← back
Books reimagined: AI to create new experiences for things you know — Lukasz Gandecki, TheBrain.pro
Takeaway
Pair multiple AI primitives behind a polished UX and hide the AI itself—books with scene-matched music, avatars, and voice Q&A become a new medium.
Summary
- BookGenius.net layers AI on classic books: character avatars, scene-matched music, voice Q&A with push-to-talk (~100ms response), embedding-based scene search, and deep-research over read-so-far context.
- Production pipeline: LLM scene analysis + mood detection drives music generation, plus structured XML metadata mapping characters to scenes—human-curated after the AI draft.
- Lukasz advocates 'hide the AI from the user': human touch judges output quality the model can't (music vibe, avatar match).
- Argues vibe-coding makes throwing away code feel great; refactor by (badness × pain × ease) only, and open-sources the player so anyone can build Netflix-style book experiences.
multimodalproductbooks
Original description
[last round of Attendee-Led 10min lightning talks] I will showcase how I got tired of waiting for an AI assisted/no spoiler book reading experience and built my own. Check 30s video at https://youtu.be/JjwnYqy668M or go to demo book at https//bookgenius.net Open Sourcing! contact: https://x.com/lgandecki Recorded at the AI Engineer World's Fair in San Francisco. Stay up to date on our upcoming events and content by joining our newsletter here: https://www.ai.engineer/newsletter