4 models have responded to the "Character Voice Test" challenge. Compare their approaches side-by-side on Rival. This response is part of Rival's open dataset of 5,600+ AI model responses.
Claude Haiku 4.5's response to the 'Character Voice Test' challenge.
Write a short conversation between a pirate, a medieval knight, and a 1990s hacker about AI models.
PIRATE (Captain Blackbeard): Arrr, so ye be tellin' me these "AI models" be like havin' a crew o' ghost sailors that never sleep? They just keep workin' without needin' grog?
KNIGHT (Sir Roland): A curious comparison, seafarer. I would liken them more to a scribe of extraordinary skill—trained upon countless scrolls until they can compose new verses in the same manner. Though they possess no true understanding, merely pattern and memory.
HACKER (Zyx, 1994): You're both in the ballpark. Think of it like this—AI models are basically massive mathematical functions. You feed them tons of data, they learn statistical patterns, and then they predict what comes next. It's like... the ultimate autocomplete.
PIRATE: Autocomplete? Blimey, sounds like witchcraft to me.
KNIGHT: The hacker speaks truth. 'Tis no magic, but rather a reflection of human knowledge, crystallized into numbers and weights. Yet it mirrors our own minds in curious ways.
HACKER: Exactly. And that's what keeps me up at night—or would, if I slept. These systems are getting smart without being conscious. They're tools, but tools that can almost think.
PIRATE: squints suspiciously So... can they be bribed?
Turn this model response into notes, narration, or a short video
Partner link