Hacker News through time: AI evaluates decade-old discussions

A developer has created a project that leverages modern language models to retrospectively analyze Hacker News discussions from December 2015, assessing how accurately the observations made a decade ago have withstood the test of time. The project, named "hn-time-capsule" and published on GitHub, processes 31 days of front-page discussions using ChatGPT 5.1 Thinking to evaluate the prescience of user comments through contextual analysis. Development proved remarkably efficient, completed in approximately 3 hours using Claude Opus 4.5, at a cost of roughly 58 dollars for 930 model queries. The resulting platform features a "Hall of Fame" highlighting the most insightful commenters from December 2015, ranked by an IMDb-style scoring system based on prediction accuracy. By examining how technology discussions have evolved over the past decade, the project demonstrates the effectiveness of language models in contextual and historical analysis, offering a novel perspective on Hacker News' intellectual trajectory.