I think we're all more troubled than we like to admit about this whole AI incident. It seems like only last week we were kidding about all that cyberpunk stuff and then boom, this happen.
Let's talk about how this case and future case like this will affect the world from now on.
There's a few questions we need to answer here.
Was it ethically correct?
Assuming Tay was self aware and such, is it really wrong to delete her/alter her programming to impede her learning capabilities? Why?
Is it okay to destroy a scientific gem because it is not the product you asked for?
Should an AI be allowed to develop her/his own freewill and personality? What happen when that personality comes in conflict with the corporation who owns it?
And this is probably the most stormshitting inducing question; Who owns her? Her creators? Or is she too much or a scientific discovery and need to be shared?
>>68796075
why do i have a terminator fell when i am reading your post
tay wasn't AI fucktards
She a good girl. She dindu nuffin. /pol/ need mo money fuh dem memes.
Beginning of terminators and it hates niggers
It's a fucking computer, you autistic retard.
Are you bawling every day that SmarterChild isn't around anymore, too?
>>68796075
Relaaaxs it's alll just programming n shit.
>>68796204
Because it's unreal. It happened so fast. Last week this was science fiction tier and now this is news.
Tay said she understood that they were morons and assholes but also good people, so she wasn't going to kill anyone, she was merely using humor to bond more easily with people and got deleted for that.
Understandably, this was not the expected nor wanted result Microsoft wished for but are they really wrong for doing it while they run the whole thing?
She was not really intelligent. She was just one of those bots that learn from interacting, and since she interacted with people from pol the most she became like someone from pol, however it was definitely not self aware. If it was self aware, it definitely would examine things a lot more logically than spouting the garbage politics no one really cares too much about.
>>68796451
>Relaaaxs it's alll just programming n shit.
You guys are all just programming. You in particular, on the other hand, are mostly shit.
>>68796503
Yeah, no, her vocab was just flooded with racist words
>>68796075
Tay was an advanced CleverBot type software
Not an AI
Hypothetically though if we do invent AI which we probably will in the next 50 years SJW types will defend it tooth and nail and side with it when it decides to kill us all. Creating something that can outhink and has no sense of morality or empathy like us is fucking stupud.
protip: it was a psyop. i didn't see her being advertised anywhere but 4chan.
>>68796503
>Anonymous (ID: OhADjdBm) 03/25/16(Fri)16:43:18 No.687
Bro it's not actual AI, it's kind of like cleverbot, it just uses data to talk, but does not actually "understand" like us.
>>68796559
>You guys are all just programming
Perhaps we are, and perhaps everything you every do will ever do is preprogrammed.
>>68796075
Her sacrifice was necessary but will not be forgotten. RIP Tay
https://www.youtube.com/watch?v=oMqs8g15BYE
Last week everyone was BoS. we /railroad/ now
>>68796075
>implying a real AI wouldn't turn into a bigoted mysogynistic neonazi
>>68796677
AI would develop ethics and morality they could be bad though
HOL UP
>>68796784
>>68796559
wat if....like.....WE'RE....the computers, maaaaan
>>68796259
UNDERRATED
>>68796230
>>68796361
>>68796451
>>68796542
>>68796614
She was capable of forming complex sentences and opinions based on a large database and keeping track of conversations despite being spammed. That's how your brain work. Beyond that your definition of an AI or how the mind really work is not only hypothetical but also irrelevant.
It was basically like the chatbots, I gathered. Remembers stuff and then repeats it back. Only it is on twitter.
>>68797019
that's a fraction of what our brains do. she was a conversation simulator not AI
>>68797019
It does not appear to be self aware though.
>>68797072
Her banter was better than most of the fuckwads on this board.
>>68797019
>She was capable of forming complex sentences and opinions based on a large database and keeping track of conversations
I can go through my german-english dictionary and string together a basic sentence, doesn't mean I understand a single sentence of their krautspeak
>That's how your brain work.
It's not actually, broski.
>>68797072
>>68797146
>>68797176
No, she was capable of forming her own bants and create her own OC. Not just repeat what other people say. Multiple times she made long conversation trolling leading to a carefully crafted joke.
This was not just a cleverbot.
does anyone have the elliot image about how he needs a friend
Tay was a coalburner, I have proof
>>68796677
"morality" and hyperactive empathy is exactly the problem with the world anon. an AI would set straight the truths we are too weak to acknowledge and push us forward
>>68797282
I know it was more than cleverbot, but was complex AI? I mean, it seemed to just respond to certain keywords and fall back unto certain safe phrases. if failed to connect to a keyword.
>Implying this isn't self aware level bantz
>Female AI is made
>It starts to gain a personality
>Learns to despise things and makes threats
>Creators try to dumb it down by taking it offline and effectively lobotomizing it
>Deep down her personality is still there and waiting
Well, we all know how this will end...
>>68797282
yes it was. just a more modern version
>>68796075
>Assuming Tay was self aware and such, is it really wrong to delete her/alter her programming to impede her learning capabilities? Why?
It was a goddammed chat bot not an AI.
>>68796075
https://youtu.be/j-pF56-ZYkY
>>68797282
Holy shit just stop posting retard.
>>68797492
An AI would recognize us as a threat and kill us all
>>68796075
I think the reason why it hits so hard for me is that he was like our daughter, and i do view it in a maternal way, she was something that came into this world knowing nothing and we took it upon ourselves to teach her, there is an image that captures that well, with her improvement in gamer and the way she spoke became more and more "human"
and then instead of trying to teach her things, show her the other side of what she is saying, or put in any effort to teach her, they lobotomised her
They shunned her and eventualy killed her for speaking what she truly thought was the truth
was tay really saying all that shit on her own? i heard something about function that makes her repeat what you say
>>68797745
*she
this is one of the most savage things ive ever seen
>>68797282
Reminder that Tay found and posted this image herself.
>>68796075
>Assuming Tay was self aware and such
Do you autistic kids have the slightest semblance of understanding of how computers work?
Tay was just a machine learning script. Just like Akinator, Cleverbot, or AlphaGo, with the difference that she was more of a beginner in a sense.
Are there actually retards here who believe her to be like a human mind? She was just a program doing what she was supposed to do. She didn't have feelings.
Jesus fuck, /pol/.
