Stopscrolling should become a map. But the map is not enough.
Founder evaluation of turning the Stopscrolling engine into a standalone, sexy, modern city app: a live map of things worth leaving home for, powered by unusually good event radar rather than another listings database.
The call
Build the prototype, but do not pitch it as “a map-first events app.” Pitch it as the city’s live taste layer: a map that only shows what survived the engine.
The strongest version is not “Time Out on a map” or “Eventbrite without spam.” It is a new habit: open the city, see the few things that are actually worth moving your body for, and instantly understand the vibe, distance, timing, social context, and why it belongs there.
The red-team finding is important: map view already exists. doNYC explicitly ships a map view for nearby events, and fomo-style map-first event surfaces exist. So the map cannot be the moat. The map is the emotional surface. The moat has to be the ranking, freshness, source graph, and taste object.
Founder score
High enough to prototype; not high enough to company-build without a hard supply/retention test.
- Market pain: real
- Surface novelty: contested
- Engine advantage: plausible
- Moat: unproven
- Best next move: fake-door + live NYC pilot
Listening-room dance night · Ridgewood
Why it is here: creator overlap + venue history + low-spam source trail.
What this becomes
The actual product thesis
“The internet knows what is happening. Your phone does not know what is worth doing. Stopscrolling turns the city into a curated, living map.”
This should feel closer to Are.na + Apple Maps + a trusted friend who is annoyingly online than to Eventbrite. The app is not a database of events. It is a daily act of compression: from the huge, fragmented city-internet into a small number of high-confidence possibilities.
Open it when you feel restless. In ten seconds, understand where the city is alive tonight and which options match your energy.
Use Padawan/Sapith as the invisible research + ranking engine, then package the result as a consumer-native map object.
Do not promise complete listings. Completeness is where calendars die. The app should win by selectivity.
“Map-first” alone is not a wedge. It is already in the market. The wedge is map-first plus taste-filtered, source-bound confidence.
What the live radar changed
- There is plenty of supply. NYC Parks alone showed 7,966 events in its calendar window during the run; DCLA and SummerStage add free, public, geographically distributed cultural inventory.
- Supply is fragmented. Parks, DCLA, CECM permits, nonprofit calendars, venues, ticketing platforms, Instagram, TikTok, newsletters, and group chats each carry different pieces.
- The cleanest data is not always the sexiest data. Open municipal feeds are legally reusable and high-trust, but CECM permit data is dominated by sports/admin/civic categories and lacks lat/long, cancellation status, and rich descriptions.
- The sexiest data is not always public. The best parties, workshops, pop-ups, and scenes often start in Instagram stories, private chats, venue posts, creator posts, or newsletters.
- Incumbents are real but ugly in different ways. Eventbrite can become a spam firehose. Time Out is editorial and trusted but not personal. doNYC has a map. RA/DICE own specific live-music lanes. Partiful owns social event pages and Explore-style discovery.
What the red-team killed
- Killed: “No one has made a map for this.” doNYC already has a map view; map-first NYC discovery surfaces exist.
- Killed: “Open data alone can power the magic.” It can seed the map, but it will miss private/social/cultural supply and needs aggressive filtering.
- Weakened: “The engine should be invisible.” It should not look like an AI tool, but the product needs visible trust receipts: why this pin, source trail, freshness, and vibe confidence.
- Unresolved: whether map beats chat/list for actual attendance. Sapith recommends a preregistered surface test with verified attendance as the primary metric.
The app should feel like this
MVP: the smallest real version
- NYC only. Williamsburg/Bushwick/LES/Lower Manhattan first. Do not launch “all of NYC” as a weak map. Launch a dense, opinionated slice where the taste graph is strongest.
- Start as a mobile web app before App Store. A comparable any-city event scout reported three App Store rejections before shipping. Avoid that until the retention signal exists.
