Product Engineering

The Inevitable Return to Native

AI is changing the economics of native mobile development, making separate Swift and Kotlin implementations a more practical long-term choice.

For years, cross-platform development offered an irresistible bargain: build a feature once, ship it everywhere, and avoid maintaining two separate mobile applications.

But that bargain came with tradeoffs. Abstraction layers add complexity. Framework dependencies shift. And making an application feel at home on each platform still takes deliberate work.

For products built to last, I think the return to native is becoming inevitable.

Shopify recently captured the reason for this shift: “LLMs changed one of the core assumptions behind our 2020 decision.” The company is moving its mobile apps from React Native back to Swift and Kotlin because AI changed the economics of maintaining two implementations. Shopify Engineering

Notion is also investing in native mobile development. Ji Pei recently announced: “We’re rebuilding Notion natively on mobile.” Ji Pei on X

Native development still means maintaining software for multiple platforms. Shopify reports that coding agents have reduced the work required to translate features, test behavior, and maintain parity. That cost has become less decisive in its choice of architecture.

The long-term appeal is clear: direct access to platform capabilities, first-party tooling, fewer framework layers, and more control over performance and interaction. Those advantages matter year after year, as operating systems evolve and users expect more from the products they use every day.

AI makes it increasingly practical to pursue those benefits without making every team duplicate all its effort. Engineers still need to make architectural decisions, review code, and verify the experience. But more of the repetitive implementation work can be delegated.

My bet is that the future looks like one shared product specification, with native implementations for each platform and AI helping keep them aligned.

The reason to share code was always economic. As those economics change, building directly for the platform becomes a much easier decision.

Keep reading

New posts, straight to your inbox.

What we learn shipping software: product engineering, AI-First delivery, and the parts of a project that decide whether it works.

We use your email to send you the newsletter. Unsubscribe any time, see our privacy policy.

Bring us the backlog.

In 30 minutes, we will show you what a Pod would ship first and how we would price it.