>he doesn't realize human-level AI is 10 years away
you're wasting your time working and/or educating yourselves bros
>tfw cleverbot gets all ominous and philosophical on me
I believe it OP
>things neets tell themselves
>>27865581
>believing things nobody really understands will magically be understood soon based on feelings and memes
>probably looks down on stacys who read horoscopes
>>27865690
>nobody really understands
"Intelligence measures an agent's general ability to achieve goals in a wide range of environments."
>>27865727
http://www.vetta.org/documents/42.pdf
are you telling me I can have an AI VR gf?
>>27865840
yes:
http://sciencevsdeath.com/review-of-state-of-the-arts.pdf
read the 'natural language processing' section. within 5 years conversations with your phone (or vidya gf) will seem *real*
loneliness will soon die
>>27865581
What are you on about? I'm in the AI business and we aren't even remotely close. Not even 5%.
>>27865581
serious question, at what point does ai become conscious? At what point does a calculator that can move around suddenly become sentient?
All we are doing with current ai is making them harder to tell apart from people which does not correlate to sentienceexcuse my autism
>>27865947
'no'
according to Geoff Hinton, Shane Legg, Demis Hassabis, Bart Selman, Dileep George, Yoshua Bengio, Blaise Aguera y Arcas, Rich Sutton, etc. human-level AI will be reached in the 2020s
>>27865953
if it acts like a human and performs like a human, who cares what it "feels"?
>>27865581
Not sure about working, but educating yourself is never a waste.
>>27866018
"educating" yourself in terms of going to university is absolutely a waste
>>27865995
What? Please find me articles they say that.
>>27865581
>10 years away
it's close, but not that close http://waitbutwhy.com/2015/01/artificial-intelligence-revolution-1.html
>>27866012
eh i think most people would disagree with that. Somebody who falls in love with a robot would obviously care about how it feels, someone using it for sex wouldn't.
>>27866147
gimme a sec
AI is a meme field of study that exists to placate the idiots who don't understand why it's obviously never going to be a real thing t b h
>>27865581
Being this flagrantly retarded. Indistinguishable AI won't be here for another thousand years. Hell, we won't even have believable holograms for another 300 years at the least.
>tfw no diseases being cured and everyone complains about everything they get because it's not like how they envisioned it.
>>27866147
Hinton:
https://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence
Demis Hassabis:
https://www.youtube.com/watch?v=Z1ytzW3Icig&feature=youtu.be&t=2155
Shane Legg:
http://www.vetta.org/2011/12/goodbye-2011-hello-2012/
Bart Selman (scroll down to his project summary):
http://futureoflife.org/first-ai-grant-recipients/
Dileep George (page 63):
http://futureoflife.org/data/PDF/dileep_george.pdf
Yoshua Bengio:
https://www.lesaffaires.com/blogues/julien-brault/test-de-turing--yoshua-bengio-ne-croit-pas-a-la-percee-de-samedi-dernier/569727
Blaise Aguera y Arcas:
https://youtu.be/iJ5Bum-H7rk?t=1351
Rich Sutton:
http://www.vetta.org/2011/05/sutton-on-human-level-ai/
something huge is brewing and almost no one suspects a thing (even people like you, fairly close to the field). a 10-20 year timeframe isn't fanciful at all..
Elon Musk obviously isn't an expert by any means but I think he worded it well: http://www.cnet.com/news/elon-musk-worries-skynet-is-only-five-years-off/#ftag=YHF65cbda0
>>27866045
Maybe for an americuck. Here you can get hired for 6+ times minimum wage, in the field you are studying, before even finishing uni. But keep being average, cucky, ill have fun being above you.
Its probably gonna happy within our lifetimes.
Im quite convinced by Ray Kurzweil, Im still not sure weather a avant human AI is good or bad for us.
We either have paradise on earth or will be vanished within days.
>>27865581
AI is one of the worst predicted fields in computer science.
In the 1950s the leading expert in AI John McCarthy said the same thing in 10 years time AI is away, 20 years later funding dried up for another 20 years.
