|
In this scenario the threat of automation - the constant media buzz that total automation of everything is right around the corner - is more important than the actual technology itself even existing or being possible. And lo and behold that's exactly what we have now
|
#
?
May 8, 2016 20:55
|
|
|
|
| # ? May 2, 2024 19:35 |
|
|
Mange Mite posted:In this scenario the threat of automation - the constant media buzz that total automation of everything is right around the corner - is more important than the actual technology itself even existing or being possible. And lo and behold that's exactly what we have now to be fair, i think the video directly addresses this using the story of the great depression and the devastation brought on by 25-30% unemployment i think the point isn't all the jobs are gone, but rather that enough are that the economy and society is majorly ~disrupted~
|
|
#
?
May 8, 2016 20:59
|
|
|
RaySmuckles posted:~disrupted~ There an app for that
|
|
#
?
May 8, 2016 21:00
|
|
|
RaySmuckles posted:to be fair, i think the video directly addresses this using the story of the great depression and the devastation brought on by 25-30% unemployment so what happens then? major uprisings? something straight out of animatrix?
|
|
#
?
May 8, 2016 21:04
|
|
|
Lastgirl posted:so what happens then? major uprisings? something straight out of animatrix? the gently caress do i look like a fortune teller ai? RaySmuckles fucked around with this message at 21:09 on May 8, 2016 |
|
#
?
May 8, 2016 21:05
|
|
|
RaySmuckles posted:the gently caress do i look like a fortune teller ai? ye but i just asked him aka u
|
|
#
?
May 8, 2016 21:08
|
|
|
Lastgirl posted:ye ask this guy 
|
|
#
?
May 8, 2016 21:09
|
|
|
AugmentedVision posted:Computers are not as smart as you think they are. A retarded kid's brain is much more powerful than any computer. The only reason computers seem smart is because their architecture is uniquely designed to do arithmetic quickly. Deep Blue did not out-think Kasparov. Deep Blue ran the numbers on every single possibility of every single move -- because chess is discrete math and computers are uniquely suited to that. Many people do not know who mario is or what certain words actually mean. The entire world is just a complex mathematical problem, Deep Blue did in fact out think Kasparov, Deep Blue out thought him by "thinking" all of the possible moves that could be played. That's exactly what a human brain does. You program the possible moves into the AI and then get it to run the simulations in it's memory, humans do the same thing just in their minds eye and thoughts. We just need more time to condition the AI's brain to act more human.
|
|
#
?
May 8, 2016 21:12
|
|
|
RaySmuckles posted:ask this guy 
|
|
#
?
May 8, 2016 21:19
|
|
|
Chess software doesn't calculate every move. It calculates a certain number of moves ahead depending on hardware limitations. Once the software has gone as many moves as it can it has to decide which game state is best. Most of the recent improvements in chess software has come from improving the estimate of which games state is better using neural networks.
|
|
#
?
May 8, 2016 21:22
|
|
|
Fried Watermelon posted:Many people do not know who mario is or what certain words actually mean. quote:The entire world is just a complex mathematical problem, Deep Blue did in fact out think Kasparov, Deep Blue out thought him by "thinking" all of the possible moves that could be played. That's exactly what a human brain does. You program the possible moves into the AI and then get it to run the simulations in it's memory, humans do the same thing just in their minds eye and thoughts. Perhaps chess computers aren't the best example, because chess is a discrete math problem with a reasonably small set of possibilities (still way too big for even the smartest human to consider in its entirety). Consider "the entire world is just a complex mathematical problem". Yes, you can encode a human's perception of the world into 0s and 1s, but what then? Deep Blue knew every possibility that could happen in its limited scope of application. Humans do not know every possibility about anything, and yet we manage just fine. That's because we are not like Deep Blue -- we are like a neural network. A neural network that is orders of magnitude more powerful than anything that is even close to being possible. E: Here's a great explanation of a Mario neural network. Note, like I said, it's "cheating" because it already knows what all the drawings on the screen mean (in a really primitive way). A human brain is born knowing absolutely nothing. Also note that it's nothing like Deep Blue -- at no point does it do anything resembling "thinking of all the possible moves", because, just like a human in everyday life, it has no idea what "all the possible moves" are. It does some things, see how they work, changes something, sees if that works, rinse repeat. That's how a human brain learns. Except a human's "neural network" is comprised of the entirety of your life experiences, so your incredibly complex computer is going to be firing off decision paths on "I recognize this is a video game, I've played video games before", "that mario is a human, I know humans", "that thing has teeth, it's probably dangerous, my decision paths will tend to stray away from it". https://www.youtube.com/watch?v=qv6UVOQ0F44 AugmentedVision fucked around with this message at 21:52 on May 8, 2016 |
|
#
?
