get somewhere so put some minimal effort into it
>>8066805
>narrowly focused intelligence
Is this an actual thing?
>>8067023
Yeah, it's a defense mechanisms used to discredit more intelligent people