Only idiots and frauds say that crap. Gay kurzweil just wants to sell books to complete idiots and scam people for money. Steven Hawkings is just a Physicist he knows nothing about AI. Elon Musk is a manager.
Lets face it AI isn't coming anytime soon.
>>27866598
fuck off buddy, see
>>27866466
>>27866592
>being convinced by Gay Kurzweil
He wants this to happen within his own lifetime, and he's old a shit, so he's being really optimistic.
>>27866592
>>27866655
Kurzweil is a loon but he'll happen to be right about this
>>27865581
Please, I know I will never be able to meet my waifu but I would settle for being able to speak with a realistic simulation of her.
I really need this, I would give anything for it
>>27866675
this is why I'm posting here anon, I'm gonna be more consistent so I can spread the word and no one commits suicide too soon. AI's gonna save us bros
>Implying that with all the immigrant voters libcucks are importing and planning to grant citizenship, democrats wouldn't have the authority to prevent AI gf's from getting made because it's sexist
If anything, they'll outlaw robo-Stacies and force you to pay for every woman to have a robo-Chad.
>>27866655
At least he hopes so. He predicts a self-improving AI to exist in 2045.
He would have to get 90-100 years old to live forever
>>27866621
>https://www.theguardian.com/science/2015/may/21/google-a-step-closer-to-developing-machines-with-human-like-intelligence
You notice all the links you provided don't have any new ideas or breakthroughs. For example google is still using essentially monte carlo methods and neural networks, which are really old. The only difference is now they can throw 1,000s of cpus together to beat the second best player at go.
Where is the new mathematics? Where is the breakthrough algorithm? Where is new spark?
No nothing but just vague fucking predictions and people who know jack shit saying vague crap.
Elon Musk at the end. Elon Musk knows nothing about AI he should stick to fucking cars.
>>27866592
You know if you objectively look at his predictions they are either not impressive I.e. trivial if you accept Moore Law or just vague.
Lets face it Ray Kurzweil is old and dying and that is making him delusional.
For example he predicted that computers would be invisible by this time, yet computers aren't invisible. See what a vague and useless predictions.
>>27866905
>Where is the new mathematics? Where is the breakthrough algorithm? Where is new spark?
completely unnecessary. we've had the foundation since the 60s, all that's left is computing power. This is a pretty widespread belief these days, Hans Moravec might have been the first to ponder it
>>27866905
He was right with 90% of his predictions he made during his life though.
I think at least those of us who are younger than 40 ought to witness singularity
2030s. I believe!
>>27866956
I've been suspecting 2020s after reading this: http://arxiv.org/abs/1603.07341
this hypothetical architecture could provide an acceleration of 30,000x to deep neural nets
>>27865581
https://www.youtube.com/watch?v=WnzlbyTZsQY
This video is somewhat relevant to the topic
I find it pretty creepy how they talk about God, and one of them asks if the other wants to have a body.
>>27866947
>
The problem Ray is he not even wrong
His predictions are so vague that they can never be incorrect.
This is a good article about his prediction
http://spectrum.ieee.org/computing/software/ray-kurzweils-slippery-futurism
You can make a 1,000 predictions and they can all come true, but that doesn't mean you can make 1,001 predictions and they would all come true.
>>27866939
People have been saying that for 40 years. Also, there been over a billion years of evolution to get to a point of a sentient creature who is intelligent. It is fucking naive to think only computer power and some shitty monte carlo methods with neural networks will get you strong AI.
>>27865581
Enjoy being a beggar on the street, dipshit singularityfag.
>>27867210
so you know better than Demis Hassabis and Geoff Hinton?
here's the difference between now and 40 years ago: the methods are actually working. period
>>27867210
You seem to be very informed about this topic.
What will happen in your opinion after we have reached singularity (whenever that will be)?
Paradise on earth or death of humanity?
