Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

theverge.com

Post reply on HN
Print view View original

31–40 of 47 posts

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#31

Post by pvg »

PeterisP wrote: post #21
pvg wrote: post #18

Earlier quoted context omitted.

The AI isn't finding bugs by somehow examining the game's source code The title doesn't say that.

It kind of implies that - "AI Cheats at Old Atari Games by Finding Unknown Bugs" would be an accurate title, but the extended "AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code " tells that it's actually finding something in the code, as opposed to simply unexpected/emergent behavior.

It's mildly ambiguous (like most things) but it's not misleading or inaccurate. It finds bugs. Which are in the code. It doesn't say what kind of code and it certainly doesn't say it finds them by looking at the source code.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#32

Post by ianferrel »

camgunz wrote: post #20

Can we put AI to work on proving that we live in a simulation? I would never enter/exit my apartment 38 times alternating between forwards, backwards and each side, but an AI would. Maybe then all the walls start flashing and then we'll know!

People with Obsessive Compulsive Disorder are just depth-first-searching for an overlooked maximal strategy.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#33

Post by gkilmain »

maxander wrote: post #29
pbhjpbhj wrote: post #25

Earlier quoted context omitted.

Lol, that Youtube video, at the end the AI pauses the game of Tetris forever so as not to lose.

Another case where “the winning move is not to play.”

I have a friend who I play chess against and every time he's about to lose he offers me a draw. And will continue offering me a draw until I win. Not happy to see the AI performing like that bitch MukyMuky.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#34

Post by soared »

andyjohnson0 wrote: post #14

The title seems misleading to me. The AI isn't finding bugs by somehow examining the game's source code, it's trying random gameplay and exploiting any advantages that emerge. That it's finding previously unknown bugs seems to be almost entirely down to trying things that human players wouldn't think to do.

Did you read the article?

>It’s important to note, though, that the agent is not approaching this problem in the same way that a human would. It’s not actively looking for exploits in the game with some Matrix-like computer-vision.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#35

Post by andyjohnson0 »

soared wrote: post #34
andyjohnson0 wrote: post #14

The title seems misleading to me. The AI isn't finding bugs by somehow examining the game's source code, it's trying random gameplay and exploiting any advantages that emerge. That it's finding previously unknown bugs seems to be almost entirely down to trying things that human players wouldn't think to do.

Did you read the article? >It’s important to note, though, that the agent is not approaching this problem in the same way that a human would. It’s not actively looking for exploits in the game with some Matrix-like computer-vision.

I did. "looking for exploits in the game with some Matrix-like computer-vision." is a fairly meaningless phrase.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#36

Post by notahacker »

gkilmain wrote: post #33
maxander wrote: post #29

Earlier quoted context omitted.

Another case where “the winning move is not to play.”

I have a friend who I play chess against and every time he's about to lose he offers me a draw. And will continue offering me a draw until I win. Not happy to see the AI performing like that bitch MukyMuky.

In repeat games your friend's strategy might actually be counterproductive, assuming he's playing against people that are rational enough to figure out that he always starts offering a draw round about the time he expects to lose, but not so good at chess that they sometimes don't sometimes miss opportunities for a quick checkmate...

I'd be disappointed if an AI chess computer invariably let me know that I had a better path to winning the game than it did.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#37

Post by gkilmain »

notahacker wrote: post #36
gkilmain wrote: post #33

Earlier quoted context omitted.

I have a friend who I play chess against and every time he's about to lose he offers me a draw. And will continue offering me a draw until I win. Not happy to see the AI performing like that bitch MukyMuky.

In repeat games your friend's strategy might actually be counterproductive, assuming he's playing against people that are rational enough to figure out that he always starts offering a draw round about the time he expects to lose, but not so good at chess that they sometimes don't sometimes miss opportunities for a quick checkmate... I'd be disappointed if an AI chess computer invariably let me know that I had a bett…

That is exactly what happened. He played a friend of mine whose much lower rated. Eventually the lower rated player found himself in winning position but when the draw was offered he assumed he must be missing something and accepted it. So it worked once.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#38

Post by aaron695 »

How about - AI not as stupid as humans who believe "AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code"!!!!!

Fuck me, seriously?

"Fuzzing on a platform not in use commonly by humans finds something cool, human intelligence writes a paper about it."

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#39

Post by bcoates »

andyjohnson0 wrote: post #35
soared wrote: post #34

Earlier quoted context omitted.

Did you read the article? >It’s important to note, though, that the agent is not approaching this problem in the same way that a human would. It’s not actively looking for exploits in the game with some Matrix-like computer-vision.

I did. "looking for exploits in the game with some Matrix-like computer-vision." is a fairly meaningless phrase.

Seems like a perfectly obvious metaphor? It's just providing GA-generated input, not doing tooled exploration like afl-fuzz or a symbolic execution analyzer.

It literally-and-metaphorically can't see behind the UI to the internal state of the game.

Re: AI Cheats at Old Atari Games by Finding Unknown Bugs in the Code

#40

Post by BatFastard »

tantalor wrote: post #17
andyjohnson0 wrote: post #14

The title seems misleading to me. The AI isn't finding bugs by somehow examining the game's source code, it's trying random gameplay and exploiting any advantages that emerge. That it's finding previously unknown bugs seems to be almost entirely down to trying things that human players wouldn't think to do.

You confuse bug (unintended behavior) with its cause (bad code).

We called them "Unintended features", and they were usually quite popular with users.
Post reply on HN