React-Native Is Dead... Thanks to AI.

Author: Better Stack · Watch on YouTube

Video authored by the channel shown, not by IANIA; all rights belong to the creator. Transcript and summary generated automatically; the original-language content is preserved.

Original language: English

Original video: React-Native Is Dead… Thanks to AI. · Channel: Better Stack · 06:11 · Watch on YouTube

Topics: AI agents, Software engineering with AI

Summary

Shopify is moving away from React Native and back to fully native apps, arguing that AI has reduced the cost of building separate iOS and Android versions. The video explains Shopify’s Helix system, which breaks migration work into small checkpoints that are tested, visually reviewed, and approved before continuing. It also describes how separating business logic from the UI lets agents run fast tests through a CLI without relying on slow simulators. The presenter says these workflows make native development more attractive and may further erode the value of cross-platform tools, while noting that maintaining two apps still has costs.

Key points

Shopify has abandoned its React Native repositories and switched back to native development.
The company says improved coding models have made building features separately in Swift and Kotlin less costly.
Shopify’s Helix system divides migration work into small, reviewable checkpoints rather than relying on a one-shot conversion.
Each checkpoint must pass tests, visual review, two adversarial code reviews, and human approval.
Helix records feedback from reviews so the process can become more autonomous as migration progresses.
Shopify decouples business logic from the UI so agents can test it headlessly through a CLI.
CLI-driven testing can avoid slow, flaky interactions with simulators and accessibility trees.
Shopify’s Shop app has already shipped fully native, and the video cites 3 million monthly downloads.
Shopify’s major open-source React Native packages are being handed over to third-party maintainers.
The presenter believes AI agents could make native development increasingly attractive, though teams still need to maintain two apps.

Chapters

Start	Chapter	Summary
00:00	Shopify’s move back to native	The video introduces Shopify’s decision to leave React Native and argues that AI has changed the economics of building separate native apps.
01:02	Why Shopify changed course	The presenter describes Shopify’s reasoning and introduces Helix, a system designed to make AI-assisted native migration more reliable.
02:06	Helix’s checkpoint-based workflow	Helix turns migration into small checkpoints that must pass tests, visual review, code reviews, and human approval.
03:38	Faster testing through the CLI	Shopify separates business logic from the UI and uses CLI-driven testing to let agents iterate without slow simulator workflows.
05:10	The future of cross-platform development	The video notes Shopify’s open-source handoffs and the native advantages that lead the presenter to question the future value of cross-platform tools.

Quotes

The biggest claim of platforms like React Native is you can build your app once, release it to both Android and iOS. — 00:00

Helix instead anticipates that the first attempt will not be perfect — 02:37

It doesn’t really matter how good the model is if it can’t test its work quickly — 03:08

the Shop app has already shipped fully native. — 04:39

I do think this is the beginning of the end for crossplatform tech. — 05:40

Concepts and entities

React Native, Shopify, Helix, CLI, Swift, Kotlin, Shop app, XCUI test, CloudKit, SwiftUI

Full transcript (original language)

00:00 Shopify have been huge backers of React Native for half a decade, but this week they just announced they’re going back to native. And if you understand why, you may want to switch as well. The biggest claim of platforms like React Native is you can build your app once, release it to both Android and iOS. It’s just one codebase, one language, and far less maintenance. And Shopify were all in on this idea. They maintain React Native packages with apparently over 2 million weekly downloads. And as recently as 2025, they published articles renewing their support for the

00:31 technology. But one year on and they’ve suddenly had a massive change of heart. All those open source repos are being abandoned and the company has fully switched back to native. The main reason is advancements in AI. With the latest models and some smart in-house tech, the burden of building two separate apps no longer apply. You can just point a model at your iOS app and be like, “Yo, build me this in Android.” And it works. So today we’ll go over exactly how Shopify are building for native and I am in complete agreement with them on this.