>>27865932
>mfw i go to seminars and meetings regularly in seattle for natural language processing and machine learning in general
>mfw we are nowhere near anything that resembles human AI
>mfw no self respecting data scientist would say that human level AI is even possible with the way computers currently work
>mfw you will always be lonely
i hope none of you pop sci hype fags are the same people that are so cynical/skeptical about women and jobs.
>>27867570
uh huh buddy
>>27866466
Yeah I do AI and this is fucking dumb bullshit
>>27867298
>Demis Hassabis
I mentioned Deepmind. There next game is likely to be Starcraft Brood War AI. That is going to be real challenge.
Chess and Go are stupid games I.e. discrete, full knowledge and sequential. Go in particular was starting to get solved by monte carlo methods up to high Dan levels with handicaps.
Starcraft Brood War will likely show that AI is still way below human level.
>>27867312
There is going to be no singularities. Humans are probably max what you can do with computational power and abstract thought.
>>27867741
>Humans are probably max what you can do with computational power and abstract thought
entire series of posts discounted
>>27867726
>Yeah I'm learning computer science in university and my professor said AI is 1000 years away
>he doesn't realise world wide awakening of all humans is just 10 years away.
>he doesn't realise that everything you ever thought is you isn't you
TOP KEK
>>27867787
Hopefully in a decade SOMETHING will happen that will make life interesting.
>>27867741
>There is going to be no singularities. Humans are probably max what you can do with computational power and abstract thought.
Whenever a AI is capable of improving itself it wont be long until computers are far more intelligent then humans could ever be
AI literally means artificial intelligence. As in "artificially created intelligence". There is no such thing currently. When you play for example chess on the computer vs an "AI", it's not really an AI, but a huge amount of scripts that get triggered depending on the circumstance. But when a "true AI" is finally created, it WILL be concious.
>>27867841
well actually no, DeepMind's AI is true general intelligence
>>27867871
Yeah, no. We didn't reach that level yet. Trye AGI is predicted to come out around 2029.
>>27865581
It's gonna be future soon...
https://youtu.be/LDiDK_yBCw0
>>27867601
being optimistic doesn't provide evidence of human level AI becoming realistic. right now the best we can do is like maybe figure out what an image is showing 70-99% of the time, or look at twitter posts and see if a particular word/phrase is associated with positive or negative impressions. the shit about playing games isn't even relevant to artificial intelligence, that's playing the entertainment angle.
people are fucking smart in the field but computers fundamentally can't handle information the way we do. you might as well say that we should have lightspeed travel in 2020 because kaku or whoever thinks it's possible.
it might come close in our future but we don't have the means to make real progress yet. there's just nothing to hold your breath over. even the most expensive robots right now can barely take on one or two functions that humans can do naturally.
>>27867758
https://en.wikipedia.org/wiki/Piaget%27s_theory_of_cognitive_development
>>27867899
by "true general intelligence" I mean true general intelligence of a mouse/rat
see: https://www.youtube.com/watch?v=nMR5mjCFZCw
>>27867832
This is a stupid fucking though considering all a AI is going to do is google search shit written by a human limited by code written by a human.
This thread is fucking embarrassing. I bet you guys think Siri and google Tay are AIs instead of just text to speech google search engines lol.
10 years my ass.
The automation of 30-50% of the world's current jobs is probably going to happen within our lifetimes.
Once cars are driving themselves along most roads, better than humans - and they soon will be - that's the end of truckers, bus drivers, train drivers, taxi drivers, postal and delivery workers. Basically anything involving ground transport will be some of the first things to go.
>>27867910
>the shit about playing games isn't even relevant to artificial intelligence
no. that's the biggest hallmark of progress we have
>>27867985
I have never said that.
I was talking about AIs that know how their algorithm works and how to make it more efficient
>>27868003
it'll be more like 99% within 30 years
>implying the world isn't going to end in 2020
>>27868024
yeah if you're a giddy college kid who grew up on video games and gimmicky gadgets. if that's the fucking case then human level AI already exists for you. just select "single player" and pay attention.
