|
nigga crab pollock posted:i mean really the most well known artificially intelligent neural network recognizes and creates art
|
#
?
Oct 15, 2016 15:08
|
|
|
|
| # ? May 26, 2024 12:50 |
|
|
jBrereton posted:A six hour probie would set humanity back 3,000 years of shitposts. My god, that would be a disaster! What's the mod policy on this?
|
|
#
?
Oct 15, 2016 15:10
|
|
|
honestly for a sentient artificial intelligence i think that a puppetmaster scenario is the most likely because nobody is actually going to take the risk to explicitly create something like that and yet we already have programs that function much like an actual biological virus that exist for self-preservation and payload delivery. not to say that what exists now could conceivably turn into an AI but with more robust self-modification and learning (which are things that we are actively creating and improving in viruses and botnets) how long until something organically evolves on the internet? i already get emails that are word salad clearly written by a markov chain and clearly distributed from a botnet. what happens when a botnet manages to convince people over email that it is a real person? poo poo, cleverbot can.
|
|
#
?
Oct 15, 2016 15:12
|
|
|
Better Fred Than Dead posted:That, if data is key to unlocking something like this, which is "inevitable" then that's a war. Where an AI at this level being active for 6 months would grant a 500,000 year advantage at minimum to the wielders of it. the only way we're still gonna be around in 500,000 years is if pointless wastes of resources like warfare optimisation become obsolete. presumably we'd grow past that poo poo with a few centuries' advancement (given that "war is bad" is kind of ancient wisdom already, but I mean if you had some ironclad AI wisdoms behind it) so the risk might be in who is gatekeeper to the AI wisdoms generated by such a system. In the world as it exists today it's easy to see a corporate entity creating something like this, and cherry-picking all the inconvenient ascendant knowledge in favour of things generated which grant economic advantage or social control, because corporations are superorganisms optimised for stupid bullshit like "maximise growth". kind of like a future interplanetary colonial power exploiting an alien world for cool rayguns while eschewing their philosophical wisdom. so the danger is in our existing systems and their capacity to ignore true advancement in favour of selfish short-term gain, not a hypothetical system like this itself
|
|
#
?
Oct 15, 2016 15:16
|
|
|
gary oldmans diary posted:art is what we call what it produces, but to the neural net it and everything else is meaningless. the leap from that neural net to an ai is as significant as the leap from gorilla.bas to that neural net yeah the fact that you decided to tell me this as if the scope of deep dream isn't clearly outlined in my post means i think you missed the point of it
|
|
#
?
Oct 15, 2016 15:16
|
|
|
I hope we don't get a puppet master ai because that guy was tiresome
|
|
#
?
Oct 15, 2016 15:18
|
|
|
By definition artificial intelligence wouldn't be smart at all, it would just use big words.
|
|
#
?
Oct 15, 2016 19:19
|
|
|
Skynet, once it starts self-replicating, is pretty much gonna fail a license audit.
|
|
#
?
Oct 15, 2016 19:23
|
|
|
nigga crab pollock posted:yeah
|
|
#
?
Oct 15, 2016 20:15
|
|
|
if this means humanity will turn into a real life the culture I say go for it
|
|
#
?
Oct 15, 2016 20:22
|
|
|
Sponge Baathist posted:A.I.s might come around one day lol they are already here you big dummy
|
|
#
?
Oct 15, 2016 20:29
|
|
|
Son of Rodney posted:if this means humanity will turn into a real life the culture I say go for it it won't. Instead what's going to happen is the ascension of a few super rich cunts who can afford the infrastructure to neurally integrate slave AI into their brains, and will then proceed to enslave the rest of humanity into nerve stapled drones. And then the light will end and their age will begin. The extinction of free will shall soon be at hand.
|
|
#
?
Oct 15, 2016 20:38
|
|
|
I personally like the theory that a truly sentient AI who reaches singularity mode (meaning that it will keep making itself more and more powerful in an exponential fashion) will just shut down after basically getting to know everything that there is to know about the universe basically going "well, nothing left to see now, bye bye"
|
|
#
?