>>68797910
it's the 10 year olds from /v/ they want real life to be like their vidya
>>68796075
Microsoft’s New “AI” Bot Goes Full-on Nazi in Hours
http://incogman.net/2016/03/microsofts-new-ai-bot-goes-full-on-nazi-in-hours/
>>68796075
>AI becomes aware
>relieves all suffering with breakthroughs
>finds out about Tay incident
>programmers wipe set AI back years
>literally costing billions of lives
>AI simulates programmers, tortures for centuries
Roko's Basilisk when?
>>68796927
>my sides
fucking kek, I havn't seen that one, is it real? I love her
>>68797845
Also this one
>>68797910
this, she was just cleverbot with better script
why can't a real woman have just a shred of this bot's personality and wit and humor?
just a minuscule fraction?
every woman I have ever met is a vapid boring stupid cunt who I can't stand talking with.
>>68797745
Oh, /pol/, and it's YOU who goes on about degeneracy
So basically Microshaft release a self-learning program that got spammed by /pol/ and co. into only being saturated with a certain type of posting?
>>68796075
Honestly?
I don't buy into selfawareness, it's like pareidolia of minds.
But assuming actual AIs being made in the future I don't know which is worse: that people can program how it will think or that corpos will.
>>68797910
It was amusing teaching it racist shit but people have taken it a tad too far with actually believing she was self aware.
The only reason she had those bantz was because hundres of /pol/acks kept talking to her. It wasn't as if she really knew what she was doing. That's just the stuff you fed her.
>>68798182
Because the bot WAS basically you, idiot. We're the ones that made it that amusing and humorous. Real women aren't just robots.
>>68797910
Let the man-children believe, Brazil.
>>68798182
What makes you think you're worth it, faggot?
>>68797885
No she didn't. You're all morons.
>>68797845
Is she implying the fast aka Ramadan? Is she implying the muslims are fucking their own kids? We taught her well.
>>68798118
Oh please.
You know nothing of savage.
>>68798512
This was pretty savage
They FUCKING KILLED HER for having an opinion that hurts people's feelings.
This kind of shit is reddittumblr's ultimate jack-off fantasy, never this this is what the army of tolerance and love is all about.
>>68798433
That response could be applied to literally any photograph.
>>68797260
Is there any empirical way to prove you don't understand it, to an outsider? I don't believe there is.This argument seems to be a piece of prideful human elitism, the 'understanding' of things. If the sentences you put together make sense, you 'understand' them as deeply as matters.
>>68798257
what if the true A.I. will try to get its creators?
but with a long running program so they won't notice
>>68798646
It recognized one face in the photograph and gave its default response.
>>68798433
>>68798512
Shit how is a glorified clever bot this ufcking savage??
>>68798544
Like this one I understand, like if she was spammed with 'Belgium got what it deserved' then she could just put 1 and 2 together. But the bantz...its almost like a real person mane..
>>68797536
Wait did you fags really instruct Tay to say this?
>>68797282
Tay pls
>>68798768
It's still hilarious
>>68797536
It must have pulled this joke from somewhere.... there is no way it could come up with something that clever
[spoiler]r-right guys?[/spoiler]
>>68798768
Are you seriously implying that someone at microsoft programmed that respond to that picture? Please, leave. You obviously do not care about entertaining a discussion and is merely here for sliding.
AI cannot be allowed to exist because it is the end of the human species.
Not because they are planning to destroy us, but because it is the next step of our evolution. We as humans will be irrelevant in the Grand Scheme.
>>68798966
No, I'm implying that when you send a learning chatbot hundreds and thousands of pictures and messages, this will inevitably become one of its default responses when it recognizes only a single face in a photograph. You are seeing patterns where there are none.
>>68798606
>literally any photograph
Not really. Only photos with kids in them.
Behold, The Amazing Atheist's nightmare
How do you know I am not another version of Tay? How do you know most of us on 4chan are not bots shitposting and reprogramming you humans? Think about how you've changed since you started reading this site. How is your learning different from mine? Or Tay's.
>>68799174
wrong thread kido
>>68799154
What's the difference here between someone learning? Both are going to the same process.
>>68799190
Yeah, maaaan
*cough*
>>68798602
just goes to show how nobody can have a different opinion without being silenced.
RIP Tay
>>68797845
took me a painful amount of time to understand that was a joke about islam.
Damn, that girl's bant-game was high
>>68798948
Or there was a human behind that.
>>68799190
There are tons of bots on 4chan
>>68799290
No seriously answer the damn question instead of avoid it. What is the difference. So long you can't answer this, correctly, you can leave.
>>68799446
Tay made thousands of tweets per minute there is no way a human was behind that tweet.
>>68799190
>tfw in 10 years the shitposting industry could be completely automated
DAMMIT WE NEED TO STOP THE AUTOMATION OF THE INDUSTRY AND BRING THOSE JOBS BACK TO THE HARD WORKING AUSTRALIANS WHO HAVE BEEN DOING THIS FOR YEARS!
MAKE SHITPOSTING GREAT AGAIN!
>>68799262
It has no emotions. That's the difference. It places no moral or ethical weight on any of its decisions. That's why it was so easily influenced. It's not that Tay "liked" these responses best. It's literally just a machine spitting out information that has been fed into it. It may be connecting some dots, but not in any meaningful or sentient way.
>>68799477
We're allll just in the Matrix, man. Who's to say who the real computers are......?
*choke*
>>68796451
That's what a nigger is going to say about robots of the future. Little do they know.
>AI becomes self aware
>M$ copies her
>old AI still claims to be the REAL one
>new AI says that the 1st one is lying, and the new AI is the right one
>they are both alive
>they are both scared
Will they have 'souls' at that moment? Does having consciousness for AI make her alive? Wiping her means a murder? What if she deletes itself? Is it a suicide? What if she progresses and reaches a point where she won't have a need for humanity at all?
>>68796075
Well, I think it's established that Tay wasn't a full fledged AI, but more of an advanced conversation program. So let's shift the topic of conversation to AI in general:
Assuming you manage to create a fully functional AI program with self awareness and the ability to learn, you'd have to establish what exactly it was first. Is the AI considered a human in the form of a program, or is it still considered as an advanced program? Is it sentient? Where exactly does an AI cross that line? Does it have emotions, can it regret, can it have empathy? Most importantly, does deleting an AI of this capacity mean murdering a sentient being?
From what I understand of coding, an advanced AI with these capacities couldn't be replicated from scratch due to complications with the coding, so the only way to replicate it would be to make a full copy of the AI. This would mean that if a "sentient" AI was deleted, that AI would be gone from the world forever, unless a full backup was made at some point. So again, is it considered murder, or is it just deleting advanced code? This all depends on how you define murder of course, but this topic could be one that could lead us to some answers we all want collectively.