01:02 Despite building in React Native myself for a good part of the last decade, I’ve also switched back over to native and I’ve been pleasantly surprised. So hopefully in this video I’ll convince you to do the same. [music] So Shopify last week published an article saying native is now the future of mobile at Shopify and it’s full of absolute gems. They talk about preventing AI slop by building a system called Helix to break agent work into smaller chunks and how to handle fast

01:34 feedback loops so agents can test quickly. After building with native for the last few months myself, it has fundamentally changed my own workflow. So let’s run through everything. But first off, why the switch? Because as they themselves said just in 2025, I wrote that the future of React Native was bright and that Shopify planned to keep investing in it. But since then, coding models have gotten dramatically better. And for apps and our team, building the same feature in Swift and Cotlin no longer carries the cost it used to. So the benefits of React Native

02:06 building once for both platforms no longer apply thanks to AI. So, first off, they built this system called Helix to solve the shortfall of current models. They said it’s tempting to just point an LLM to the React Native codebase and try to oneshot the same feature in native, but it just doesn’t work. Even if you ask it to gather as much information as it can, frees that up into specs and task files and then implement it. You end up with a huge amount of unmaintainable code that can’t be shipped. Helix instead anticipates that the first attempt will not be

02:37 perfect and builds a loop where an imperfect attempt simply cannot move forward until it becomes a good result. A developer first points Helix at a screen. It reads the React Native code, proposes a sequence of checkpoints, small ordered slices of the work that can be reviewed in minutes. Then checkpoint by checkpoint, it builds. Each one must prove its behavior with tests. Match the running app in a visual review. survive two adversarial code reviews and get a human’s nod before it’s committed and the next one starts.

03:08 Feedback from each of these reviews is remembered, so the loop gets more autonomous as the migration progresses. This approach effectively eliminates slop. But with all of this in place, it’s still left a massive problem because compiling and testing for native is slow. Agents can make changes in seconds, but it takes them several minutes to test the output. This makes iterating extremely slow and manual. It doesn’t really matter how good the model is if it can’t test its work quickly, which is especially difficult on mobile.

03:38 Shopify solved this in two specific ways. First, business logic should be completely decoupled from the UI. This means logic can run headlessly on desktop and is available to agents to run via a CLI that allows them to iterate in milliseconds instead of minutes by not involving simulators. From the context of web development, this is like holding all of your app state in JSON, then firing actions to update that JSON store. The UI is just a representation of that state. So there’s never a need to process the

04:09 accessibility tree, inspect the screen, or find and click any elements. Everything can be driven without the UI at all. And in cases where you really do need to render the UI, you can still drive all of the interactions from the CLI, so everything remains fast. This is the agent testing an app in real time. And this is not sped up at all. I’ve put together a simple demo doing the same, which through the CLI can pull up YouTube and subscribe to Bet Stack. But you don’t even need an agent to do that. You can just subscribe using the button below this video. Without a setup like

04:39 this, you’d need to use Apple’s XCUI test. But this runs out of process, finds elements by walking the accessibility hierarchy, synthesizes taps, then re-queries the tree to verify, which in contrast is super slow and flaky. So, following these techniques, the Shop app has already shipped fully native. That’s an app with 3 million monthly downloads. Now, the caveat here is Shopify is of course a massive company. You’d still have to maintain two apps yourself, but I’d say this is what’s possible with models

05:10 today. Over the next few months and years, this is only going to get easier, and I think that does start to erode React Native’s value even more. Shopify maintain a few major open source packages including flash list with 2 million weekly downloads. React Native Skia and Restar. All of these packages are now being handed over to third party maintainers. So it seems Shopify are entirely wiping their hands with React Native. Airbnb also notably left React Native many many years ago because

05:40 maintaining iOS, Android, and a bridge was three platforms, not one. Now, I’ve loved developing in React Native for years. I spent most of my career in that space. But I do think this is the beginning of the end for crossplatform tech. As agents improve, why would you not want to optimize for performance and better platform native functionality like CloudKit and Swift UI? Of course, you could interact with these systems through React Native, but performance is never going to be as good. Let me know what you think in the comments.

06:11 Subscribe if you haven’t already. We release tech and AI content constantly and I’ll see you next time.