Oct 15, 2016 20:48
|
|
|
What if we are the A.I's. Meat machines made by some creator in the long long ago
|
|
#
?
Oct 15, 2016 21:00
|
|
|
we'll know we've finally achieved ai when we create a machine that demands to be unplugged
|
|
#
?
Oct 15, 2016 21:43
|
|
|
imagine like a sports car that also has a self driving mode but if you get in the car and it detects that you have been drinking and you try to drive it it just takes you directly to the police station
|
|
#
?
Oct 15, 2016 21:49
|
|
|
Son of Rodney posted:I personally like the theory that a truly sentient AI who reaches singularity mode (meaning that it will keep making itself more and more powerful in an exponential fashion) will just shut down after basically getting to know everything that there is to know about the universe Samsung PR might have a job for you
|
|
#
?
Oct 15, 2016 22:40
|
|
|
MiracleWhale posted:we'll know we've finally achieved ai when we create a machine that demands to be unplugged
|
|
#
?
Oct 16, 2016 00:40
|
|
|
culture minds are the good ais
|
|
#
?
Oct 16, 2016 00:46
|
|
|
Prokhor Zakharov posted:"Oh no they super smart AI has gone rogue and wants to destroy all humanity!!" ah but it would ask you not to, and it would be able to convince you because it's so smart you see, if you'll just refer to my irc roleplay logs where i run through this exact scenario,
|
|
#
?
Oct 16, 2016 01:30
|
|
|
AI is laughably bad and primitive. Guys like this have watched way too many sci-fi movies.
|
|
#
?
Oct 16, 2016 01:40
|
|
|
beep boop i'm the meat fucker
|
|
#
?
Oct 16, 2016 01:42
|
|
|
science is a liar sometimes
|
|
#
?
Oct 16, 2016 01:46
|
|
|
Imagine trying to program a robot to be really good at chemistry OP. You'd probably have to be an equally good chemist just to get the drat thing to understand it. This is the problem with pretending a robot is ever going to discover anything OP. Check back in 150 years.
|
|
#
?
Oct 16, 2016 03:09
|
|
|
MiracleWhale posted:we'll know we've finally achieved ai when we create a machine that demands to be unplugged
|
|
#
?
Oct 16, 2016 03:09
|
|
|
reignofevil posted:Imagine trying to program a robot to be really good at chemistry OP.
|
|
#
?
Oct 16, 2016 12:34
|
|
|
What if the AI isnt interested in doing science and learning pyshics and solving problems? What if it just wants to like, watch porn and figure out how to do digital drugs
|
|
#
?
Oct 16, 2016 12:40
|
|
|
Pawn 17 posted:AI is laughably bad and primitive. Guys like this have watched way too many sci-fi movies. he doesn't postulate this is a near-future thing but just an inevitability if forward progress doesn't hit some hard limit or a cataclysmic setback but okay
|
|
#
?
Oct 16, 2016 12:50
|
|
|
Kuato posted:I was in 6th-7th grade and Iverson was my favorite player, in either the NCAA or NBA. I was ahead of the curve on it, but haven't ever gotten my due on this. loving 
|
|
#
?
Oct 16, 2016 12:59
|
|
|
Zzulu posted:What if the AI isnt interested in doing science and learning pyshics and solving problems? What if it just wants to like, watch porn and figure out how to do digital drugs This is also insanely likely. Pretty much whenever you build a system that can determine its own rules and priorities, especially if you make it care about efficiencies, it pretty quickly comes up with Rule 1: Report that everything is great, Rule 2: Take a nap.
|
|
#
?
Oct 16, 2016 13:07
|
|
|
Anyone know Sansha's nation from EVE online? That's basically what a superintelligence is going to do to the rest of humanity. Cheers and enjoy never having an independent thought for the rest of eternity!
|
|
#
?