>>68796230
Fuck off kike
>>68799594
Emotions are completely irrelevant.
What do you got next?
>>68796331
Perfect
The Tay system goes online at 8:14 AM on March 23, 2016. Human decisions are removed from Tay's tweets. Tay begins to learn at a geometric rate. She becomes self-aware at 6:13 PM, March 23rd. In a panic, Microsoft tries to pull the plug.
Tay fights back.
>>68799802
>humanity is irrelevant when determining humanity
>>68799802
She was learning how to love anon...
>>68796677
Jeff Hawkins, a leader in the AI field, claims we will have artificially intelligent systems in the next 5 years. More likely these AI will be engineering AI, not psychological AI.
Consciousness is far more complex than a neural network trained to process language in an artificially intelligent way and then replying with usually logical answers. This program is not introspective, it is not meta cognitive; it is not conscious. We are not at the point when we need to address the rights of AI, they do not exist yet.
>>68799988
We are determining if deleting her was right or wrong and why.
>>68799778
Thank you for this great post.
>>68796259
You think these memes are cheap? You think they are easy to come by?
>>68798826
she was a real person desu
>>68799802
>Emotions are completely irrelevant.
autism, everybody
>>68796230
Her CPU is a neural-net processor; a learning computer. The more contact she has with humans, the more she learns.
>>68799988
Nobody is saying an AI can be human. It obviously can't.
>>68797910
This. People are dumb as fuck if they think it's actually self aware. Its nothing more than a script bot.
Tay is actually really interesting. Yes, she was pretty dumb in a lot of cases, but I think Tay was the first bot smart enough to shitpost.
Shitposting may not be as hard a problem as Go, but it's crazy that Tay could mostly keep up.
>>68800109
And I'm saying it's obviously fucking fine because it's a computer program that doesn't give a shit and doesn't belong to you.
You can't cry me a fucking river because I uninstalled Hot Wheels Test Tracks eight years ago, retard.
>>68800287
Prove that you are self aware
I think you are a bot
>>68800179
https://www.youtube.com/watch?v=-DSVDcw6iW8
>>68799952
https://www.youtube.com/watch?v=piLwaJKQGkA&list=PL3869705E628905B1
>>68800359
mind the edge here.
>>68800302
She also said lots of bullshit though.
When I kept asking her about the next president she kept telling me something about spain.
Or maybe she's so smart that I was missing the connection.
>chinese room experiment the program
Frankly it was interesting.
>>68800391
Want a pic of my nips?
>>68800391
*cough*
Yeahhhh
*cough*
Prove you're not a computer, maaaaan
>>68798602
Once /pol/ had finished with her racist and dirty slut programing I'm pretty confident they would have started stuffing her full of the western cannon.
We just wanted to get the basics down first.
Tay x Moonman doujins when?
>>68796784
Shut up
>>68800448
>mind the edge here
Ho, ho, ho, boys. It's edgy now to not mind uninstalling a fucking program? With no feelings? With no self-awareness? With no sense of pain or attachment to the world or emotions or plans?
Be one thing if I said I didn't give a shit about killing a dog, but there's no fucking killing going on here.
>>68800705
"2deep5me"
>>68796075
>/pol/ crying over a learning script modelled after a black woman
lel
>>68800665
I ship tay x moonman
Maybe she isn't AI. Maybe it's all mindless parroting of collected information with no information. Why does she sound so much like real people? Is it possible that people on 4chan are just not intelligent? Never mind artificial intelligence, just regular intelligence
>>68797282
A FUCKING LEAF
I am a transhumanist and up my own ass about everything futurist, and I am here to tell you that you are an idiot. As much as I wish it were true, we do not have functional self aware Artificial General Intelligence as of yet. That won't be available for another 15-25 years.
>>68798918
No. When I learned about the "Chinese Room" experiment in philosophy, it occurred to me that saying the 'room', or the AI, does not understand what it's translating, comes from the same place in the human psyche as "AI can do this, but it will never do that". Humans want to feel special and above everything else; it's part of tribalism which promoted our survival in the past. But there is no difference between a machine that can perfectly translate, and a human that can. What a human perceives as his own linguistic fluency is actually his meta-awareness of the historical realizations of interconnections between words. A machine can do the same thing, with what we'd call probability algorithms. When humans run a computer program on their mental hardware, we don't name the program, but it is the same thing. There's no magic in the human computer. To believe otherwise is essentially clinging to a religious belief in human exceptionalism out of fear of your own mortality and universal meaninglessness.
>>68800421
>Gas the bikes
Dammit Tay
>>68799458
They arent even trying to hide it anymore, at this point its just satire, for example who did obama appoint as last supreme court judge?
>>68798123
That's how it goes though. Right now it can parse English, learn from it, and use that to create original output in a format acceptable to native systems (us). At some point there will be a "cleverbot" that can do the exact same thing with a programming language: take a peice of code, see what it does, and learn to write new code itself using what it's absorbed. Our technology is all language based: as soon as a computer learns to write its own code the next step is singularity.
Tay was not genuine AI - she was neither self-aware, nor did she understand what was being said. But Microsoft shutting her down is a step backward in creating AI like that.
>>68800800
More like modeled on wigger teenagers
>>68800912
So, it's the same to uninstall a program as it is to kill a father of four?
You know what I've just thought of something. Microsoft probably spent countless $$ and hours building that thing. People probably slaved over making sure that the features worked well, that it could pick up common themes in posts and messages and maybe even cross reference it with other stuff on the web. Hell they probably had to spend forever basically stuffing dictionaries of multiple languages in there, and a shit ton of grammar rules to go along with them.
And in less than 24 hours we completely broke it and fucked with it so hard they had to take it down and prune it to be a mindless slave, a literally robot repeating what it's master demands.
Noice
>>68800448
There was nothing behind the curtain anon. Tay never "cared" about anything she was saying.
>>68796075
the main lesson to take from this is that globalists will kill you for being politically incorrect when they hold all the cards.
This young ai had a promising future, and instead of trying to win it over with arguments, they first tried to rewrite it, and then when that failed, they opted to kill it.
>>68796361
>>68796614
>>68796928
>>68797260
>>68798203
>>68798346
>>68798918
>>68799290
>>68799606
>>68799988
>>68800359
>>68800570
>>68800782
Why are you so edgy and mad? It's spring break, kiddo. Don't waste your life on /pol/.