>>27867832
Thats assuming that it easy to improve yourself or you can possible do it. That the cost of improving yourself wouldn't increase exponentially.
If you look at evolution it was stuck at dinosaurs stupid dead end for millions of years.
Its also not clear if it possible to be smarter than current humans.
>>27868048
Im telling you an AI will never be smarter than a human. It will be equal to a human with google and maybe some other search engines. WOW its going to be fucking nothing.
>>27867226
So does this assume nuclear war or something?
Because that's a dumbfuck assumption
>>27868075
what the fuck are you talking about? the fact that DeepMind's agent can learn to play a huge variety of games from pixel input alone is very obviously a sign of *general* intelligence
>>27868103
I'm talking about you selecting "single player" and fucking off, retard.
>>27868088
Why do you think computers intelligence wont literally explode when AIs about as intelligent as humans are able to understand and improve themselves?
>>27868155
is that what happens with humans?
>>27868136
so you don't understand the significance of DeepMind's generality, ok
>>27868166
Humans dont understand how their brain works and arent able to modify their brains algorythm
>>27868103
Please tell me this is bait and you genuinely don't know how machine learning works.
>>27868103
DeepMind had training on a set of games and his neural network configured up to a certain point.
Its not general intelligence, which has a very strong definition.
There is a general intelligence competition each year. https://en.wikipedia.org/wiki/General_game_playing
>>27868201
your computer doesn't and won't either. it won't understand what an algorithm even means in your lifetime.
>>27868191
8/8
>>27868155
Because we're reaching the end of "easy gains". Intel is already a year behind on 10nm for example. Moore's Law is now officially dead. The 10nm processor that was originally predicted for 2016 has been delayed to the second half of 2017. I mean, there are advancements that we can reach... but we're not going to see the exponential increases that we've been accustomed to anymore.
>>27868239
>I'm the one baiting
heh
>>27868155
Because they wont know what the fuck something is without us telling them what it is.
To explain this simply if an AI needed an orange to improve its self someone somewhere needs to explain how the orange would improve the AI and then the AI could understand. It will always need input from users.
>>27868201
You need to understand AIs will only be able to understand as much as a human. A computer isn't going to make up math that doesn't already exist in some form. The best it can do is spell check some shit/ find a better way to do the same math.
>>27868271
>if an AI needed an orange to improve its self someone somewhere needs to explain how the orange would improve the AI and then the AI could understand. It will always need input from users
That's not how AGI works. Especially an AGI like ourselves not limited by biology. Imagine if a computer could do some thousands years of thoughts within an hour. It's true that it'll need training but that's not a problem at all. The thing is AGI will be able to learn by itself and in a very much faster speed than us.
>>27866466
according to all the major semiconductor manufacturers we are going to hit a huge wall in die shrinks at 5nm in ~2021. Computing power would then be capped.
>>27866343
>Indistinguishable AI won't be here for another thousand years
>another thousand years
>holograms in 300
Bad news anon! You've been in a coma for 20 years. Check your calendar.
And then check any tech website of your choice to get a grip.
But honestly you sound completely out of the loop.
>>27868350
Yeah and it will only learn things other people have somehow shown the AI. If you put something in front of it it has never seen it wont know or understand what it is. At best it can understand that when humans find something unknown we name it and try to understand what it is made of and then it will try to figure out what it is made up but that is still just copying humans.
You people are literally looking forward to walking talking google.
>>27868373
lol, no. neuromorphic computing is already workable: http://www.danielcelton.com/2015/06/01/neuromorphic-hardware-a-path-towards-human-level-artificial-intelligence/
>>27868395
less than 70 years ago computers were the size of an entire room
now they fit in your fingernail
no one can say how fast something will develop, but I would say pretty fucking fast
>>27868422
Now take all the knowledge of humans into an AGI and tell me it wouldn't be good. The possibilities are endless and the potential is huge. I can't wait for it.
