Concept B: Book 2.0 as a generative and unique visual reading experience

[truedat101]

This is Yoda-san's concept. The idea is to prompt in text from a book either through OCR or QR code reading of remote text, and kick off a pipeline of text encoding and generative AI to generate assets.

I imagine even a sketchy or hazy version of the gist of a story being generated with each page turn is incredibly engaging.

As a general rule in english, you take the first and last sentence of a paragraph to gather the meaning of a paragraph. Or, we ask the LLM to take the paragraph or a page (context length is an issue... also speed) and give us a summary. From there we need a pipeline

• 3D asset gen
• 2d asset gen
• audio asset gen
• a textual headline for the scene