Skip to content

--case-sin-fomoSide project · B2C · 2026 – now

Sin FOMO.

Newsletters, picked for your goals.

Sin FOMO reads the newsletters piling up in your inbox and picks, with Amplified Intelligence (AI), what best fits your goals and what you want to learn: one short daily selection that learns from what you actually read. My own product: I discover, design, build and launch it solo, with AI as copilot and the same method I use at work.

Role
Everything: product, design, code, go-to-market
Status
Private beta · 10 testers
Stack
Next.js · Supabase · Claude API · Gmail API
Timeline
2026 – now

Fig. 1 · Side project · B2C

My role

Solo maker: every decision and every line of code is mine.

Built with AI as copilot and with the same gates I design at work: challenge, metrics and hypothesis before code, and a kill-switch date set before the first invite.

  • 01

    Discovery with my own method: problem statement, activation metrics and a kill-switch date, before building.

  • 02

    Brand and design system: its own identity, dark by default, contrast checked with a script, not by eye.

  • 03

    The product end to end: onboarding, goals, the AI selection engine, feedback and the Gmail connection.

  • 04

    Go-to-market in small batches: private invites first, close to the first users.

Context

It started as a personal bot. Every weekday at 7:00 it sent me three pieces on AI and design, and learned from what I opened.

It worked for me. The real question was whether it worked for anyone else, so I turned it into a web product around one idea, your goals decide what you read, and gave it a name: Sin FOMO.

Today it's in a private beta with 10 testers from product teams: product owners, product designers, a design system specialist, two managers and developers.

Problem

Newsletters pile up. The good ones get buried with the rest, and “I'll read it later” quietly becomes never. What does get read is whatever arrived last, not what helps you grow.

  • Noise

    Too many subscriptions and no time to curate them. One early tester had 150 in their inbox.

  • Setup

    Adding sources by hand is the wall. Nobody wants to subscribe again to what they already get.

  • Proof

    It worked for me. Before charging anyone, it had to work for people who aren't me.

Key decisions

  1. 01

    Read the headers, not the inbox.

    To bring in the newsletters people already get, Sin FOMO connects to Gmail with metadata-only access: it spots newsletters by their unsubscribe header and follows their public feed. It doesn't read your emails. I ruled out a forwarding address: more invasive, and still manual.

    trade-off: Newsletters without a public feed can't be followed yet. They're shown, disabled and explained, not dropped in silence.

  2. 02

    Zero-friction feedback, with no rewards attached.

    One tap per piece, useful or not, with no reasons to fill in. Rewards come from product surveys instead, so the signal that tunes each selection stays honest.

    trade-off: A thinner signal per tap. The rich feedback lives in the surveys.

  3. 03

    Ship small, close to the first users.

    No onboarding tours built up front: the private batch goes out as it is, and I join the first sign-ups live. Help gets built only for what actually confuses people.

    trade-off: The first testers see rough edges. That's the point.

Craft

Your inbox in, one selection out.

The loop behind every selection: who you are, the sources that matter, your goals, and a track that tunes itself every day.

Fig. 2 · The learning loop, recreated with fictional content. Scroll to go through it.

Reflection

On my own product nobody enforces the gates. That's exactly why I keep them.

Next case · 04 / 05LandbotIntegrations without a single webhook.Web + Product · Integrations · 2021 – 2023▶ Continue