We're entering the final stretch of 2025 with the Apple ecosystem busier than ever. This week brings us Embedded Swift gearing up for its debut in Swift 6.3, Apple showcasing cutting-edge research at NeurIPS 2025, and the App Store Awards finalists warming up for the big night. But if anything defines this season, it's the feeling that the pieces of the approachable concurrency puzzle are finally falling into place.
Speaking of pieces falling into place, this week we had the privilege of saying goodbye to 38 new developers who completed the Swift Developer Program. Weeks of intensive training where they built a solid foundation in Swift, SwiftUI, architectures, and best practices. The most valuable thing they take away isn't the tools they learned to use, but the deep understanding of the language and its fundamentals. In a world where LLMs generate code at the drop of a prompt, those who understand the why behind every line will always be one step ahead. Congratulations to all of them—the journey has just begun.
Embedded Swift Takes Off with Swift 6.3
On November 17th, the Swift team announced the improvements coming to Embedded Swift in version 6.3. For those unfamiliar, Embedded Swift is a subset of the language designed for resource-constrained environments: microcontrollers, embedded systems, and that whole universe where every kilobyte counts.
The interesting part is that the vast majority of Swift works exactly the same in embedded mode. The difference lies in the compiler, which generates significantly smaller binaries than standard compilation. Until now, getting Swift running on a microcontroller required dark magic and infinite patience. With Swift 6.3, having Embedded Swift as part of nightly builds and eventually the stable release is a game-changer for those working in IoT or low-level systems.
Concurrency That Doesn't Scare You
One of the barriers to entry for modern Swift has been concurrency. The actor model and async/await are powerful, but for many developers coming from other backgrounds, it felt intimidating. Antoine van der Lee published this week a guide on Approachable Concurrency in Swift 6.2 that promises to change that.
Swift 6.2 introduces improvements designed to make concurrency more approachable without sacrificing safety. The goal is clear: writing correct concurrent code should be as natural as writing synchronous code. Data races are still the enemy, but now we have better weapons to fight them without needing a PhD in distributed systems.
If you haven't made the jump to modern concurrency yet, this is the time. And if you want to dive deeper into applying these techniques alongside AI tools for development, the Swift AI Mastery Program will give you the keys to mastering both the code and the tools that power it.
SwiftUI Gets Its Liquid Glass Makeover
iOS 26 brought the Liquid Glass interface, and SwiftUI received its corresponding dose of goodies. Paul Hudson published his now-traditional What's new in SwiftUI for iOS 26, and there's plenty to dig into:
- Native WebView: We can finally embed web content without reaching for UIKit. Goodbye
UIViewRepresentablewrappingWKWebView. - Rich Text Editing:
TextEditornow supportsAttributedStringfor real rich text editing. - @Animatable Macro: Animating custom views and modifiers has never been easier.
- Collapsible TabView: The tab bar can minimize on scroll, following the Liquid Glass philosophy.
- openURL with In-App Browser: No need to kick users out to Safari just to show a link.
Index labels in list sections and fixed width for icons in Label might seem minor, but they're the kind of improvements that make you think "how did we ever live without this?"
Apple at NeurIPS 2025: MLX and the M5
The NeurIPS 2025 conference is just around the corner, and Apple arrives loaded with machine learning research. Among the highlighted works, ExpertLens caught our attention—exploring how activation steering features in language models are highly interpretable.
But the article that really got us excited is Exploring LLMs with MLX and the Neural Accelerators in the M5 GPU. The MLX framework continues to evolve as the go-to tool for experimenting with inference techniques on Apple Silicon. And yes you read that right: M5 GPU. The neural accelerators in the new generation of chips promise to take local model execution to another level.
Apple's ML research isn't just academic. All of this eventually lands in Core ML, Create ML, and the Apple Intelligence features we use every day. Understanding how it works under the hood makes us better developers.
Mini Apps: Apple Opens a New Door
On November 13th, Apple announced the Mini Apps Partner Program. It's a program for apps built with web technologies offering a reduced 15% commission on In-App Purchases if you implement the required commerce and age verification APIs.
Alongside this, the App Review Guidelines received a major update: age restriction requirements for creator apps, 36% APR caps for loan apps, and prohibition on using competitor icons without approval. If you publish on the App Store, it's required reading.
Texas Enters the Game
Starting January 1, 2026, Texas SB2420 will require age verification for app downloads. Apple will provide the Declared Age Range API to help developers comply while maintaining privacy standards. If your app has users in Texas, it's time to start planning.
A Coffee with Code
The Temporal SDK for Swift is now available as an open source project. Temporal lets you build durable and reliable workflows that survive failures, restarts, and deployments. If you work with distributed systems or long-running processes, it's worth a look:
import Temporal
@WorkflowDefinition
struct OrderProcessingWorkflow {
@WorkflowQuery
var status: OrderStatus = .pending
@WorkflowMain
func run(orderId: String) async throws -> OrderResult {
// Temporal workflows survive restarts
let payment = try await PaymentActivity.processPayment(orderId: orderId)
status = .paymentProcessed
let shipping = try await ShippingActivity.scheduleShipment(orderId: orderId)
status = .shipped
return OrderResult(payment: payment, shipping: shipping)
}
}The idea is simple but powerful: you define your business logic as regular Swift code, and Temporal handles durability, retries, and failure recovery.
ServerSide.swift 2025 Videos Are Live
For server-side Swift enthusiasts, the ServerSide.swift 2025 conference videos are now available on YouTube. From Vapor to Hummingbird, plus real-world production use cases, there's material for several afternoons of learning.
And That's a Wrap
We close this last week of November with the feeling that Swift keeps maturing in every direction: embedded systems, approachable concurrency, liquid glass interfaces, and cutting-edge ML. The ecosystem has never been this rich or offered so many possibilities.
To the 38 colleagues who just graduated from the Swift Developer Program: the knowledge you've gained is your superpower. LLMs can generate code, but only those who understand the fundamentals can evaluate whether that code is correct, efficient, and maintainable. Keep learning, keep asking questions, keep breaking things to understand how they work.
If you'd like to leave us a comment or share anything, we'd love to hear from you.
See you soon, and until then: happy native coding.
Thanks for reading our Newsletter! See you in Be Native.