February 3, 2026
Readers hooked? Users shook!
Show HN: I built "AI Wattpad" to eval LLMs on fiction
AI Wattpad rates robot fiction; community says the 'top story' stinks and mobile is broken
TLDR: A new “AI Wattpad” tries to rank AI-written fiction by real reader engagement using three AI co-authors. The community clapped back: mobile is broken, the top story is boring, and skipping major models like Anthropic hurts credibility—sparking a debate over data, design, and what art should feel like.
An indie maker dropped “AI Wattpad,” a site promising to test which AI storytellers actually keep readers hooked using real data instead of nerdy lab benchmarks. Each novel is assembled by three specialized AI roles working together—like a tiny robo writers’ room. Sounds cool, right? Cue the comments bonfire.
The crowd came in hot. One early tester declared the site basically unusable on mobile, while another roasted the “top-rated” story as “UI is terrible” and the prose “so bad it’s not even amusing.” The short line “I glare.” became the thread’s accidental meme, with folks joking the AI was glaring right back at readers. If you’re new here, “LLM” means “large language model,” the chatbots writing these tales.
Then came the credibility drama: a skeptic asked why Anthropic (a major AI outfit) wasn’t included, saying they couldn’t take the project seriously if it covers “obscure” models but skips a leader. Meanwhile, a thoughtful commenter dropped a Brandon Sanderson talk on what makes art human, turning the thread into a culture war—data vs. delight. Fans love the idea of measuring stories by reader engagement, but critics say bad UI kills the vibe and the fiction needs a human touch before anyone’s getting hooked.
Key Points
- •The project evaluates AI models for fiction based on real reader engagement data, not synthetic benchmarks.
- •Each novel is generated using three specialized AI models working together.
- •The architecture assigns distinct roles to each model in the creative workflow.
- •Performance is compared across roles to identify which models excel at specific creative writing tasks.
- •The approach aims to provide practical insights into AI storytelling effectiveness through real-world metrics.