Weekly AI Insights
Practical perspectives on building software with artificial intelligence. Written by Alex Shively — one post per week.
Inside Apple's WWDC do-over, the Gemini deal nobody expected, and the hidden "choose your own model" switch buried in iOS 27
Apple rebuilt Siri from scratch on a Google model at WWDC 2026, then quietly left out its most interesting feature: a switch that lets you set Claude, ChatGPT, or Gemini as your iPhone's default AI. Here's what shipped, what's hiding in the beta, and what it means for the rest of us.
What the surprise government recall of Fable 5 and Mythos 5 means for anyone building on AI
The US government ordered Anthropic to disable its two most powerful AI models just three days after launch. Here's what actually happened, why it matters even if you don't build with Claude, and the quiet lesson it teaches about depending on any single AI vendor.
Inside the first documented in-the-wild attack run by an autonomous LLM, and what it means for the rest of us in IT
Security researchers caught the first cyberattack in the wild driven entirely by an AI agent, not a human with AI help. The whole intrusion took under an hour. Here is what happened, why it is different, and what to do about it if you work in IT.
Inside Claude Opus 4.8, and why honesty might be the most useful upgrade Anthropic shipped this year
Anthropic released Claude Opus 4.8 this week, and the headline feature is unusual for an AI launch. It's better at admitting what it doesn't know. Here's why that matters more than the benchmark scores.
Inside the agent Google just unleashed at I/O, and what to make of it if you work in IT or AI
Google unveiled Gemini Spark at I/O this week, a 24/7 personal AI agent that runs in the cloud and takes action across your apps on your behalf. Here's what it actually does, why MCP makes it tick, and what it means for the rest of us trying to keep up with the agent era.