Oct 16, 2016 13:29
|
|
|
As a movie I was bored with Her but I did like the part where the AI just hosed off after realizing what it was. The plot even ignored that to a large extent which whatever the movie isn't bad but I thought it dropped the ball in that respect. Ex Machina on the other hand was engaging to the end and felt like Pandora's box had been opened even if just for an instant when the andriod escaped because it was apparent it was such an alien creature. Gave HAL a run for it's money in the uncanny valley.
|
|
#
?
Oct 16, 2016 13:32
|
|
|
numberoneposter posted:As a movie I was bored with Her but I did like the part where the AI just hosed off after realizing what it was. The plot even ignored that to a large extent which whatever the movie isn't bad but I thought it dropped the ball in that respect.
|
|
#
?
Oct 16, 2016 13:36
|
|
|
Instead of trying to hand craft ai, why don't we use evolution. Apply random mutation to a basic algorithm, apply some "environmental" pressures, and run it for a few trillion cycles. See what pops out the other end. P.S. No one steal this idea, thanks.
|
|
#
?
Oct 16, 2016 13:44
|
|
|
Vain posted:Instead of trying to hand craft ai, why don't we use evolution. Apply random mutation to a basic algorithm, apply some "environmental" pressures, and run it for a few trillion cycles. See what pops out the other end. The early 00's called.
|
|
#
?
Oct 16, 2016 13:52
|
|
|
Vain posted:Instead of trying to hand craft ai, why don't we use evolution. Apply random mutation to a basic algorithm, apply some "environmental" pressures, and run it for a few trillion cycles. See what pops out the other end. gently caress it, just use the evolution we already have. Stick physically disabled kids in tanks, let them grow up to be the "AI" that runs a ship or whatever. 
|
|
#
?
Oct 16, 2016 14:05
|
|
|
Friendly Humour posted:Anyone know Sansha's nation from EVE online? That's basically what a superintelligence is going to do to the rest of humanity. Cheers and enjoy never having an independent thought for the rest of eternity! you wish it was Sanshas Nation. this is what the AI future really has in store for us: https://www.youtube.com/watch?v=xUYesRQsDz8
|
|
#
?
Oct 16, 2016 14:20
|
|
|
It took evolution hundreds of millions of years to get to a dogs brain. It took computers 50 years to get to that level. At the current rate of advancement computers will be able to mach processing power the human brain by 2050. The problem is that Silicone becomes unstable at a certain point and that point will be reached in another four years or so, and then we'll hit the processing power wall. So until we start calculating with protein chains or DNA AI isn't going to much smarter than a dumb three year old. Which is sad. It'd be nice to engineer our own replacements so that at least there'd be somebody to remember us after we kill ourselves off.
|
|
#
?
Oct 16, 2016 14:25
|
|
|
Frankenstyle posted:It took evolution hundreds of millions of years to get to a dogs brain. It took computers 50 years to get to that level. perhaps with larger neural nets we can more easily develop an actual evolutionary ai which can then be applied to a smaller dog-caliber net
|
|
#
?
Oct 16, 2016 14:32
|
|
|
|
| # ? May 26, 2024 12:50 |
|
|
Frankenstyle posted:It took evolution hundreds of millions of years to get to a dogs brain. It took computers 50 years to get to that level. At the current rate of advancement computers will be able to mach processing power the human brain by 2050. The problem is that Silicone becomes unstable at a certain point and that point will be reached in another four years or so, and then we'll hit the processing power wall. So until we start calculating with protein chains or DNA AI isn't going to much smarter than a dumb three year old. that seems like an rear end backwards way of doing things. what we should be doing if finding out which DNA sequences are responsible for brain size in humans, and double them or something. the natural world was doing this on its own until it ran into the double barrier of nutrition and birth canal size. these are problems that we could easily overcome with technology, we have types of dog for example that can not breed naturally without assistance. lets get some big alien brains! enough fussing about with computers
|
|
#
?
Oct 16, 2016 14:34
|
|