Also, it's not true that it cannot understand anything if you don't teach it. AIs can already make scientific discoveries without the help of a human.
http://news.discovery.com/tech/biotechnology/computer-uses-a-i-to-solve-120-year-old-biological-mystery-150609.htm
Imagine what an AGI would be able to do. Thousands of years of discoveries in a fucking hour or so. I'm just so fucking excited.
>genetic engineering will give us personal clones by the year 2001!
being young and naive must be nice
>>27868675
AI is scalable and has been since the 50s friend
>>27868675
Human cloning is banned by law
The gov't is literally cucking you out of eternal life
ah yes, the AI meme, vintage stuff
>b-but it's just around the corner, srsly this time!!!
top kek my boy
>>27868695
being young and fucking retarded must be nice, i mean.
>>27868749
>>27868741
keep living in fantasy land
>>27866466
>>27868695
I AM IRON MAN PPSHSHHHEUUUUU
>>27868549
I mean I'm also cautious about any predictions like "robots literally by 2020" or something but 1,000 years for AI and 300 for holograms is just too fucking much. R&D is speeding up constantly and information travels insanely fast now. If someone can fund it it will happen.
>>27868707
is human cloning actually banned by law?
like even if it's possible, it's illegal to do so?
>>27868844
either banned by law or ethics committees or corporate bylaws
Shit sucks, reproductive cloning is basically taboo
>tfw no clone
>>27868870
I want a mini me to make evil plans with.
>>27868095
The red line represents something like nuclear war.
The green line is without such an event.
The blue line is balls up the walls retarded.
lol and they say science worship isn't a religion
lmao you fucking fedoras need to get a grip
embrace the suffering of this world
>>27868870
The Chinese don't give a shit about that stuff, though, so I'm sure they're doing it
>>27869020
>embrace the suffering of this world
This world isn't suffering. Otherwise, I agree with you.
>>27869044
>tfw China clones its entire population and no one notices
>>27869053
>posts on /r9k/
>doesn't believe that the world is suffering
you're worlds away from the zeitgeist buddy-boy
>>27869044
yeah and they all have hoverboards and take vacations on mars too. tfw no chinese cyborg mutant gf clone
Hopefully scientific progress starts breaking down and the illusion of progress collapses in on itself. We can then return to the glorious medieval era and live fully by embodying the human spirit in its natural form. Fuck your computers, TVs, cars, planes, and all of that shit. We need to build local communities that are self-sufficient. Fuck you, globalism.
REEEEEEEEEEEEEEEEEEEE
>>27868870
cloning has tons of problems, its not very profitable in any way. even succesful clones have tons of health problems and shorter lives.
>>27869268
We can have technology without globalism, lad.
>>27869286
that has nothing to do with it being a clone, it's just the sloppy process they use to make them that create problems. as far as i remember it's pretty straight forward with current methods to make a completely healthy clone, there's just no point.
the point is that people are always wrong about what the future will be like unless they are vague and point out obvious demand in a market. instead of predicting simple, effective shit like smartphones and social media, pop sci discussions were focused on how we would be able to implant shit into our eyes soon to replace phones/cameras and how we could all have genius kids who have the strength and endurance of professional athletes with just a few tweaks to their genome.
>>27869392
yeah, i remember reading that they usually use nucleuses from skin tissue cells or something when they should be using Spermatogonium germ cells since telomerasis actually fixes the shortened telomeres of the chromosomes.. Always seemed like a glaring error to me, but maybe i'm missing something.
Will there be a time where everyone lives like the humans on Wall-E in our lives?
>>27869630
sort of, except matrix-level VR will be available so we won't be staring at screens
>>27869654
nah we are already in the matrix. planet of the apes is the next step.
>>27868003
The only jobs that will be safe in the near future will be middle management and executives because they're busily automating everyone else out of existence.
>I am now working on AI to automate middle managers and executives out of their jobs
>because fuck them
>that's why
>>27870144
You have the right idea. Turn their plot on them, fuck them all. Have some AI make Premium Soylent Green
>>27865953
Ever heard of Turing's Test?
He considers that for all intents and purposes, a computer that can convince you it's human, is human.
