Workflows
LAD — List A Day
A fresh ranked top-100 list every day.
Each morning, /lad picks a never-done-before topic, researches it via NotebookLM, ranks 100 items, generates the markdown + HTML + infographic, and persists to AIDB.DAILY. ~810 lists and counting.
- Cross-checked against historic lists
- Multi-view HTML + portrait infographic
- Upsert to AIDB.DAILY
Why
Top-100 lists are evergreen, shareable, and rank well — but only at scale. One a day, deduped against history, with a portrait infographic and HTML views, compounds into a moat.
How
- Pick a never-done-before topic, dedupe against MongoDB AIDB.DAILY
- NotebookLM research → 100 ranked items → md + HTML + JSON
- Portrait infographic + upsert to DB + log to fleet
Proof
- Lists in the archive
- ≈810
- Topics deduped against
- every prior LAD
- Outputs per list
- 5 (md · HTML · JSON · infographic · DB row)
LAD — list a day, deduped against history
Pick · Research · Rank · Persist
Hover or tap a node to see details.
FAQ
- How do you avoid generating a list you've already done?
- Hard dedupe against AIDB.DAILY before research kicks off — same topic, same angle, same ranking criteria all caught at step 1. A new LAD never repeats a prior topic shape.
- Where do the source items come from?
- NotebookLM is the primary research surface, with WebSearch fallback for niche topics. Both feed the same scoring rubric, so the source layer is interchangeable per topic.
- What does the portrait infographic actually look like?
- A single PNG sized for social — title, top 10 items in a numbered column, a stats footer, and a QR back to the full HTML list. Shareable in one swipe across every channel.
In production
- yb100 archive
810+ top-100 lists, all deduplicated against history, all rendered to md + HTML + JSON + infographic.
See it - Multi-output per list
Each LAD ships five artefacts: Markdown master, multi-view HTML, JSON, portrait infographic, Mongo row.
- Never-repeated topic picker
Topic picker hard-dedupes against AIDB.DAILY — no day ever lands the same LAD twice.