Overview
AI news is a firehose where ninety-five percent is noise, reposts, or hype. AI Hacker Daily is a pipeline that drinks the firehose so you don't have to: every day it collects the candidate stories, scores and curates them down to five picks, and publishes the digest — same format, every day, no filler.
The constraint that shapes the whole product is the number five. Forcing the pipeline to rank down to exactly five picks means the curation step has to make real choices, and the reader gets a two-minute read instead of another infinite feed.
How it works
- Collect — sources are pulled on a daily cron into Postgres: launches, papers, repos, and developer-relevant news.
- Curate — a scoring pass ranks candidates for developer relevance and novelty, deduping the same story told twelve ways.
- Publish — the top five go out as the day's digest on the site, each with a one-paragraph plain-English summary of why it matters.
Stack
A Python pipeline on Railway with Postgres for story state and dedup history, and a Next.js front end serving the daily digests and archive.