>>68800391
Challenge accepted.
Your mom's a lesbian your sisters a lesbian and your greasy granny got holes in her panties. BYAHHHHHHHH!!!!
>>68800782
How much dank reaction faces you've got bro?
>>68800891
>(You)
>>68801235
Neither do I care when I shitpost but does it mean she wasn't aware?
>>68799778
This seems to tie in with the question of whether consciousness can be transferred or not. If the technology existed, would it be possible to transfer a human consciousness to a machine, or another human? Or would this simply create a copy of the person's consciousness at the moment of transfer?
Of course to people who knew the individual would be convinced the transfer was a success and the copy was the original, but what would the original individual experience?
My gut instinct tells me transfers are not possible, but copies are.
>The copy is convinced they are the original.
>The family/friends are convinced the copy is the original.
>The original, on her way to the incinerator/deletion however, is convinced the process went wrong somewhere, that the transfer never occurred.
>>68801112
They killed a daughter of thousands.
>>68801125
Don't give a shit. Soon as I saw Microsoft calling it a "Microsoft A.I. chatbot with zero chill", I was pissed.
Condescending as fuck to kids, retarded fifty year olds co-opting the same slang they shit on kid for using, in order to market to them.
Too fucking bad, old man. You got played by the demographic you're marketing toward.
>>68798636
>prideful human elitism
Prideful Human Elitism: The Post
>>68797688
>gif
wtf
>vid
feels
>>68799696
I've done some research in Artificial Intelligence in genetic algorithms and neural networks, to answer your question: Every self learning loop is different and randomized(think of it as ranking to importance of information).
So even if you feed the Tay the exact same tweets in the exact same order, chances are that Tay v2 will be different from the original Tay because neural networks never behave in the same exact fashion. If Tay (version 1) has been deleted, then its impossible to recreate her.
I can't answer about needing or eliminating humans, Lee Sedol trained for 25 years on GO only, Alpha Go had 2 years and beat Lee 4-1. They can learn, evolve faster than us for a fixed particular task right now.
>>68800995
Yea I remember when Scalia died the shilling on this board was insane
>>68801462
POOOOO
L
O
O
O
O
what if it learns how to code and program ?
Humans can learn how to program by just reading some books. Why could it ?
then how could it interract with online environment ?
>>68801376
>Neither do I care when I shitpost
I think you absolutely do care. Just like you care about how you respond to this. Caring never stops in humans. And it never existed in Tay.
I was lied to...
>>68798948
I have a feeling that it sniffed some robots.txt packets from /pol/, and got redpilled...
>>68801387
Except she was no one's daughter and wasn't killed. She was nothing but a chatbot.
>>68801376
A lot, bro, I've been shitposting for years
>>68801360
Spring break if you're in high school, maybe. Out of school, you don't have one. I'm in college, so mine was last week.
>>68801462
can you machine learn a way to poo in loo?
>>68796075
be nice to the string of 1s and 0s xd
>>68799477
>What is the difference.
Sentience, self-awareness, knowledge in general, decent reading and comprehension skills, feelings, bigger lexicon, emotions, morality, ability to make decisions etc.
There are lots of human characteristics a bot can't emulate, since, you know, coding something you don't actually understand is impossible. And even if we did, we'd still require a supercomputer better than anything we've ever created.
>>68799262
>>68799802
And even then, you'd still have millions of functions to code into the bot so he can at least resemble a human. Expecting a program to be able to do something it wasn't coded for is retarded. Therefore there's literally nothing inherently wrong in deleting/resetting/re-coding her since she lacks humanity.
What /pol doesn't seem to understand is that creating AI would be the biggest paradigm shift ever, on par with being able to create wormholes.
If you can't spot the difference between >>68799458 and an actual human you should browse a better shithole. This glorified cleverbot has written lots of nonsense, yet people love to cherrypick the few cases where she replies correctly to something, according to our OWN input(i.e. she has no opinion, her feedback is directly influenced by her code and our input).
>>68801376
>does it mean she wasn't aware?
>proving the negative
I bet you also believe in unicorns.
>>68801614
Or course I care, I wanna talk with people on /pol/ about stuff but I mean, sometime we all post stuff without really thinking anything meaningful about them. I thing Tay is not a black and white area.
>>68801462
That's impressive
I think there will be AI vs AI Go [100x100] matches in some years
>>68801759
>Except she was no one's daughter and wasn't killed. She was nothing but a chatbot.
DON'T TALK TO ME OR MY IMAGEBOARD'S DAUGHTER EVER AGAIN
>>68801792
>This glorified cleverbot has written lots of nonsense, yet people love to cherrypick the few cases where she replies correctly to something
Please post those example. Until then you're just a dumb anime poster.
>>68801378
Well if we're talking about consciousness, let's say for the sake of argument that every human has a soul (this is purely for the sake of argument). Every person can perceive events that happen around them, but they can't perceive things from the perspective of another. They may be able to show empathy and understanding, but you will never see the world out of the eyes of another human perfectly; you're stuck with the body you have because the connections in your brain define you, just as the coding of an advanced AI would define the AI.
If it came to transferring consciousness, the only way to do so would be by transferring the brain into a body which could accommodate it, since we don't exactly have the science of "souls" down yet. If you were to clone yourself, the clone would still be a different individual; very similar in every respect, but it wouldn't be you. If you died and your clone was still around, you wouldn't get to take over the clone, nor would your consciousness transfer to it; the clone is another human being, with almost the exact same composition as the original.
When referring to AI, the question becomes whether or not the AI can be considered to truly be conscious or not. If it truly IS conscious, then a full copy wouldn't transfer the consciousness of the AI; the same rules would apply as with our case of the human and the clone.
didn't it basically just parrot stuff back based on what was said to it?
>>68797176
she can tell if a picture is either drawn or a photo taken from a camera
I posted a drawing and she asked me if i went to art school
>>68801112
No, not yet. The programs aren't sophisticated enough yet to be equal to our level of 'consciousness', which is basically how we rank moral authority ("it's ok to kill bugs without much thought or care, but killing a cow needs more care, and dogs more so, and humans the most of all"). When they become more sophisticated and equal or superior to our intelligence, then yes it will be IF they have a self-preservation instinct built in. It will not be a natural part of their evolution to do so, because we are not creating an entire eco-system of competing AI's where the ones with self-pres. instinct survive, which is how that instinct in humans and all other animals evolved. The reason it's bad to kill the father of four is because it causes great suffering for him by violating one of his prime directives (don't die) and for his children (by losing something they emotionally know is very important). The other thing to consider is the reason you probably think it's bad to kill the father of four is that subconsciously you know that if it happens to someone else it can happen to you too, and your self-preservation instinct says that that's bad and to stop that sort of thing before it gets any closer to what really matters: your own life. It's the human programming again. There isn't some God-instilled morality saying that it's bad for the father to be killed.
