Everything that landed over the last few days, where the whole project stands today, and a full, plain-language map of what's left to build.
In a few days The Brick Drop went from 28 graded sets and no personalization to 3,600+ sets and counting, graded through four buyer lenses, with a one-tap "for you" switcher — while a background sweep grades the entire buildable catalog and a richer wave of blog reviews feeds every grade. The core product is built and live. What's left is depth: trust, coverage, and refinement.
Three big things shipped at once: personalization, a catalog explosion, and a deeper well of reviews.
The app is called "how a set rates for you" — and now it actually does. Every set is graded four ways, and a one-tap switcher lets you see it through your lens.
One general grade. The engine could grade per-lens, but nothing stored it and there was no way to pick a lens. "For you" existed only in code.
Every set carries a display / builder / parent / investor grade. A segmented lens switcher (with a first-run nudge) sets it via URL + cookie; the card shows "A− for you · B+ overall" and the reason that fits your lens. Even the share image respects your lens.
A background grading sweep is working through every buildable LEGO set from 2026 back to 2017. It runs safely and can resume anytime, grading in parallel where the data sources allow.
| Year | Graded so far | Notes |
|---|---|---|
| 2026–2021 | 3,596 | Complete through 2021 — the years people are shopping right now |
| 2020 | in progress | The sweep is here today, marching backward |
| 2019–2017 | queued | The remaining back-catalog; ~5,754 buildable sets total is the target |
A harvest of Brick Insights added 5,883 review references across the catalog — and confirmed your hunch: there's more written blog coverage (3,916) than YouTube (1,711). The mining engine now distills blog and forum articles alongside video transcripts and comments, so grades rest on a broader base of real opinion.
All of this held the quality line — 552 tests green, clean type-checks, clean build, each feature shipped as its own reviewed pull request, and the live site never broke.
Three things are actively in motion as you read this.
And Slice 17 — per-lens hot-takes — is designed and spec'd (the current working branch). Today the grade changes per lens but the words don't; this writes a take for each buyer, so the investor view leans on resale and cites the investor grade, not the overall one.
The Brick Drop is a real, live, installable web app with a deep engine behind it. In plain terms:
The whole remaining project, grouped into four tracks. Each item is tagged by status so you can see what's done, moving, next, or later.
The grade is only worth what people believe it is. This track is about making every letter defensible.
| In flight | Smart Price fix — finish and ship the pricing edge-case fix so licensed sets grade fairly. |
| In flight | Deep re-mine pass — distill the 5,883 harvested reviews into every set, so grades reflect the full written record, not just what existed at sweep time. |
| Next | Validate the sentiment engine — build a small hand-labeled set of reviews and measure the AI's read against it, so we can retire the "unvalidated" caveat and trust the scores. |
| Next | Value calculation v2 — a sturdier "is it worth it" model that handles licenses, exclusives, and thin markets gracefully (the Smart Price fix is step one). |
| Later | Prediction model — an honest, clearly-labeled "Early read" grade for sets before they release. |
A grading tool is only as useful as the sets it covers, and grades go stale as new reviews land.
| In flight | Finish the sweep — grade the remaining 2020–2017 back-catalog to reach the full ~5,754 buildable sets. |
| Next | Living grades — a routine that re-grades sets as fresh reviews appear, so a hot new release sharpens from "Early read" to a confident grade over time. |
| Next | Full-catalog search — let people find (and grade-on-demand) any set by name, not only the ones already graded. |
| Later | New-release ingestion — automatically pull and grade sets as LEGO announces them. |
| Later | Source expansion — more review sources beyond Brick Insights + YouTube for even broader coverage. |
The grade is personalized; the rest of the experience isn't yet. This track deepens "for you" from a number into a point of view.
| Done | Per-lens grades + switcher — the four buyer grades and the one-tap "for you" view are live. |
| In flight | Per-lens hot-takes (Slice 17) — the words match the lens, not just the number: the investor view talks resale and cites the investor grade. |
| Next | Per-set tweak — a light "actually, I care more about play value here" adjustment on a single set, without a full profile. |
| Later | Profile — a saved identity (your lens, your history, your watchlist) so the whole app greets you as you. |
| Later | A personalized "For You" home — curated picks that match your lens, not just a search box + shelf. |
The craft that makes people want to use it and share it.
| Done | Share loop — branded link previews + a downloadable 4:5 card, both lens-aware. |
| Next | Crop-perfect share cards — bring your Crop Curator framing into the generated 4:5 card so wide sets sit right (today it uses a generic crop). |
| Next | True Value, expanded — grow "Is the price fair?" from one line into a real price/parts breakdown people can dig into. |
| Later | Onboarding polish — make the first-run lens pick feel like a moment, not a nudge. |
| Later | Discoverability — SEO and a share-driven growth loop so public result pages pull people in. |
If it were up to me, I'd order the near term like this — trust first, because everything else rides on the grade being right.
The hard part is behind us: the engine, the app, personalization, and the share loop are built and live, and the catalog is filling itself in. The road ahead is about trust (prove the grade), coverage (every set, always fresh), and depth (make "for you" go all the way down). None of it is a rewrite — it's finishing well.
The Brick Drop · built with receipts, not vibes