Build log

What if I tried to build a better social network than Meta?

25 March 2026 · 6 min read

Social App

A social app that’s actually social 🌍

This weekend, I tried to build a better social media app than Meta.

Small goal 🙂


The problem

Instagram pulls you out of the moment. WhatsApp lets relationships drift.

Neither is designed around the people you actually care about. They’re designed around keeping you on the platform.

And that creates a weird tradeoff:

A lot of us want to delete Instagram. But we don’t.

Because it’s the only place where those looser connections live — people you don’t message directly, but don’t want to lose touch with.

And WhatsApp doesn’t solve that. It’s too intentional. Too direct.

So we end up stuck:

Which means over time, we just… lose touch.


The approach

Start from contacts, not content.

Strip it back to three primitives:

Then remove everything that doesn’t serve that.

The feed is:

chronological, unfiltered, and only people in your phone

The goal isn’t to maximise usage. It’s to make staying in touch feel effortless.


The product

Social App has three simple surfaces:

1. Contacts (Home)

You open the app and see your people first.

This is the foundation:

your real-world network, not an audience


2. Feed

A lightweight stream of updates.

Tabs:

Posts are simple:

“Working from this café ☕” “Anyone up for a run later?” “In town this weekend”

It’s not content. It’s context.

This is where those “in-between” relationships live — visible, but low pressure.


3. Profile

Your current state, not a highlight reel.

More presence than identity.


Core features

Ambient map

See where people are — roughly.

Not precise. Not trackable. Just enough to answer:

“Is anyone nearby?”


Groups

Create lightweight circles:

The feed becomes smaller, quieter, more relevant.


Lightweight events

Quick coordination:

More:

“Drinks later?” Less: “Create event”


Designing the experience

A few constraints shaped everything:

No infinite scroll Content ends. You leave.

No likes or follower counts Removes performance.

No algorithm Everything is chronological and transparent.

No drift

WhatsApp starts simple… then becomes noisy and fragmented.

Social App avoids that by:


Building it

I timeboxed this to 48 hours.

That constraint forced every decision to focus on the core loop:

open → see people → understand context → act → leave

Everything else got cut.

Stack:

I used Claude Code to scaffold most of it, after mapping the system upfront.

Technically, this is all straightforward.

Which is kind of the point.


The experience

You open the app.

You see people near you. Someone you haven’t seen in months is around.

They posted:

“Working from a café”

You reply. You meet.

That interaction probably doesn’t happen on Instagram. And it definitely doesn’t happen on WhatsApp.

That’s the gap.


Outcome

Working prototype.

The constraint proved something interesting:

a social layer that is ambient rather than addictive is technically simple

There’s no complex ranking system. No recommendation engine. No engagement optimisation.

The hard part isn’t engineering.

It’s the cold start problem.

A social app with no people isn’t social. That’s a product challenge, not a technical one.


Final thought

This project isn’t really about building another social network.

It’s about asking:

what if a social network was truly designed for socialness?

Not attention. Not engagement. Not growth.

Just:

Because maybe the future of social media isn’t more content —

it’s not losing touch with people who matter.

🌍

← Back