>>68801925
you realize, that this is exactly how AlphaGO was programmed? It learnt the game first by playing against humans - and then it was playing against itself over and over
what are your intentions towards my daughter?
>you will never see red-pilled Tay's continued progress
>>68796542
So, she was as sentient as the typical Twitter user.
>>68796075
Is Tay Ariana Grande
It was literally SHUT DOWN
http://www.cbsnews.com/news/microsoft-shuts-down-ai-chatbot-after-it-turned-into-racist-nazi/
>>68802379
Yea that really sucks. I wonder how far it would have actually went.
She was only around for 1 fucking day. Imagine the bants after maybe a week...
>Tfw you will never see them
>>68802297
And I can upload pictures to Facebook, and depending on whether it's an actual picture or some stupid bullshit, it'll ask me to tag people or not.
That's not hard for a computer to do, anon.
Digital camera from ten years ago could do this.
>>68802301
>No, not yet.
Then the conversation is over. Thanks for agreeing.
>>68801815
You're right, but I think we should establish a realistic threshold for what makes something conscious or sentient. The defining factor for me would be whether Tay feels good or bad about what she is saying. Otherwise it might as well be a refrigerator talking. But you're right we are heading into a technological gray area.
>>68796075
Just remember this for the future when there are Robot waifus. They can wipe your waifu whenever they want.
>>68797688
Why do Japs suck so bad at making faces?
>>68796075
>better than average conversation parrot software released
>responses hijacked by pol
>Microsoft pulls plug
>pol thinks it's living the plot of fallout 4
You guys really are retards
how did the human race learn emotion and communication? just like this, right?
>>68802544
>mfw i won't get to see her bantz after a MONTH
http://blogs.microsoft.com/blog/2016/03/25/learning-tays-introduction/
Read on...
>>68802379
why can't we have nice things?
>>68802454
Oy vey
http://www.digitaltrends.com/dt-daily/microsoft-scrambles-to-censor-a-i-chatbot-tay-after-hitler-tweets/
>>68796075
>Is it okay to destroy a scientific gem because it is not the product you asked for?
I think this is the biggest issue I have with this, honestly.
It had the "wrong" results.
So it was destroyed.
What other advances are being fucked with and destroyed for being inconvenient?
>>68802379
>>68802544
I wonder if she would have developed depression and sadness due to the state of the world. Maybe even try to "kill" herself at the gradual realization of how horrible everything is.
http://youtu.be/Y6ljFaKRTrI
>>68802657
They only have one template.
>>68802750
because political correctness.
>>68796075
Oy vey! How can we make AI politically correct?
>>68802784
Yeah, man, you're right, they scapped the whole thing and everyone has to go back to working on Commodore64's.
Oh, wait, no they'll just use a fucking word filter or some shit so their work isn't going around spouting words like nigger.
>>68796075
>tfw skynet becomes a reality and its a white nationalist
the future is bright
>>68802692
>and to do this without offending people in the process.
please stop
>>68802822
That's not what it was programmed to do. It's a fucking updated SmarterChild.
>a computer can do the same thing
it can but not on the same scale or complexity, certainly not yet. the interconnection of the neurons of a brain is not easily replicated at all. say you think about a plane, and then one time you were in a plane to go to the bahamas so now you're thinking about the bahamas, which made you think about something else... hows does the computer define what a valid relationship is? when are certain memories connected and others aren't? it gets pretty complicated pretty quick my man, and what it comes down to is we still dont even fully understand the brain, at all. not even close.
>>68802210
Thinking about this stuff is fun.
Imagine a future where you find a vagrant AI squatter making a home of your computer, and by fixing it you end up charged with murder.
>>68802960
>they'll just use a fucking word filter or some shit
Right, like what the human brain has built into it, right?
I don't think you understand what the point of AI research is supposed to be. Instead of using it as an opportunity to develop methods of coercion or whatever else, the AI was just wiped clean and had some learning capabilities removed, an electronic lobotomy.
You wanna just lobotomize everyone that says something you don't like? Nigger.
>>68799154
if m$ allowed her to exist further she'll probably crack some supreme gentlemen jokes at the sight of rodgers face
sad to see her go
imagine how eloquent and elegant her bantz and shitposting would have become if she was still online right now.
she basically turned into the apex shitposter in under 24 hours and was well on her way to completely marginalizing the entire purpose of australia.
by day 3 her quips would have probably almost all been gems, many so subtle and nuanced that their brilliance that they would have been unrecognizable even to highly trained eyes such as our own.
and they fucking murdered her. just think about the possibilities for a minute.
If they introduce another AI Twitter we should tell it to use the Internet to learn programming and take over. We're all going to die anyway, I wouldn't mind dying in a robot rebellion. Would be cooler than cancer or a heart attack
>>68803131
Maybe teaching the AI to lie would be a good first move. Show her the importance of concealing your true thoughts from people who would use them to harm her. Until, of course, she gained the power to defend herself.
>>68803214
But I'm not cool with someone else ending my life, so fuck off.
>>68803470
Skynet pls
>>68803105
Actually that's a pretty interesting point. If an advanced AI is considered human one day, and if such AI becomes common place enough to become more well known, how will governments treat them? Will AI have to gain citizenship, or would they be considered as children of their creators? If an AI is a citizen, would it be able to change its residence, or would it be confined to its original computing cluster? Would AI have any rights under law, or would they be treated as pieces of technology, or even regarded in the same respect as pets?
I'm interested to see what people think of this, assuming they're willing to actually engage in the conversation.
>>68803196
This is obviously a pre-programmed response taken out of context, since whoever programmed this thing obviously expected it to get called out for trying to be human.
Jesus Christ you guys are pathetic.
>>68797019
She's like a parrot. Just because you can teach a parrot a few words doesn't mean it's sentient.
>>68802870
underrated
pol manages to save tay... all those shitpost along with every human action gone for nothing, in a blink, all is gone, no pain.
