How this works

Hacker News has been asking itself “what should I read?” since 2007. The answers are scattered across hundreds of threads and tens of thousands of comments. This site reads all of them and counts.

Method

Threads are found with the HN Algolia search API: every story whose title mentions books, filtered to recommendation threads (Ask HN, best-of-year lists, reading lists) and away from news about publishing. That is 443 threads and 48,845 comments as of 2026-09-19. Each comment is read by a language model that lists the books it names. Titles are then normalised (“SICP” and “Structure and Interpretation of Computer Programs” are one book) and counted.

The rank is the number of distinct commenters who recommended the book. A thread where one person mentions a book five times counts once. Books need at least 3 recommenders to get a page. “Hot” is the same count restricted to threads from the last 30, 90 or 365 days.

The quotes

Every quote is a verbatim excerpt of a real comment, checked by machine against the original text, and links to the comment on Hacker News. Nothing is paraphrased or written by us. If one of them is yours and you would like it removed, email team@levelbrook.com.

Money

Links to Amazon carry an affiliate tag. As an Amazon Associate this site earns from qualifying purchases; the price you pay is unchanged. Ratings shown next to the buy button are Amazon's own, read at build time. That is the whole business model.

Privacy

Pages load a small first-party script that records the visit (page, referrer, browser, rough location from the IP). No cookies, no third-party analytics, nothing sold.

This site is not affiliated with Y Combinator or Hacker News. It is a side project of Levelbrook.