the game was rigged
when two humans at the top level play they've had time to research their opponent. if garry had the chance to observe deep blues moves ahead of the game he would have slaughtered that computer.
>After his loss, Kasparov said that he sometimes saw unusual creativity in the machine's moves, suggesting that during the second game, human chess players had intervened on behalf of the machine. IBM denied this >Kasparov demanded a rematch, but IBM had dismantled Deep Blue after its victory and refused the rematch
Third party chess players have agreed with Kasparov's assessment, though. It doesn't even have to be IBM forming some conspiracy of chess masters behind Gary's back, ten normal-good players together will beat a master.
AlphaGo is even better than Stockfish now. Chess engines are so unbelievably good that they actually are doing things humans find puzzling.
2 years ago
Anonymous
I guess that even though one thing may have been staged back then, It really doesn’t matter these days. It’s over.
2 years ago
Anonymous
Yep. Not only has the software improved massively but obviously the hardware available has too. Computers can evaluate a ridiculous number of positions in comparison to humans.
Starcraft like most RTS games relies on executing the meta at the highest APM (actions per minute) possible, so obviously that's how a Starcraft AI would win. It isn't doing anything special here beyond being able to move the moust instantly and click at several kilohertz.
2 years ago
Anonymous
you didn't read the very short article before commenting on it
when they made their fancy computer account for human bottlenecks like only being able to see one part of the map at a time and only being able to click a mouse so fast, the best human won.
this was against an ai that had beaten the best chess ais, and had beaten every human before its unfair physical advantages were nerfed
they remade the machine and designed it specifically in mind to defeat garry, studying in depth his every game, but then not letting garry study a single game done by the now remade machine. they didnt give garry a chance to better understand his opponent. garry lost. and then they tore the computer apart so there couldn't be a rematch.
>When Kasparov requested that he be allowed to study other games that Deep Blue had played so as to better understand his opponent, IBM refused, leading Kasparov to study many popular PC chess games to familiarize himself with computer gameplay
what's a make work problem I can use for a cluster of processors?
I believe there is a way to calculate portions of pi without iterating from the beginning but what else can I do?
I want to make a super microcontroller.
I've got a ton of various mcus and components I ordered at the beginning of the year because I was worried about the supply chain
>I believe there is a way to calculate portions of pi without iterating from the beginning
There is. Square a random huge number, subtract one, and check if it's a (marsienne) prime.
Literally what mprime does though, so just run that. Another good project is folding@home, they're trying to cure cancer or whatever. There's also SETI@home, if you just want to throw away electricity.
>mprime
I wrote a better prime finder, in C.
#include <stdio.h>
#include <limits.h>
int main () {
unsigned long long int x=0;
while (x<ULLONG_MAX) {
int isprime=1;
for (unsigned long long int i=2;i<(x/2);i++) {
if (x%i == 0) {
isprime=0;
break;
}
}
if (isprime) {
printf("%dn",x);
}
x++;
}
return 0;
}
the most impressive computer ai in the world lost a game of starcraft to a korean, after the computers advantage of being able to click a million times a second and see the entire map at once was nerfed, the human won. as impressive as computers are don't be so quick to assume humans automatically inferior on a level playing field. starcraft is the new chess btw and google and facebook spent a lot of time and money training that ai.
Just a bunch of vidja and games. It’s not really what people use ai or supercomputers for. Consider it a side gig. The big boys (top 500) are focused on solving problems that would be impractical without the help of ML
also as impressive as computers are, they were made by humans. so the real winner is the people who made the ai. so even when the computer wins it still proves humans are impressive.
