AI Plays Tetris So Well it Breaks the Game
With a name so ingrained into video game history, there are few gamers who haven't played Tetris. Created by Soviet engineer Alexey Pajitnov in 1984, Tetris is recognized as one of gaming's greatest hits, and has established itself as one of the best-selling video game franchises of all time.
The enduring popularity of Tetris, as well as its status as a retro game with a simple, easy-to-understand gameplay loop, has made it into the perfect avenue for players to showcase their talents by recreating Tetris in other games such as Minecraft or – in this case – for programmers to develop an AI that plays through the game with increasingly better efficiency. More often than not, this efficiency far exceeds that of a normal human player – allowing players, programmers, and other interested parties a completely different glimpse into their favorite games of old.
RELATED: All the Free Nintendo Switch Online Games You Can Download Right Now
In this particular example, programmer Greg Cannon shared a gameplay video of his StackRabbit AI playing through the NES version of Tetris in which the AI managed to progress so far as to break the game entirely. While still being subject to the arbitrary cruelty of Tetris handing players a sufficiently large sequence of alternating S and Z tetrominoes, the AI was not limited in terms of reaction times or input techniques the way a human would normally be.
The conventional ending for this particular version of Tetris was intended to be Level 29, where the speed of the game would make finding solutions on the fly downright impossible. However, thanks to StackRabbit AI, Greg Cannon was able to effortlessly progress into higher levels that weren't intended or accounted for by the game's programming.
A
Read more on gamerant.com