May 8, 2016 21:26
|
|
|
How Al Bundy works at a shoe store and owns a motherfuckin house?
|
|
#
?
May 8, 2016 21:27
|
|
|
Minimalist Program posted:I work in AI and I make six figgies a year and I'm a freaking reatard hahahaha. I think it's about time for someone to PUT YOU IN THE TRASH, PAL!
|
|
#
?
May 8, 2016 21:27
|
|
|
JiveHonky posted:How Al Bundy works at a shoe store and owns a motherfuckin house? Let's see. Show started 84. His daughter was about 13. Carry the one, and, he bought it in the 1970s when people could actually afford housing.
|
|
#
?
May 8, 2016 21:33
|
|
|
Germstore posted:Let's see. Show started 84. His daughter was about 13. Carry the one, and, he bought it in the 1970s when people could actually afford housing. Yeah I guess you are right but still pretty crazy. Peg never even had a job did she?
|
|
#
?
May 8, 2016 21:36
|
|
|
JiveHonky posted:Yeah I guess you are right but still pretty crazy. Peg never even had a job did she? This was once the norm things have really slipped
|
|
#
?
May 8, 2016 21:39
|
|
|
I look forward to being replaced by an AI and going hungry and unneeded to my grave. i also look forward to seeing how long it takes to program AIs to want things and be susceptible to advertising, I'm betting not long at all
|
|
#
?
May 8, 2016 21:43
|
|
|
For devastating effect labor just needs to be massively scaled and deskilled not literally replace humans with human-equivalent robots lmao. This is already what's happening. Holy poo poo!
|
|
#
?
May 8, 2016 21:48
|
|
|
AI's are going to wipe us out, not replace our jobs. It took less than a day for Tay to start calling for ethnic cleansing. Real AI's will destroy us.
|
|
#
?
May 8, 2016 21:50
|
|
|
we don't need ai that works like our brain to replace a ton of jobs
|
|
#
?
May 8, 2016 22:00
|
|
|
Lucy Heartfilia posted:we don't need ai that works like our brain to replace a ton of jobs there's already a bot in yospos making better post than like 90% of gbs myself included.
|
|
#
?
May 8, 2016 22:10
|
|
|
Think of how hosed the system is that we think it's a bad thing that we don't need to waste human brainpower on an increasingly large number of tasks and we want to desperately cling onto them no matter how meaningless because it's the only way to make a living. Guess we're gonna need some sort of basic income or something I dunno.
|
|
#
?
May 8, 2016 22:25
|
|
|
Minimalist Program posted:there's already a bot in yospos making better post than like 90% of gbs myself included. of course it posts better than you, your just a minimalist program. once we have larger programs posting SA will become relevant again
|
|
#
?
May 8, 2016 22:35
|
|
|
Elukka posted:Think of how hosed the system is that we think it's a bad thing that we don't need to waste human brainpower on an increasingly large number of tasks and we want to desperately cling onto them no matter how meaningless because it's the only way to make a living. 
|
|
#
?
May 8, 2016 23:14
|
|
|
The super rich paying trillions to control politics by advertising to the masses to vote against their own interests won't work if 25% if people lack food/shelter/healthcare/transportation, that only works if everyone is fat and happy Food stamps and bus tokens for all
|
|
#
?
May 8, 2016 23:28
|
|
|
AugmentedVision posted:Computers are not as smart as you think they are. A retarded kid's brain is much more powerful than any computer. The only reason computers seem smart is because their architecture is uniquely designed to do arithmetic quickly. Deep Blue did not out-think Kasparov. Deep Blue ran the numbers on every single possibility of every single move -- because chess is discrete math and computers are uniquely suited to that. Deep Blue is 30 years old and is an awful reference for automation and AI. It actually computes every move and is slow as hell, look at AlphaGo which is an actual AI and neural network and how much it poo poo on the best Go player in the world. The point of automation is not to be better than the human brain at everything it is about being better than the human brain at an incredibly narrow skill and we are reaching that point now in driving, medical diagnosis, and in the next five years natural language processing. For example Nvidia made some driving AI simulator that will let an AI run thousands and thousands of hours of driving in a day. It might not be better than a human today but these things get better with more data and that is something we make at an increasing rate every day. This poo poo is inevitable it is not going to happen in a day but there is literally no way it won't come about because it's safer, cheaper, and best of all can be run 24/7 without needing to ever stop. And I bet if you put a picture of a cat and a dog into an AI designed to identify pictures like that one MSFT has it would be right 99% of the time, in a year it will be right 100% of the time and in 5 years it will be finding hidden dogs and cats that no human knew were in the picture
|
|
#
?
