Sen 選
A social media feed with an end — by design, not by willpower. You set what the session is for and how big it is; an agent pulls from every platform, then stops.
A note on scope
Platforms wall off their feeds — no open “For You” algorithm, no cross-platform search — so this isn’t buildable today. Sen is a vision piece: the experience if platforms allowed it.
The problem
Infinite scroll is a design choice.
People who want to be intentional get pulled into doomscrolling anyway, their attention scattered across apps built to keep them. Intention has no home.
The challenge: one feed that gives you what you came for, then stops — no willpower required.
Who it’s for
Someone caught in the scroll who wants out — intentional by wish, scattered in practice.
The voice is empowerment, not comfort. Sen is a tool, not a comfort blanket. It should make you feel capable and in control — never coddled or scolded.
Who it’s not for
- The already-disciplined, who don’t have the problem.
- The deep-diver who lives inside one interest and is better served going straight to the source.
- The quit-social-media crowd, who want no feed. Sen offers a better feed — the opposite promise.
What this person needs
- To choose their focus — a topic, or the people they care about.
- Control over the session’s shape, including when it ends.
- A real stopping point, and to feel good about stopping.
- One place for both going deep and keeping up.
Ideate
Aggregator or new app → new app. The first instinct was to re-display each platform’s feed. But those algorithms are each platform’s crown jewels — they aren’t exposed. If Sen can’t inherit their personalization, its own AI has to become the algorithm. That pushed it from aggregator to a new app.
Convenience or control → control. “One place for all my apps” is a convenience play, and convenience means frictionless infinite scroll — structurally a better doomscroll machine. Leading with control is the only framing that fixes the problem instead of worsening it.
What stops the scroll → a finite feed. The centerpiece. The feed is finite: the scroll ends because there’s nothing after it, which kills doomscrolling structurally rather than leaning on willpower. The user sets the boundary — “top 30,” or “30 minutes” — with a sensible default for the majority who never change one.
Cut from v1: in-app posting and a portable taste profile. Both are later layers, not the core story.
Prototype
The core loop:
- Set intent. You say what you want — a topic (“women’s health”) or people (“what my friends have been up to”), plus a boundary — or take the default.
- The agent curates. The AI assembles a finite set across platforms to fit that intent and boundary.
- It ends. The session finishes at the point you defined.
Intent screen
You open into a choice, not a feed. Before you see anything, you say what this sitting is for.
The finite feed
One scroll of mixed-format content assembled for the session, with visible progress toward the end. Because the feed is finite, you can always see how much remains — the opposite of an infinite feed, where hiding the bottom is the whole trick.
The ending
When the set is done, you reach a designed, affirming close — “that’s your set,” not “you’ve been cut off.” From there you can start another session or leave satisfied.
The feed is a session you enter and finish, not a place you live inside. One intent input handles both jobs, so keeping up and going deep share one front door.
Design decisions
| Decision | Serves the need for |
|---|---|
| Visible progress toward a known end | A genuine stopping point |
| User-set boundary + sensible default | Empowerment — capable, not coddled |
| One input for topic or people | A single home for two jobs |
| One consistent experience, no modes | The “no doomscroll, ever” promise |
Test
Sen is a concept, so there’s no shipped data. Here’s how it should be validated.
What to validate
- Does starting with intent feel empowering, or like friction before the reward?
- Does a finite feed satisfy, or frustrate? The central bet — do people reach the end feeling done, or annoyed and reaching for more?
- Does it change behavior over time? A diary study: does Sen displace mindless scrolling once the novelty fades?
- What default earns its keep? Most users never change one, so compare boundaries directly.
Metrics — including the ones to ignore. Time-on-app and sessions-per-day are the wrong target. For Sen, those going up would signal failure, not success. Instead:
- Session completion — did users reach the end they set?
- Post-session satisfaction — did you get what you came for?
- Regret — do users feel good about the time afterward, versus the guilt after a doomscroll? Low regret is the real win.
- Displacement — less time in the other apps, a sign Sen is becoming the default.
The name
選 (Sen) means “to choose.” Every session begins with the user making a choice, and the product’s whole argument is that a good feed is a curated one — someone chose this, and it ends.
Outcome
If the bet holds, Sen’s users consume less but feel more satisfied — they get what they came for, feel in control, and stop without guilt. A hypothesis, not a measured result.
What I learned. The project got sharper by confronting its hardest constraint instead of avoiding it. “Everything in one feed” kept colliding with the fact that platforms won’t expose their feeds, and each collision forced a better decision. What started as “all your apps in one place” ended up as something more specific: a feed that structurally can’t become a doomscroll.
What I’d do next.
- Portable taste profile — a user-owned profile that persists and improves across sessions. The most exciting next layer, held out of v1.
- The format-collision problem — how genuinely different content types share one coherent scroll.
- The default boundary — research it properly, given its outsized effect on behavior.
- Validate that finitude satisfies — run the ending test first, since the whole thesis rests on it.