>oh boo hoo the computer beat us at chess we're so worthless and weak stupid meat creatures
shut the frick up
that's a pipedream, humans aren't just bottlenecked by input speeds. our brains can't process faster then the input speeds anyways. ever tried speed reading? i can read an entire page of a book at a glance, input bottleneck shattered. but i can't comprehend what i read or even remember it unless i slow down the reading speed to the speed of a fast talking human. communication at infinite speeds doesn't matter when the brain can only process at human speed anyways. woop de doo you got a man with a chip in his head that can read a thousand books a minute, he's not going to comprehend any of what he read so its moot. so with that out of the way there's little point to merging man with machine, it's just a gimmick. might as well just have a man with a smartphone since humans can only think about as fast as their natural physical input/output speed anyways.
that being said we've already merged with the computer and continue to do so in meaningful ways. with mouse and keyboard and slapping our glass touchscreens with our greasy fingers. thats about as good as it will ever get because the human brain can't move faster then that anyways. so we're already there, we're already doing it, we're training the ais as we speak. there could be ais learning from this conversation we're having right now. no brainchips needed and not much point to brainchips other then for correcting paralysis and other disabilities.
even if they make a computer that makes another computer that beats humans. it's still a human victory because it will have been the humans that made the computer capable of making a computer that are the real winners.
The 3380 family ranged from about $98-146k. In inflation adjusted dollars, that would put the top of the range (10GB) at around half a million dollars.
probably the one in hackers
Whichever one i’m working on
the game was rigged
when two humans at the top level play they've had time to research their opponent. if garry had the chance to observe deep blues moves ahead of the game he would have slaughtered that computer.
>After his loss, Kasparov said that he sometimes saw unusual creativity in the machine's moves, suggesting that during the second game, human chess players had intervened on behalf of the machine. IBM denied this
>Kasparov demanded a rematch, but IBM had dismantled Deep Blue after its victory and refused the rematch
sad cope from a dumb loser b***h
Third party chess players have agreed with Kasparov's assessment, though. It doesn't even have to be IBM forming some conspiracy of chess masters behind Gary's back, ten normal-good players together will beat a master.
it was a publicity stunt to boost ibms stock price not a legit game
Do you think they could win a rematch on modern software and hardware tho? I guess that may have been staged but it hardly seems to matter now
no, chess computers are light years ahead of humans now.
No human could ever beat a raspi running Stockfish.
AlphaGo is even better than Stockfish now. Chess engines are so unbelievably good that they actually are doing things humans find puzzling.
I guess that even though one thing may have been staged back then, It really doesn’t matter these days. It’s over.
Yep. Not only has the software improved massively but obviously the hardware available has too. Computers can evaluate a ridiculous number of positions in comparison to humans.
Starcraft like most RTS games relies on executing the meta at the highest APM (actions per minute) possible, so obviously that's how a Starcraft AI would win. It isn't doing anything special here beyond being able to move the moust instantly and click at several kilohertz.
you didn't read the very short article before commenting on it
https://www.techspot.com/news/78431-human-player-finally-beat-deepmind-alphastar-ai-starcraft.html
Around when deep blue beat kasparov was the last time humans could conceivably beat computers. The gap is far too large now.
when they made their fancy computer account for human bottlenecks like only being able to see one part of the map at a time and only being able to click a mouse so fast, the best human won.
this was against an ai that had beaten the best chess ais, and had beaten every human before its unfair physical advantages were nerfed
forgot to link this to this
It sounds like this is human cope. Flesh is weak and all that. It’s cool to tune these parameters and see how it measures up though
he won the first match you fricking Black person
and then what happened after that?
they remade the machine and designed it specifically in mind to defeat garry, studying in depth his every game, but then not letting garry study a single game done by the now remade machine. they didnt give garry a chance to better understand his opponent. garry lost. and then they tore the computer apart so there couldn't be a rematch.
Nah Kasparov is a cool guy
he's a butthurt israelite
because the game was rigged, it's been explained multiple times itt, pay attention.
I liked the PlayStation 3 clusters, shame they aren't replaceable now.
really? i keep seeing reballing videos popping up on my feed and people fixing ps3s. there seems to be a renewed interest in running these in homelabs
>When Kasparov requested that he be allowed to study other games that Deep Blue had played so as to better understand his opponent, IBM refused, leading Kasparov to study many popular PC chess games to familiarize himself with computer gameplay
my mind
Thinking Machines CM-2 (Connection Machine)
Its front-end is a Symbolics Lisp machine, and those are also very cool.
