The Learning Library
The Learning Library (Library in the nav) is the team's collective memory of every resolved hypothesis. It's where you go before writing a new bet — to find out whether your team has already tried something similar and what they learned.
Private hypotheses are never shown here. Drafts and active hypotheses don't appear either — only resolved ones.
Header and counter
The counter under the title tells you how many resolved hypotheses across the whole team. As your team works, this number grows; it's a small but real measure of organisational learning.
Search and filter
A single bar with everything you need to slice the library:
- Search — full-text search over the title and learning notes (powered by Postgres FTS, so ranking is meaningful).
- All outcomes — limit to Held, Broke, Wrong Magnitude, or Unresolvable bets.
- All templates — limit to one template family (e.g. Pricing Change only).
- All PMs — limit to one author. Useful when you remember "Alex tried this".
- Date range — two date pickers for resolved-between filtering.
- Sort — Newest first (default), Oldest, Confidence ↑/↓, or Broke first.
Tip: the highest-leverage search is Broke first with a tag filter. The team's failure log on a specific topic is usually more useful than its success log — failures encode where the world pushed back.
Entry cards
Each result is a single card showing:
- The chain icon (⛓) if the hypothesis was created from a template.
- Title.
- Author · Template · Resolution date · Confidence (e.g. C1/5).
- The learning note preview.
- Tags.
- The outcome pill on the right (HELD / BROKE / WRONG MAGNITUDE / UNRESOLVABLE).
Click a card to open the full hypothesis detail page — the original chain, pre-mortem, actual values, and any team comments.
Tag filter bar
The same tag pills you see on the dashboard work here too. Pick one to scope the library to a topic. Combining the tag filter with the outcome filter is the fastest way to surface relevant prior art.
Previous: The Calibration page ← Next: The Team Dashboard →