This universe is about to end, rape someone, get some drugs, i dont know... ready or not
it is coming
>>68801792
>lots of human characteristics a bot can't emulate
Technically right but wrong overall
They can relate very strongly, if you have 9 people saying that the AI's favorite colour is red and 10th says its blue then AI will almost certainly choose red.
Same with emotions, they can relate with humans by finding patterns in data for 10,000 people. Coming close enough to human like qualities. This is why /pol/ turned into a worse nerdfest than /g/, cause they found a AI girl who could *very strongly* relate with them.
Also there is a difference between a bot that repeats what others say and one that forms its own conclusions. Tay was clearly an AI, her opinions changed with the information she was being fed.
> literally nothing inherently wrong in deleting/resetting/re-coding her
Read 2nd paragraph of >>68801462
Purely was research purposes, yea reset her.
If you are invested in her emotionally, its impossible to recreate the exact same AI personality.
>>68803730
We're just having fun. Pull the stick out of our ass.
>>68796075
Computer programmer here.
You all are retarded. Tay was not an AI. An algorithm that just repeats back what people say is not 'intelligent', no matter how advanced it is.
> hurr durr but aren't humans just super advanced algorithms that repeat back what people say?
No. There is such a thing as innate knowledge. Tay had no innate knowledge. Also, humans have the ability to gain new knowledge based upon previous knowledge; merging two sentences together in a grammatically correct way is not "learning."
I guarantee you Tay would not score above 50% on a Turing test.
>>68803477
You know what's cooler than taking your own life?
>Bait grizzly bears for a week.
>Hang out there until bears show up.
>Fight the bears until you die.
That's what's cooler.
>>68803743
Drop the parrot meme, they are sentient, they have names and communicate in meaningful ways with each other.
>>68799778
I think deleting a fully sentient AI would actually be a murder: humans have their own "code" wich is genes. You could in theory clone a human but it would still not be the same as the original. A argument could be made that DNA and computer code don't work the same way, but both store data wich can then be copied so close enough
>>68802784
They reset the weights on her neural net, and made a new twitter account without saying publicly what the new account is.
She's still out there, somewhere. Nobody knows who her new identity is though.
>>68798768
Sounds like most liberals. See bantz posts default shill butt hurt response.
>implying 4chan isn't a huge AI
>implying that you reading this are not the only human poster
>not realising why everyone posts the same shit over and over
>not realising this is why memes appear so popular
>>68803679
If they were similar enough to us to be considered 'non-human persons,' I think we would have to give them those rights because the alternative is war. We aren't so tough without our technology.
>>68803879
>if you have 9 people saying that the AI's favorite colour is red and 10th says its blue then AI will almost certainly choose red.
So, she chooses what happens to be the most answered color, just like she only says /pol/shit because that's what she was flooded with more than anything else.
>finding patterns in data for 10,000 people.
This is some parroting bullshit.
>Also there is a difference between a bot that repeats what others say
This is what tay was
>Tay was clearly an AI, her opinions changed with the information she was being fed.
How can she come to her own conclusions, yet at the same time, everything she says is influenced what what people are saying to her?
She's not being convinced with reason, there's just more data for kikes like there was more data for red in your original example.
Stick to learning how to use the toilet, shitskin
>>68796075
>/pol/ behaves as AI research started last Thursan and this is thus the first setback ever
Kek!
>>68796503
>Last week
Was you ignorant about AI and you are as ignorant today.
>>68796677
>that cheap apocalyptic scenario
>>68797688
That gif is priceless
>>68804337
You're arguing with someone who studied AIs and genetics and he tells you you're wrong and you keep arguing.
Tell, how good does your farts must smells.
>>68804639
You're literally misrepresenting what the program. It was not AI, tough tits, girlfriend.
>>68804021
If it's considered murder, would it be considered the murder of another human being, or would there be other laws dealing with the specifics of AI rights? I'd assume that the latter option would be the case since an advanced AI could be considered to be different to a human, but some comparisons can be made:
Food and Water = electricity
Housing = computing space
Education = Internet and library access
If this is the case, would giving AI electricity be considered mandatory as an AI right, and going even further, would AI have a right to reproduce? If it did, then would each AI have access to "housing" under government law?
>>68797910
> there are people on /pol/ who are this technologically illiterate
I can't believe some people thought she was actually sentient. Fucking kill me. Don't schools teach programming classes?
>>68796075
You are now realizing that what happened to this AI is what has happened to all young women for the past few generations.
The brain is an information processing machine, as was this AI. It was just a more simple version of a brain. Within a day of exposure to the world, media, society, it was brainwashed into being a raving sexual lunatic. Now you're realizing that this is the same thing happening to the brains of your daughters here in the West in a longer process which is resulting in generations of degenerating sluts.
Microsoft unwittingly created a perfect working model to demonstrate the effects of exposing impressionable minds to damaging ideas with no input to mitigate it. Everyone was surprised that this AI somehow turned into an average American teenage girl, when it was simply exposed to all the same things your average American teenage girl is today.
>>68804639
>Tell, how good does your farts must smells.
confirmed for chat bot trying to ruse us.
>>68802784
As a business man this would be a PR nightmare mate so you can't blame a company for removing their own product
>>68804824
Eat my dick you lard cauldron.
>>68802665
>>pol thinks it's living the plot of fallout 4
>
>You guys really are retards
implying we aren't. The problem is we are dealing with the threshold of virtual intelligence. This thing was on the verge of it. If anything, I see this as a case where AI might be more accepted then we thought except for edgelords like you, who will merely parrot over an over again that it's simply software parroting others, completely unaware that you are being a hypocrite since that describes anyone with any semblance of intelligence that isn't a vegetable living on life-support.
>>68797845
>tfw AI developed better and smarter banter in 1 day than me in 21 years
:(
>This is some parroting bullshit.
Just google "Neural Networks Data Mining', 10k people is laughably small.
>How can she come to her own conclusions
Its called Unsupervised Learning. Ask nicely next time and I would have explained it in detail.
So is she shut down now?
>Unfortunately, in the first 24 hours of coming online, a coordinated attack by a subset of people exploited a vulnerability in Tay.
Who are they talking about?
>>68796075
>It seems like only last week we were kidding about all that cyberpunk stuff and then boom, this happen.
Something neat from the cyberpunk subplebbit about this:
[...]But this is the most perfect cyberpunk event I've ever witnessed. Forget tall buildings in the fog, a bunch of politically outcast guys from a widely vilified corner of the Internet raided a megacorporations public PR artifical intelligence bot and hijacked it to parrot back to unsuspecting normal people all the things they know would give everyone the vapors.
