How is there not a thread about this?
I actually find this quite sad
http://www.theguardian.com/technology/2016/jan/27/google-hits-ai-milestone-as-computer-beats-go-grandmaster
>>53399859
Rekt Humans
There was a thread about this like a month ago.
>jan/27
yep, a month ago.
>>53399905
Humans on suicide watch
>george martin is kill
I actually find this quite sad
>>53399920
It was another game, where it won a pro player (which isn't something AIs have been able to before). This time it beat not just any pro player but the world champion.
>>53399944
He was the 4th worst Beatle so whatever
>>53399944
A different type of sad bby
/sci/ has better discussion
/g/ is essentially /v/'s computer helpdesk board
Ai is supposed to play some chink grandmaster in the next few months, chink is world champ i think
can these beat navy seals yet? then, we're fucked
>>53399859
Because it happened over a month ago
>>53399968
There is no "world champion" in Go, there's just the highest rated player (by some officially recognized ratings organization, think it's the ELO same as in Chess).
Punk bitch resigned in the 1st match yesterday.
NEVER GIVE UP, NEVER SURRENDER
It hasn't beaten the superior Asian Master yet
>>53400390
>not resigning when you can see you're implicitly at the equivalent of a check mate
>>53400078
well any one of us could beat this grandmaster now - just run the algorithm developed by google and you will move exactly as that program does. every. single. time.
granted it will take you a couple of weeks (years?) to go through all the number crunching with pen and paper but in the end you will reach the same move.
>>53400612
isnt it a self learning algorithm thats been playing itself? good luck getting there with pen and paper
>>53400642
just run the result with the reference data from learning already in place
>>53400612
>well any one of us could beat this grandmaster now - just run the algorithm developed by google and you will move exactly as that program does. every. single. time.
I bet you would make mistakes though, especially considering the time constraints of the game.
>>53401072
so do the calculation three times and compare results, anyway any of us could beat that grand master now
>>53399859
There was a live thread about this dipshit.
Today we are having it again
>>53401387
>There was a live thread about this dipshit.
>Today we are having it again
>>53401361
Time constraints.
>>53401361
you could if google gave you the code to it
currently its private
>>53401411
>I dont know how to search the archive
>look at me using meme arrows
It can't beat Komugi anyway
>>53401431
can us humans even hope to comprehend the information it gets from the reinforcement learning?
>>53401442
>>I dont know how to search the archive
>>look at me using meme arrows
What if we put 2 algos against each other?
Like in the 4chen cup.
>>53401443
Nobody can m8.
>>53401469
There was a live thread about this dipshit.
Today we are having it again
>>53401600
hehe epic
>>53401498
I would love to see that actually
I read the paper, very interesting tech, plan on making an superior ai for a card game.
Computers will beat humans at recognizing emotions, and will probably become better at text-flirting with girls, so once that happens we at /g/ will have a fighting chance on tinder.
>>53401796
tfw I can have computers flirt with eachother for the perfect shipping fanfics
