March 6, 2026
Pew pew meets meow meltdown
Show HN: 1v1 coding game that LLMs struggle with
AI flops in cat-blasting duel as commenters cry “unfair” and LOL
TLDR: A simple 1-on-1 coding game where you blast enemy cats made name-brand AI stumble. Commenters loved the chaos but argued it was unfair since humans can watch and iterate while AIs can’t simulate, turning it into a lively “humans vs robots” moment about AI’s limits in reactive strategy.
An indie 1v1 coding game—move(), pew(), and a screen full of cartoon cats—just handed big-name AI a hairball. The dev pitted large language models (LLMs), which are text-predicting AIs, against a human-made “Clowder” bot, and the machines whiffed. Cue the comment section fireworks. One camp is poking the bear: are these supposedly “powerful” AIs just bad at strategy? Another camp says hold up—this test seems stacked. As one commenter put it, without a way to watch and simulate games, the AIs are fighting blind while humans iterate and adjust.
There’s also some user-experience side-eye: “Took me a while to understand the home page,” admitted one early fan. Meanwhile, resident historian dang “macroexpanded” the drama with throwbacks to similar programmable strategy games like Yare 2 and Yare.io, reminding everyone this cat fight has a long lineage. The meme energy? Off the charts. Think: “pew pew vs. meow meow,” “AI allergic to cats,” and a scoreboard reading Humans 1, Robots 0. Underneath the jokes, the real debate claws at a big question: does this show the limits of today’s AI in fast-moving, reactive tasks—or did we just bench the bots with house rules? Either way, the crowd is entertained and the cats are undefeated.
Key Points
- •Objective: eliminate all enemy cats to win.
- •Players interact via functions move() (movement) and pew() (attack).
- •The game exposes all relevant game data for decision-making.
- •Instructions are intentionally minimal, emphasizing player-devised strategies.
- •Success depends on using move(), pew(), and game data effectively.