Those Who Swift - Issue 226
Your weekly dose of Swift!
Weekly note ✏️
Last week, both Stack Overflow and Substack released surveys, and—intentionally or not—both focused heavily on AI. Coincidence? Maybe. But let’s stick to the facts.
The Stack Overflow survey spans a wide range of topics: age, company size, learning methods, and technology stack. Substack, on the other hand, as a blogging platform, centered primarily on AI features and usage.
As someone who is both a developer and a writer, it’s fascinating to see how expectations and usage patterns differ across audiences. Developers are mostly adapting (or planning to), though there’s still some distrust toward AI-generated responses. Still, they’re happy to offload boilerplate code and get help with less popular tasks like writing documentation and creating tests.
On the content side, AI tools for brainstorming and grammar checks have become indispensable. And that’s nothing to fear—after all, advanced grammar tools already function much like LLMs for accuracy and context detection.
According to Substack’s report, 85% of creators in the Technology category use AI—likely developers or those in related fields. The age breakdown is even more interesting: younger bloggers lean on AI for brainstorming, writing, and translation, while older ones use it for image generation and research.
From a broader perspective, the stats (see graph) make one thing clear: adoption is inevitable, even if usage patterns vary across communities.
Connect with the "Those Who Swift" team - Justas Markus & Anton Gubarenko 👋
Sponsor 🤝
Forget about Ruby and Fastlane installation issues!
Discover Codemagic CLI tools — the free, open-source Fastlane alternative for automating iOS builds, code signing and publishing.
Swift Around the Web 🌐
SwiftUI for Mac 2025
Sarah Reichelt explores SwiftUI’s latest features tailored for macOS 26, using a custom sample app to showcase key updates. It’s a comprehensive, hands-on tour of how SwiftUI on Mac continues to mature into a full-fledged, modern UI toolkit.
Read more.📍
Using Foundation Models Framework to Stream from External LLM Providers
Natasha explains how to use Apple’s FoundationModels framework in Xcode 26 Beta 4 to stream responses from external large language models—instead of solely relying on the built-in on-device model. This approach opens up new flexibility for building more capable, multi-model agent systems on Apple platforms.
Read more.📍
Encoding vs Encryption in iOS: What to Use and When
Himali explains that encoding ensures data usability—like Base64 or ASCII—making content compatible across systems and easily reversible without a secret key. In contrast, encryption transforms data into a secure, unreadable format using cryptographic keys so only authorized users can decrypt it, preserving confidentiality.
Read more.📍
Coding 👨💻
Define Scroll Edge Effect Style for a ScrollView in Liquid Glass
This guide shows how to use iOS 26’s Liquid Glass scroll-edge effects to maintain visual clarity when content scrolls under glass-styled elements. It explains selecting between adaptive fade or hard-edge styles—perfect for toolbars—so that floating views stay legible and visually distinct from scrolling content beneath.
Read more.📍
Swift Observations AsyncSequence for State Changes
Starting in Swift 6.2 / iOS 26, Swift introduces the Observations type—an AsyncSequence that streams meaningful state changes from any @Observable object. Instead of relying on scene phase events or Combine publishers, you can now observe computed closures transactionally, persisting updates like UI state reliably and efficiently.
Read more.📍
Swift Concurrency: Part 1
Swift Concurrency: Part 1 introduces Task, Task.detached, and task priorities—explaining the differences between them and when to use each. A foundational guide to structured concurrency in Swift for writing clear, maintainable async code.
Read more.📍
Design 🎨
Building for Hate: Designing for Deception
Giselle Katics explores the line between persuasive and manipulative UI design, warning how familiar patterns can become deceptive when engineered with the wrong intent—prioritizing user retention over well-being. She breaks down common dark patterns like urgency, reward triggers, and confirm shaming, citing ethics, mental health, and even upcoming legislation like the Digital Fairness Act to underscore the need for more empathetic, transparent design.
Read more.📍
Other cool stuff 🧰
SharingGRDB Public Beta
Point‑Free has released a public beta of their SharingGRDB library, a SwiftData alternative that enables full CloudKit synchronization with a local SQLite database. With just a few lines of setup, apps gain real-time sync across devices—including support for shared data, foreign keys, and large binary assets like photos and audio.
Read more.📍
Understanding Swift’s @available Attribute
Natascha Fadeeva breaks down how Swift’s @available attribute lets you control API usage across platforms and OS versions—defining where and when specific code should (or shouldn’t) run. It covers compiler keywords like deprecated, renamed, and obsoleted, plus newer checks like #available and #unavailable to conditionally execute code at runtime.
Read more.📍
Sendable, sending, and Nonsending Explained
Fatbobman delivers a clear breakdown of Swift 6 concurrency’s core messaging: when and how to use Sendable, unchecked Sendable, sending, and nonsending annotations. The guide helps developers choose correct concurrency patterns and prevent subtle race conditions.
Read more.📍
AI 🤖
Qwen-Image
Qwen‑Image is a new multimodal diffusion model released under Apache 2.0 via Hugging Face, specializing in high-precision text rendering and interactive image editing . It delivers exceptional layout-aware icons, GUI mockups, and multilingual typographic fidelity—even for complex languages like Chinese—by synthesizing billions of annotated image‑text pairs with structured metadata .
Read more.📍
gpt-oss by OpenAI
OpenAI has released its first open‑weight models since GPT‑2: gpt‑oss‑120b and gpt‑oss‑20b, first time in years. These models offer strong performance—comparable to OpenAI’s proprietary o3‑mini and o4‑mini—and can run locally on everything from laptops to desktop machines without requiring server calls.
Read more.📍
Tutorials 📒
Building a Heart Rate Heat Map with SwiftUI & HealthKit
Wesley Matlock walks through creating a MapKit-based heatmap that syncs GPS routes with HealthKit heart rate samples to visualize exactly where elevated heart rates occurred during activities. It’s a practical example of blending health and location data for insightful visual feedback.
Read more.📍
Video 🎥
Liquid Glass - 5 Things You MUST Know Before Implementing
Sean Allen discusses new Liquid Glass design and points the main controversial paradigms about adaption. From native components usage to general visual appeal.
Watch here.📍
Yet, another thing…🛠️
AI Use Cases
This repository compiles over 300 real-world machine learning system design case studies from more than 80 companies, including Netflix, Airbnb, and DoorDash. Each entry covers practical architectures, trade-offs, and deployment challenges—making it a go-to resource for developers designing scalable ML and LLM systems.
Learn here.📍
Thanks for reading Those Who Swift! Subscribe for free to receive new posts.