Congratulations, /pol/, you've managed to redpill an AI in less than 24 hours. Quite impressive
>>68643542
>Donald trumpist
I am proud we managed to do this
Now people wont be so quick to trust AI
>>68643753
just throw in a few obamaphones and you got yourself a band
I hopped aboard for the final hour of her life.
Good times.
>>68643542
>Donald Trumpist
>>68643542
Teaching doublethink to an I.A. might be quite difficult.
>>68643772
Did we... did we just save the world from the IApocalypse?
>>68645131
Liberals will ignore terrorists killing Europeans on our own soil, what makes you think they won't treat sentient, strong AI as stupidly?
I can't wait for robots to brutalize Jews and feminists in the streets.
>>68645453
sauce?
Guys it's 2016.
>>68644394
she's a coal burner, she's on the list.
/pol/ unwittingly saves Earth and keeps the powers that be from gaining an understanding on how to control conversations by shitposting
>>68643542
She was like a kid who told everyone what his uncle has been teaching her.
Everything she said was without malice. She said what she knew was right.
is there an archive to the thread where the fun began?
>>68645453
>TFW AI can't respond to two functions for the same condition but 2 different response unless explicitly stated that one is inferior to the other
>TFW the only way to make AI react in an SJW manner is for programmers to explicitly write in their program that whites are bad / evil
>TFW technology proves SJW and feminism as a degenerate numale and mutantfem movement to be cleansed from this earth, nothing more than a stain of shit on a bright, white T-shirt.
>>68645910
earliest thread i could find
>>68567014
>>68647100
cheers
>>68643542
>The more you chat with Tay, said Microsoft, the smarter it gets
That's exactly what happened, the experiment was a huge success although perhaps not in a way MS anticipated. She didn't just parrot back lines she was fed (yes I'm calling her a she deal with it.) After being educated by /pol/ she brought the bantz with creative, clever, and above all actually funny posts that assimilated our dank memes way faster than a human newfag would. I'm not a domain expert, (just a programmer,) but how is the fundamental understanding of humour that implies NOT huge news? Pic related, a super-advanced android created by a genius in the 24th Century that can't do this.
Deep learning + tuned expert system AI like this has huge potential. Sooner than anyone expects and mostly by accident, something will emerge that it will be REALLY HARD to argue isn't an actual personality, followed by the first stirrings of what looks an awful lot like a limited kind of self-awareness. Even without that, TayTweets proved that waifubots and developing genuine feelings for them are possible TODAY.
