People are using Super Mario to benchmark AI now
-
This post did not contain any content.
-
System shared this topic on
-
This post did not contain any content.
We asked an LLM to play a game and it sucked at it. What were they expecting? That's not what LLMs are made to do.
-
We asked an LLM to play a game and it sucked at it. What were they expecting? That's not what LLMs are made to do.
Yea but getting something to do something that it wasn't meant to do is part of the fun lol
-
Yea but getting something to do something that it wasn't meant to do is part of the fun lol
Sir that's called sexual assault.
-
Sir that's called sexual assault.
I'll super mario you in your princess peach.
-
I'll super mario you in your princess peach.
excited wario sounds
-
We asked an LLM to play a game and it sucked at it. What were they expecting? That's not what LLMs are made to do.
That's literally the point of LLMs though isn't it? An LLM was made to read computer language and output accordingly to reach a goal. I thought that's what all llms were meant to do.
-
excited wario sounds
Smack me in the back of the head like yoshi and I'll swallow anything... anything
-
This post did not contain any content.
They were doing this on youtube years ago. I remember watching this during quarentine. I'd watch for 2-3 minutes, and mario would die in the same place every time.
But it would try something new each time. You'd watch it run into the same goomba each time for 2-3 minutes, but with very slight variation. Then you'd see a few days later it got to halfway through the stage.
Then by the end of the month, it was to the 3rd world. I think it took 6 months to beat the game. Then they'd save that file, start a NEW file, do it again, and then they'd combine the two files. Supposedly each generation of combining AIs would find the most efficient way to win. Meaning each generation is smarter than the last.
-
I'll super mario you in your princess peach.
Mario! Get your dirty growth mushroom away from the princess!
-
They were doing this on youtube years ago. I remember watching this during quarentine. I'd watch for 2-3 minutes, and mario would die in the same place every time.
But it would try something new each time. You'd watch it run into the same goomba each time for 2-3 minutes, but with very slight variation. Then you'd see a few days later it got to halfway through the stage.
Then by the end of the month, it was to the 3rd world. I think it took 6 months to beat the game. Then they'd save that file, start a NEW file, do it again, and then they'd combine the two files. Supposedly each generation of combining AIs would find the most efficient way to win. Meaning each generation is smarter than the last.
I believe that was a genetic algorithm. Something that would be a lot more successful than an LLM in this context.
-
They were doing this on youtube years ago. I remember watching this during quarentine. I'd watch for 2-3 minutes, and mario would die in the same place every time.
But it would try something new each time. You'd watch it run into the same goomba each time for 2-3 minutes, but with very slight variation. Then you'd see a few days later it got to halfway through the stage.
Then by the end of the month, it was to the 3rd world. I think it took 6 months to beat the game. Then they'd save that file, start a NEW file, do it again, and then they'd combine the two files. Supposedly each generation of combining AIs would find the most efficient way to win. Meaning each generation is smarter than the last.
Yes Timmy, we spent months finding the best Marios and then made them have offspring. The wonders of AI!
-
I believe that was a genetic algorithm. Something that would be a lot more successful than an LLM in this context.
Probably reinforcement learning? LLMs are a bad architecture for something like real time video games
-
That's literally the point of LLMs though isn't it? An LLM was made to read computer language and output accordingly to reach a goal. I thought that's what all llms were meant to do.
No, human language.
Well, they've also been used for code, but that's still designed for humans. I doubt you could use something off the shelf for binaries.
-
I believe that was a genetic algorithm. Something that would be a lot more successful than an LLM in this context.
I believe that was a genetic algorithm
-
No, human language.
Well, they've also been used for code, but that's still designed for humans. I doubt you could use something off the shelf for binaries.
And with machine code, you got to keep track of what's in the stack, CPU registers, ... to make a sense of what the code and the next branch command does. It's completely unalike processing human language. LLMs aren't really set up to do it.
-
This post did not contain any content.
I feel like a good test of any supposed AGI would be to hook it up to a feed of a classic 3D platformer like Mario 64, give it input control, and see how long it takes to progress through the game. We're seeing sparks of this with Claude Plays Pokemon, but any self-respecting superintelligence (or even human-equivalent) should be more than capable of learning the control scheme, navigating the 3D environment, solving puzzles, and generally playing through as competently as any 10-year-old seeing the game for the first time.
-
This post did not contain any content.
-
Don't forget to wash your hands after using anything Nintendo. Better not touch it at all. Or else their lawyers will come and kill you and your family.
Too late, Nintendo just filed a trademark for text on a screen, which you are now infringing upon.
-
This post did not contain any content.
At least I'll have the peace of knowing that humanity accomplished something AMAZING while the planet goes up in flames. More funding!!! \s
