We might have a chance against AI #shorts

First, the machines came for other machines that are really good at Go. Who will they come for next?

FOLLOW US ELSEWHERE
---------------------------------------------------

[haydenhayden]

Ah yes, the ultimate weapon against any opponent: *gaslighting*

[zoran]

And then AI learns that rare strategy and you no longer stand a chance.

[brentsnocomgaming]

It's like when a DOTA team figured out how to beat openAI by playing solely with meme strats.

[grinchwitadumpy]

“If you have no fucking clue what you’re doing, neither does the enemy.”

-Someone, sometime, somewhere, somehow, somewhat.

[jayjam_]

"years later"

more like 20 minutes later when you've started up the next game and the ai has added your trick to its database.

[Jamison_parker]

“The game has been dominated by AI players for 1000’s of years”

“Bro the Roman’s had AI💀

[rikuandsomeonestoleit]

"Which has been dominated by ai for thousands of years."
Bruh, my 90 year old grandma barely understands what an ai is.

[Owenzzz]

But… isn’t this adversary another AI?🤔

[noahpoobbailey]

This is the scenario in chess as well. Some computers are not the best at evaluating positions with lots going on, especially with gridlocked pawns. In some cases, you see top grandmasters finding a specific move that computers don't even see.

[ErRr_thunder]

the perfect strategy:”if I don’t know what I’m doing, they can’t know what I’m doing”

[Fudge_]

A SHORT? NEVER BEFORE HAS THIS BEEN SEEN
(for like the last year)

[anakinskywalkercrappyprodu]

"The enemy cant understand what you are doing if you dont understand what you are doing"
-Sun Zues

[TomatoChan]

"If you do not know what you are doing, neither does your enemy" - Joe Tzu

[kingwes]

I like that chess engines are unique in the way that they calculate live instead of learning the best way to play when encountering a certain position.

[graysonofthewood]

But… won’t the AI learn how to handle these rare situations when they actually experience them more frequently?? All the AI creators have to do is present a few hundred scenarios with these situations and as long as it hasn’t passed into the realm of overfitting, it should learn how to deal with them.

[rusticgiraffe]

Well when the Terminator arrives to kill me to death, I’ll be sure to challenge it to a board game.

[Real]

But now you've taught the AI the secrets!!!

[meejinhuang]

Once AI learns how to replicate itself and spread on the internet, humanity is in trouble.

[cherubinth]

adversarial examples are such a fun research area in AI

[Yarumasi]

It brings to mind the Fresh Prince of Bel-Air episode where Will was getting hustled at Pool (billiards) and so Phil came in and hustled THEM back using the same tricks to appear noobish.