🤖⌚ Cloudflare Workers AI on the Apple Watch
It has been nearly a year since I published a walkthrough of any sort. The original purpose of this blog has given way to unhinged anecdotes about Portugal and Texas.
Today I return to my roots! Apple released iOS 18 earlier this week but it was missing the biggest feature: their generative Artificial Intelligence (AI) platform called Apple Intelligence (also, cheekily, styled as AI). A*AI is scheduled for a beta launch later this year.
As someone who frequently wears an Apple Watch, I’m excited about the Apple Intelligence but also impatient. I figured I could combine Cloudflare Workers AI, Meta’s latest Large Language Model (LLM), and an iOS Shortcuts to build my own tool for the time being.
👔 I work there. I work at Cloudflare. Some of my posts on this blog that discuss Cloudflare focus on building things with Cloudflare Workers. That said, I am a real Workers customer and pay my invoice to use it.
Cloudflare’s network provides a developer platform where builders can run compute workloads and store data. As part of that infrastructure, customers can also run AI inference. AI inference just refers to cases when you take an AI model and provide it with new information, like a question about Lisbon, and the model running on some piece of hardware infers a response. In this case, the model might have been built and trained elsewhere but is now running on Cloudflare hardware. You can use this type of inference in Cloudflare Workers AI to run one-off workflows or build real applications, like this Lisbon recommendation tool I put together last week.
