A search engine for Hacker News obsessions: 18 years, 45M comments, charted

A new tool called Hacker News Trends does for HN what Google Trends does for search: it plots how often any topic, tool, person, or company has surfaced across roughly 45 million posts and comments since the site’s early days. You type in terms, overlay several at once, and watch each one’s mention frequency rise and fall as a live date-histogram. Clicking or dragging across the timeline filters down to the actual stories and comments behind the curve, searchable by term or author. The backend runs on Upstash Redis Search.

The interesting payload isn’t the engine but the patterns it exposes — a near-complete oral history of developer fashion told through community attention. Almost every category shows the same relay-race shape: one tool dominates the conversation, goes quiet, and a successor takes the baton. Bundlers go Webpack to Vite, editors go Sublime to Atom to VS Code, containers hand off from Docker to Kubernetes, ML frameworks march from TensorFlow to PyTorch to JAX, and the AI-lab narrative flips from OpenAI’s multi-year lead to a 2026 Anthropic surge that pulls level. Single-month spikes capture specific shocks: Zoom in the March 2020 lockdown, all three major game engines lighting up together during Unity’s September 2023 runtime-fee backlash, DeepSeek’s lone January 2025 tower.

For a technical reader, the value is twofold. As a navigation tool it makes HN’s collective memory queryable instead of anecdotal. As a dataset it’s a fairly honest proxy for hype cycles — useful for sanity-checking whether a ‘hot’ technology is genuinely ascendant or already past its attention peak, and a reminder that today’s default stack is usually just the current holder of a baton that keeps getting passed.