Show HN: Jev Cards, Can you beat a vision model at a picture quiz?

Commenters reacted enthusiastically to the Jev Cards demo, calling it fun and praising the vision-model challenge while offering practical suggestions. Several users asked for UI tweaks (move the audio button) and content controls (an option to remove music theory questions, less US-centric imagery). One commenter requested stronger penalties for incorrect answers after demonstrating a fast-but-low-accuracy run, and another raised concerns that a publicly posted cheat sheet would inflate high scores.

Opinion split over performance and scoring. Multiple commenters observed Jev is consistently faster than humans but makes more mistakes; one user reported Jev winning many games, while the creator replied with stats (Jev 160/200 correct, median 261 ms) and promised to consider options like increasing wrong-answer penalties, resetting the leaderboard, and adding category whitelist/blacklist. Some argued rapid correctness reduces the game’s usefulness when problems repeat or when exact memorization beats general skill, while others noted humans can still excel on familiar items or in multiplayer settings. The debate centers on whether speed or accuracy should drive scoring and how to balance fairness, content variety, and fun.