Who hurts
Knowledge workers with 900 unread saved articles and a guilt reflex.
The problem
Saving is frictionless and reading is not, so the queue grows forever. The tool becomes a monument to intentions and eventually gets deleted.
What you build
Invert it. Nothing is a list — each week you get three items chosen for what you're working on right now, in the length you have time for, and everything older than ninety days is archived without ceremony.
Why now
Matching a saved article to this week's actual work needs semantic understanding of both, which is now a cheap embedding lookup.
Validate it this week
Build it for yourself and use it for six weeks. If your own read rate doesn't move, no marketing will save it.
Why you'd keep winning
None. This is a taste-and-craft product that wins on feel or not at all.
The honest risk
The most crowded category on this page, with beloved incumbents and a user base that doesn't pay. Build it because you want it, not for the market.
The prompt is written to make an AI argue with you before it writes code — that first round of pushback is worth more than the scaffold.