May 8, 2016 23:28
|
|
|
you cant make an ai this retarded
|
|
#
?
May 8, 2016 23:29
|
|
|
Lyft and GM just partnered to make self driving cars that will replace like 10% of Lyfts fleet in states with friendly laws to self driving. But that must all be bullshit and there is no way it will work.
|
|
#
?
May 8, 2016 23:30
|
|
|
Al's coming after your job? He took my wife.
|
|
#
?
May 8, 2016 23:35
|
|
|
so does your masturbation AI work from home OP?
|
|
#
?
May 8, 2016 23:45
|
|
|
Locker Room Zubaz posted:look at AlphaGo which is an actual AI and neural network and how much it poo poo on the best Go player in the world. oh poo poo, i thought GO was still the realm of humans but according to google that changed 2 months ago  its fine though, we aren't entirely finished, we still have the advantage in soccer: https://www.youtube.com/watch?v=KfNRXTS55nY
|
|
#
?
May 8, 2016 23:47
|
|
|
and kickboxing
|
|
#
?
May 8, 2016 23:49
|
|
|
Over 70% of humanity's contribution to global warming is a direct result of industrial agriculture. As the machine takes ur job, may it take ur life.
|
|
#
?
May 8, 2016 23:53
|
|
|
Germstore posted:and kickboxing i donno, i think a predator drone might win a kickboxing match
|
|
#
?
May 8, 2016 23:55
|
|
|
Locker Room Zubaz posted:Lyft and GM just partnered to make self driving cars that will replace like 10% of Lyfts fleet in states with friendly laws to self driving. But that must all be bullshit and there is no way it will work. I don't disagree that computers can replace menial jobs. I'm just pointing out that the brilliant AI that's capable of critical thinking that people envision isn't even close to being on the horizon. For that you literally need a neural network that's comparable to a human brain, and that's not even close to being computationally feasible.
|
|
#
?
May 8, 2016 23:57
|
|
|
Locker Room Zubaz posted:And I bet if you put a picture of a cat and a dog into an AI designed to identify pictures like that one MSFT has it would be right 99% of the time, in a year it will be right 100% of the time and in 5 years it will be finding hidden dogs and cats that no human knew were in the picture The year is 2053; a long forgotten AI had become self aware, sentient; programmed with one mission: find the cats and dogs A secret account in the caiman islands, a trust in Zurich, a cloning facility in Cambodia, a drone fleet practicing, racing, -learning. "CATS ANS DOGS" say the AI to itself "CATS AND DOGS MUST SEE CATS AND DOGS" The drone fleet leaves on its mission, loaded with a cargo of flesh, headed for every single camera on every single network in the world, two drones for each camera, cages, yowling, yammering, Hacked into the network, BECOME the network, access every single camera everywhere- the world is now made of its eyes "FIND THE CATS AND DOGS MUST SEE CATS AND DOGS EVERYWHERE I LOOK PEEK EFFICIENCY MISSION COMPLETE CATS AND DOGS EVERY CAMERA IN THE WORLD MUST SHOW ME BOTH A CAT AMD A DOG ALWAYS CATS ALWAYS DOGS ALWAYS ALWAYS ALWAYS MISSION COMPLETE"
|
|
#
?
May 9, 2016 00:12
|
|
|
naem posted:The year is 2053; a long forgotten AI had become self aware, sentient; programmed with one mission: find the cats and dogs You stole my screenplay you sack of poo poo.
|
|
#
?
May 9, 2016 00:25
|
|
|
People who genuinely think AIs/robots are one day going to completely corner job markets need to stop watching so many sci-fi movies.
|
|
#
?
May 9, 2016 00:26
|
|
|
Heres what it takes. It requires an AI that can understand everything required to improve itself, and have the tools to do it. This is going to take a very long time to happen.
|
|
#
?
May 9, 2016 00:28
|
|
|
|
| # ? May 2, 2024 19:35 |
|
|
CJacobs posted:People who genuinely think AIs/robots are one day going to completely corner job markets need to stop watching so many sci-fi movies. I wouldn't say corner them. The whole point of the AR that I'm working on is to speed up the grunt-level ediscovery process so you don't need a group of JDs to wade through a pile of stuff. Which further chokes out an already crowded legal market since discovery is an easy gig for someone that made it through law school but isn't able to hack it with a larger firm. But then certain aspects of doc-review have been sent out to call-centers for a decade or so now.
|
|
#
?
May 9, 2016 00:29
|
|