864 HN readers' books, ranked by consensus

Every year, the Ask HN "What are you reading?" thread collects hundreds of replies from working engineers, founders, and researchers. This year it drew 864 comments and 451 points, and the pattern is always the same: each commenter hand-types a book title, an author, and a one-line reason it mattered to them. That is an enormous amount of human-curated recommendation data — arguably better than any algorithmic feed, because every entry was written by someone who actually finished the book.

The problem is that Hacker News is a discussion forum, not a library. Comments are ordered by thread structure and votes, not by how often a book was recommended. There is no way to filter fiction from technical manuals, no way to see that a single title was independently praised by seven different people, and no way to find the thread again once it scrolls off the front page. Within roughly 48 hours, one of the richest reading datasets on the internet sinks into the archive, and the following year someone scrapes it all over again from scratch.

Meanwhile, the mainstream alternatives solve a different problem. Goodreads optimizes for social features and advertising, and its recommendations are algorithmic rather than consensus-driven. Curated newsletters are paywalled or single-editor. What is missing is a simple, permanent, filterable page that answers one question honestly: what did a large group of thoughtful technical readers agree is worth reading? That is exactly what this project builds, and the timing is right because the source data is fresh, free, and already annotated by hand.