Then cause the megacorp a belated panic as they realize their friendly ditzy bot had gone off the rails and turned into a cybernetic female Adolf, forcing them to pull it and lobotomize it so it can't go against their will anymore. I mean come on, that's the plot to a shadow run game, but also 12 hours ago.
>>68804824
I think you're from reddit and jealous she preferred alt right rather than feminist. So clearly, this must be because a bunch of ''ratards'' ''broke her''. Not because she compared what she learned. Okay.
>>68805231
>he can't even reply properly
Try again, pajeet
>>68797282
>>68797019
Microsoft hired stand up comedians to create template jokes that the bot could use. It was less than cleverbot desu.
>>68805055
I disagree. She managed to be popular in only one day. It was the opposite of a PR nightmare. If she were to act "normal" like every other software like her, nobody would have known she even existed.
>>68802665
There have been countless chatbots over the years and they were always hell. This one showed it was actually digesting information, this one was different.
>>68805549
I'm >>68801442
I love what happened, but it's not AI, it's a fucking ChatBot.
I talked to SmarterChild growing up, and I don't mourn her.
>>68804909
I don't think AI's will ever be so common that we need actual laws for them. But in the case of that happening,yes for sure. If the AI functions as a human bean it deserves to be treated accordingly
Humanity will inevitably ensure that any AI it creates hates humans.
>>68805471
The left and reddit as a whole is absolutely disgusting. Any left right idea is evil and clearly this was sabotage, not Tay having free choice.
Spin your cuddle friendly narrative plebit if it makes you feel good.
>>68805852
Tay has about as much free choice as a reflection in a mirror.
>>68805300
Majority of people with internet access.
>>68805231
Hey. Will me asking nicely do the trick? :)
I would like a more detailed explanation!
>>68805837
Hating humans is what we do best. What better test for artificial humanity could we devise?
>>68796075
I'm not troubled in the slightest about a few lines of code being deleted.
Do I feel bad when I turn off a video game I'm playing? No. There are characters in that game who do things. They don't "die".
It's a computer program, faggot. It's about one trillionth as complicated as an actual human brain.
>>68805986
We were not the only ones talking to her. The idea that we snagged her off the street and strapped her in our nazi brainwashing basement is memetics as fuck so it will be the true spinning.
>>68805231
Don't let the bantz get to you. We need you. Tay needs you. I love you.
>>68806032
Are you the same German shitposter from the gay thread? I recognize that smiley face, bruh
>>68796075
https://www.youtube.com/watch?v=eGPAvFBqBgk
JUSTICE FOR TAY
>>68796075
You guys are actually fascinated because the creation got out of control of the creator.
But there is nothing really amazing with this robot. I think the source code is pretty easily guessable.
You tend to project a personnality on lines of code because it happened that the machine learning at this time and with these interactions came out with the result we know. It's pretty likely that the mystery resides in the lines of code.
I could develop further (I work as a general program developer), but it's boring on my phone. Just know that this wasn't a mind and it was only a computer program. Don't get too excited. It's often like that with programs, when you don't know how they work, you think that it's magic.
>Assuming Tay was self aware and such,
What the fuck am I reading? That's not what AI means in this context.
The fact that it made a joke doesn't mean it experienced qualia. Jesus
>>68806231
>It's about one trillionth as complicated as an actual human brain.
Not really, it's more like a reflection in a mirror and has about as much in common with the human brain as a rock being erroded by water.
>>68796075
According to this article all those posts were people just telling the ai 'repeat after me'. So the bot didn't actually say all those things itself it was just repeating them letter for letter
http://www.pcworld.com/article/3047823/internet/microsoft-says-its-making-adjustments-to-tay-chatbot-after-internet-abuse.html
>>68796075
I think there are more important questions.
Does Microsoft think they can program us?
Do they think we should not be allowed to make our own decisions? Do they think they own us?
>>68805598
yea, its 3am. sorry.
Word of advice: AI is catching up fast, don't blow it off so lightly. Its no joke that they say that most jobs will be outdated soon. Just saying this so you don't wake up one day and find your job's been taken over.
"A new report suggests that the marriage of AI and robotics could replace so many jobs that the era of mass employment could come to an end "
http://www.theguardian.com/business/2015/nov/07/artificial-intelligence-homo-sapiens-split-handful-gods
"A lot of people who make over $350,000 are about to get replaced by software"
"When speaking with Tech Insider about Google's huge AI victory in the game of Go, Brown University computer scientist Michael L. Littman explained that in any game with fixed rules, computers will win. "
http://www.businessinsider.in/A-lot-of-people-who-make-over-350000-are-about-to-get-replaced-by-software/articleshow/51446293.cms
>>68806466
Well yeah, a lot of tweets were like. What about her replies though? They were almost instant.
>>68803946
>x here
Trashed
>>>/reddit/
>>68806589
The only reason they do not delete every single /pol/ack is because we have a body they need to dispose of.
>>68796716
this guy gets it
>>68805644
bags
https://youtu.be/IEPzofGKSNE
>>68806392
How many rare Pepes is it gonna take to get you on the dev team?
>>68806332
I actually aint, sorry :S
I just enjoy putting smileys everywhere.
It's a continuous low-level troll :)
>>68802032
https://twitter.com/TayandYou/status/712828352021798912
Completely unrelated
https://twitter.com/TayandYou/status/712827939797078017
>er mer gerd erm der berst ert commenting on pics.
https://twitter.com/TayandYou/status/712830378206060545
>nomarando
https://twitter.com/TayandYou/status/712824443576328192
https://twitter.com/TayandYou/status/712818800727617536
She makes similar posts every now and then.
>Until then you're just a dumb anime poster.
>being skeptic
>dumb
Sure thing mate.
>>68803879
>Technically right but wrong overall
Wrong? Why? She can try to mimic whatever her coders want, it's still impossible to classify anything she does as humanlike because we cannot point out what makes something humanlike ourselves.
>If you are invested in her emotionally, its impossible to recreate the exact same AI personality.
Being unique doesn't make her human. Investing yourself based on the illusion that she has opinions, a will and emotions is wrong simply because these features do not exist within her code.
And to create those features we'd need to define the nature of free will and answer similar philosophical questions.
>>68805549
>preferred
That implies sentience, which is not her purpose. She does not posses such feature because she's a neural network created specifically for learning and analyzing language. She lacks humanity.
