iOS Newsletter
Hi folks,
So. Much. AI. It feels like it’s everywhere at the moment, doesn’t it? We’re leaning right into the hype this week with a look at the latest in Apple’s AI world. Don’t worry, we’ve also sprinkled in some excellent articles on Swift, Core Data, and keeping your codebase healthy. Grab a coffee, and let’s dive in.
Enjoy the selection!
Swift
Opaque Types in Swift: Hiding Details, Not Power
This article by Saifullah Sajid provides a concise explanation of opaque types in Swift, using the some keyword. It clarifies how they allow functions to return a concrete type that conforms to a protocol, without revealing the specific underlying type. This is particularly useful for maintaining abstraction and flexibility in your code.
Evolving a Codebase
A thoughtful piece from Dr. Beat Rupp on practical strategies for improving and evolving a codebase. It details various techniques, from low-risk compiler-assisted cleanups like reducing public interfaces and cleaning imports, to higher-risk refactorings such as dependency injection and abstracting third-party libraries, emphasizing the importance of unit tests for safe evolution.
Data
How Core Data Saves Data in SQLite
This in-depth article from Fatbobman’s Blog demystifies the inner workings of Core Data’s SQLite persistence layer. It explains the structure of the tables and fields that Core Data automatically generates, such as Z_PK, Z_ENT, and Z_OPT, and how they relate to NSManagedObjectID.
AI/ML
Getting started with Apple Foundation Models
A practical guide by Artem Novichkov to getting started with Apple’s new Foundation Models framework. It walks through the process of using on-device large language models for various tasks, with a focus on privacy and performance.
MLX-Outil
This project by Rudrank Riyam allows you to use the Qwen 3 1.7B model with the help of MLX Swift. It’s a Swift package that provides a set of system integration tools, allowing the language model to perform various actions, including fetching weather, accessing workout data, searching the web, and managing calendar events.
FoundationChat
An open-source chat application by Thomas Dimillian built with SwiftUI and the Foundation Models framework. It serves as a great example of how to build a modern, AI-powered chat experience on Apple platforms.
Tools
Sparkle: Distribution apps in- and out of the Mac App Store
This article by Antoine van der Lee details the process of using the Sparkle framework to distribute macOS applications both through the Mac App Store and independently. It covers the challenges of maintaining a single codebase for both distribution channels.
Olleh
Olleh (hello reversed) by Mattt Thompson is a command-line interface (CLI) and HTTP server that is compatible with the Ollama ecosystem, allowing developers to interact with Apple’s new Foundation Models. It acts as a bridge, making Apple’s native AI models accessible through familiar Ollama-style commands and API endpoints.