Quality non-fiction books are the antithesis of AI slop
Points and comments are a snapshot, not live.
A researcher built a searchable index of 6,500 prize-winning non-fiction books as an antidote to AI-generated content.
The author, a historian, created the Book Prize Index (book-prize-index.vercel.app) using Claude Code and GPT-5.6 to aggregate winners and finalists from major English-language non-fiction prizes. The site features semantic search that allows queries like "classic biographies that are surprisingly weird" and includes data visualizations of publishers' award performance. The author argues that non-fiction quality peaked between the 1980s and early 2000s, attributing this golden age to post-war developments like jet travel, expanded library access, and early digital cataloging systems. The project hosts 6,500 titles, funded entirely by the author.
What commenters are saying
Commenters split sharply. Several praised the project's goal, with one noting that asking LLMs for book recommendations is "much better than a google search" and can surface deep corners of a subject. Others questioned the premise: one pointed out that book prizes are often arbitrary or comically easy to win, citing the NCR Book Award scandal where judges didn't read the books. Another argued that using AI (Claude Code) to build a tool about quality non-fiction undermines the project's intention. A third commenter observed that AI could be used not as a replacement but as a guide to physical books, calling LLMs "almost purpose built" for cross-referencing print collections.