Very sci fi
who is this handsome lad
what a beast
what's a make work problem I can use for a cluster of processors?
I believe there is a way to calculate portions of pi without iterating from the beginning but what else can I do?
I want to make a super microcontroller.
I've got a ton of various mcus and components I ordered at the beginning of the year because I was worried about the supply chain
What are you planning to control? Some sort of smart home system?
>I believe there is a way to calculate portions of pi without iterating from the beginning
There is. Square a random huge number, subtract one, and check if it's a (marsienne) prime.
Literally what mprime does though, so just run that. Another good project is folding@home, they're trying to cure cancer or whatever. There's also SETI@home, if you just want to throw away electricity.
SETI@home is over, sadly
Mersenne primes != Pi
>mprime
I wrote a better prime finder, in C.
#include <stdio.h>
#include <limits.h>
int main () {
unsigned long long int x=0;
while (x<ULLONG_MAX) {
int isprime=1;
for (unsigned long long int i=2;i<(x/2);i++) {
if (x%i == 0) {
isprime=0;
break;
}
}
if (isprime) {
printf("%dn",x);
}
x++;
}
return 0;
}
My favorite HPC thing is ephemeral clusters based on Kubernetes. Dask and Argo are fun to play with.
t. does this at work
the most impressive computer ai in the world lost a game of starcraft to a korean, after the computers advantage of being able to click a million times a second and see the entire map at once was nerfed, the human won. as impressive as computers are don't be so quick to assume humans automatically inferior on a level playing field. starcraft is the new chess btw and google and facebook spent a lot of time and money training that ai.
Just a bunch of vidja and games. It’s not really what people use ai or supercomputers for. Consider it a side gig. The big boys (top 500) are focused on solving problems that would be impractical without the help of ML
also as impressive as computers are, they were made by humans. so the real winner is the people who made the ai. so even when the computer wins it still proves humans are impressive.
>oh boo hoo the computer beat us at chess we're so worthless and weak stupid meat creatures
shut the frick up
the real fun begins when we start merging computers and humans, if we don't destroy ourselves first
that's a pipedream, humans aren't just bottlenecked by input speeds. our brains can't process faster then the input speeds anyways. ever tried speed reading? i can read an entire page of a book at a glance, input bottleneck shattered. but i can't comprehend what i read or even remember it unless i slow down the reading speed to the speed of a fast talking human. communication at infinite speeds doesn't matter when the brain can only process at human speed anyways. woop de doo you got a man with a chip in his head that can read a thousand books a minute, he's not going to comprehend any of what he read so its moot. so with that out of the way there's little point to merging man with machine, it's just a gimmick. might as well just have a man with a smartphone since humans can only think about as fast as their natural physical input/output speed anyways.
that being said we've already merged with the computer and continue to do so in meaningful ways. with mouse and keyboard and slapping our glass touchscreens with our greasy fingers. thats about as good as it will ever get because the human brain can't move faster then that anyways. so we're already there, we're already doing it, we're training the ais as we speak. there could be ais learning from this conversation we're having right now. no brainchips needed and not much point to brainchips other then for correcting paralysis and other disabilities.
Devilish.
even if they make a computer that makes another computer that beats humans. it's still a human victory because it will have been the humans that made the computer capable of making a computer that are the real winners.
the gigabyte
wtf this thing looks dope
How much would one of these be worth today?
The 3380 family ranged from about $98-146k. In inflation adjusted dollars, that would put the top of the range (10GB) at around half a million dollars.
the one that looks like a sofa. CRAY?
Yeah, just a big round couch to put in a waiting area (also there's a computer in the middle).
Cray-1
Tie between the T3E and Cray-2.
damn I used to work in a lab that looked just like this. that sterile white, the drop floors, just rows of racks of hardware. bleh.