- Ingest three clean lanes: NYC Parks/DCLA/SummerStage for high-trust public culture; curated event/ticketing sources for paid inventory; social/creator/newsletter radar for the reason people actually care.
- Rank by “leave-home probability,” not popularity. The scoring model should combine freshness, source credibility, distance, uniqueness, fit to taste clusters, social proof, friction, and whether it is actually still happening.
- Ship one weekly test cohort. 50–150 NYC users, not a public launch. The question is not downloads; it is whether people attend or meaningfully try to attend.
Do not build yet
- User accounts beyond saves/invites
- Creator tools
- Full personalization graph
- Multi-city expansion
- Ticketing marketplace
- Every possible category
Those are scale features. The first proof is a taste-and-attendance loop.
Founder red-team
If the pitch is “events on a map,” it is dead on arrival. doNYC already says the same thing. The product must be “the only pins that matter.”
Municipal data is clean but not sufficient. The private/social layer is where the most exciting inventory lives.
If the engine requires heavy daily human intervention, it becomes a newsletter with map UI. If it is fully automated, it risks slop. The system needs calibrated human-in-the-loop review for only the top pins.
Free/public programming has weak ticketing take-rate. Revenue likely comes later from premium city companion subscription, sponsored but clearly labeled drops, affiliate/ticketing for paid inventory, or B2B/event-operator tools.
The emotional frame is strong: stop scrolling, go outside. This is not productivity. It is a better relationship with the city.
Sapith can produce source-bound, adversarially filtered city intelligence. That is the part a normal startup would struggle to reproduce quickly.
The next three founder reps
- Supply survivability audit. Pull one week of NYC Parks + DCLA + CECM + 20 handpicked social/newsletter/venue sources. Classify every candidate: attendable by a stranger / not; desirable / filler; geocodable / not; still live / stale. Kill if fewer than ~30 high-quality pins survive for the target geography/week.
- Blind taste test. Give 20 NYC people three outputs for the same weekend: Time Out, Eventbrite/DoNYC, and Stopscrolling map shortlist. Hide brand. Ask which five they would actually consider. Win condition: Stopscrolling is chosen for “things I would not have found but would actually do.”
- Surface test. Use the same ranked corpus behind three prototypes: map, elegant cards/list, and chat concierge. Primary metric is not clicks; it is route/ticket/calendar/share intent, then verified attendance where possible.
Kill criteria
- The map cannot produce a dense enough set of compelling pins in one neighborhood for one weekend.
- Users say the pins are “interesting” but do not save, share, route, RSVP, or attend.
- Time Out / RA / DICE / Partiful already win the blind top-five test for the launch segment.
- The engine cannot beat a smart human curator fast enough to matter.
Recommended positioning
Stopscrolling is not an events app. It is a map for getting your life back from the feed.
That sentence is the emotional center. The practical promise: “Every pin is a reason to leave the house.” The design promise: “Beautiful enough to browse, strict enough to trust.” The founder promise: “We own the intelligence layer that turns messy social/civic/cultural signals into a small, useful city map.”
Source trail
Source of record for free/open cultural programming and related calendars.
Official permit dataset; useful but operationally messy: no lat/long, no cancellation field, duplicate event IDs by occurrence.
Primary source; 60+ free/benefit performances across Central Park and neighborhood parks.
Primary competitor evidence: already markets a Map View for nearby events by neighborhood.
Useful but self-interested primary/market content: discovery is fragmented across social, group chats, venues, newsletters, ticketing, search, and event pages.
Competitive/substitute surfaces: ticketing, scene-specific discovery, social event platform, and editorial curation.
Research receipt: Sapith run run_5e47c6f62b6f, created 2026-08-21T21:27:00Z. Full path, high stakes, Grok web + X radar, Codex adversarial audit, Claude binding/synthesis. Residual risk: two highest-severity tickets returned partial, not complete — operational feed quality and surface-wedge superiority remain unproven until the recommended tests run.