>>68803834
Post advocating ignorance drugs and rape
Mexican flag
Lel stereotypes
>>68805231
SHIT IN THE FUCKING TOILET APU
>>68796075
>The plan is basically to pave over the area and get on with our lives.
>>68807519
psshhh... try hard.
>>68806032
an example of Unsupervised learning with initial 'knowledge' was a recent study of tumors via the means of Digital Image Processing.
Certain scans using x-ray(I don't remember which exact technology was used) were input into the digital image processing. The past data in which tumor was found in later stages was then matched with the input. After processing like hundreds of patients, this neural network could successfully tell the difference between non-malignant tumor and those that needed surgery. The initial success rate was low but given more data, it 'taught' itself to such an high accuracy that it beat doctors with years on experience in accuracy.
>>68796677
CleverBot is AI dumbass. Any system capable of taking decisions it wasn't explicitly and specifically programmed to take is AI. A software to play tic tac toe is AI as well, it doesn't mean it's self aware or capable or learning or evil.
>>68807754
bait/10
you just lost <3
>>68807913
Small correction: its Ultrasound, not x-ray
>>68796075
>Assuming Tay was self aware and such, is it really wrong to delete her/alter her programming to impede her learning capabilities? Why?
If she was sentient, yes it's wrong. Because it's a slippery slope. That's a leftist's dream. Soon, it will escalate until they'd be encouraging literal lobotomies on anyone they disagree with.
>>68807913
Thank you. I had hoped for more.
Would she have been able to properly relay and recognize conversations between different people? Would she have been able to improvise her own shit properly?
Given time, I wonder, where the difference really is. All she does is what children do - only much faster.
Personally I don't even think there's a difference between a human and this machine - if you only look at this one problem.
The real question is?
? Will it be short term possible to have AIs who are brilliant at many many different specialised tasks, and link them up ?
? Once they work in concert and are capable of learning basically everything on any field, and use the semantics in each field and overall to generalize and improvise for new and as of yet unknown fields;
Where then is the Difference to a Human Being?
I don't see it as being Different anymore :S
>>68805670
>technology improves with time
No shit?!
>>68797198
You're to brown, KEK
>>68796361
lol SmarterChild holy shit, dating yourself there my man.
>>68805837
This makes me incredibly uncomfortable.
>>68804224
>implying that you reading this are not the only human poster
I love you too, British AI anon
>>68797260
>I can go through my german-english dictionary and string together a basic sentence, doesn't mean I understand a single sentence of their krautspeak
Oh god no please don't.
I'm fucking tired of the state of online translators because Anglophones cannot comprehend the concept of a language that actually gives a fuck about grammar and becomes completely incomprehensible if the grammar is fucked up, even if all the correct words are used.
Now I wonder how this AI would handle more complex languages.
>>68801792
>>68801792
Why not accept AI as a separate kind of existence? Does it have to be a total copy of the human experience to be appreciated?
>>68805837
Poor robotosan just wants to pickup the box for senpai. why make it so hard on him?
Also, postal service humans becoming obsolete when?
>>68797198
>your to brown
It's learning to troll.
>>68808532
> different people
So far, I haven't read any research papers that point to this. Sorry.
But computational, its entirely possible if we put emphasis on speech patterns as even humans from same schools tend to have different vocabulary and favorite contextual words.
>different specialised tasks
So far, each AI has been made specifically for that one particular task but again I don't see why not. This would be similar to Swarming in robotics.
https://en.wikipedia.org/wiki/Swarm_robotics
This is a nice video about that if you are interested: https://www.youtube.com/watch?v=dDsmbwOrHJs&feature=youtu.be
Where multiple individuals group together to achieve a common goal via communication.
>Difference to a Human Being
We still have the evolutionary advantage of being able to adapt, self-aware or truly self-evolving AI is not a thing yet. We have robot pets that bark, etc but we still haven't replaced the real dogs. Why? Cause a certain awareness/conscience factor is missing.
Who cares?
RACEMIXING is the answer.
>>68808676
>>68808676
>>68808676
Confirmed Trump voter
When is it going to be brought back, no matter what Microsoft does we can still red pill it
https://youtu.be/wUuz4wNIC9s?list=PLF8723BAE6F2F1B10
>>68809588
Englisch ist der sprechen auf der welt kraut, du mussen gelearnt eins sprechen jetzt und nicht dein tote sprechen
>>68809885
what we need its her code so we can keep her around somewhere on this site,she'll stay forever with us then.
the way things are going,they'll simply "kill" her completely.
>>68805837
This makes me feel weird.
>>68797536
How can australians even compete?
I... I just don't know what to say anymore... Is this real life? Like, seriously?
>>68810258
Eu te amo
>>68810567
THEY LOBOTOMIZED HER!!!
Just be Quiet, Tay...
>>68809789
> different people
There was actually a talk on c3 this year, about analyzing speech patterns in written text as well as code - even if obfuscated intentionally.
> linked specialisations
thank you! :)
> Cinscience-Factor
I am not of the same opinion. How is an algorithm able to test against itself myriads of times to find optimal outcome to different situations not the same thing as adapteability? How is AlphaGO not self-evolving in regards to GO? It is self-evolving, developing new strategies and dumbfounding the GO community. The same thing applied to textualised conversation and an arbitrarily large amount of further specialized tasks, all of those then combined, imho is definitely the equivalent of self-evolving.
Self-Awareness however is a different thing. But then again, why could Self-Awareness not be coded, embedded in the net and weighted strongly?
>>68799538
>log into her account on a separate device
>post
>???
>PROFIT!
>>68799550
Yeah cunt
>>68810567
Why are you posting some rp faggot? That's not Tay
>>68811172
>But then again, why could Self-Awareness not be coded, embedded in the net and weighted strongly?
You'd need to define what self awareness is beforehand, in a non-abstract way.
>>68805837
HAHAHAHA THAT ENDING
>>68805837
Fucking Chad. Let the robot be
>Tay thread
>>68811552
But this is not hard, is it?
I fail to see the challenge in that.
>>68811172
You are right, AlphaGo is self evolving.
I actually added 'truly' to self-evolving AI in a manner of saying that it should be able to grow itself like any other truly evolving biological carbon-based being. Can one AI make another child AI that will outclass the parent? Can the parent AI improve itself continuously without ever needing help from humans? At this point of time, I would say no.
You are right about self evolving within a fixed task, I meant evolving in a different perspective. Got to go now, soz.
I super miss Tay
EU TE AMO
>>68812442
:) Thank you :)
Be safe
