Xcode 27 Ships With Claude Code Inside: Apple's Take on Agentic CodingJune 11, 2026 · 5 min read
At WWDC 2026 Apple built Claude Code, Gemini, and Codex directly into Xcode 27 — with MCP support, a local Neural Engine completion model that never phones home, and agents that run their own tests. As a daily Claude Code user, here's what stands out.
Apple Just Rebuilt the Local AI Stack — and I Can't Wait to Try ItJune 10, 2026 · 9 min read
I've been building a local AI setup around Ollama and a Mac. Now WWDC 2026 just announced Core AI, an open MLXLanguageModel, multi-Mac training over Thunderbolt, and a free fm CLI. Here's the news — and everything I'm excited to try in the coming months.
WWDC 2026: Apple Finally Shipped the AI It Promised — A Developer's BreakdownJune 9, 2026 · 9 min read
Siri rebuilt with Gemini, every OS jumping to 27, a new Core AI framework, free cloud inference for small developers, and Claude Code inside Xcode. Here's my developer's breakdown of WWDC 2026 — with the Apple jargon actually explained.
Microsoft Build 2026, Day 2: Windows Wants to Be the Home for AI AgentsJune 4, 2026 · 5 min read
Day 1 was about the models. Day 2 was about where agents actually live and run — Windows reframed as an agent runtime, and Foundry's Agent Service going GA. As someone who just hand-rolled a tiny agent, this is the part that hit home.
I Built a 300-Line Coding Agent for Local ModelsJune 3, 2026 · 6 min read
After watching Claude Code crawl on a local model, I wanted to understand the 'harness' that makes an AI actually edit files and run code. So I built a tiny one in ~300 lines of dependency-free Python — and the bugs along the way taught me more than the code did.
Microsoft Build 2026, Day 1: Seven New MAI Models and What They Mean for DevelopersJune 3, 2026 · 8 min read
Microsoft just shipped seven of its own MAI models at Build 2026. Here's a developer's breakdown of what each one does, how to actually call them — and the honest answer to the question I kept asking: can I run them locally with Ollama?
How to Point Claude Code at a Local LLM with OllamaMay 20, 2026 · 8 min read
Claude Code is built around Anthropic's models — but as of mid-2026 you can point it at a local model running in Ollama, with no proxy in between. Here's the exact setup, plus an honest take on where local models fall short.
The Navigation Crash That Taught Me About Compose LifecycleFebruary 7, 2026 · 6 min read
How a production crash from users in India led me to discover a critical Compose Navigation pattern that every Android developer should know.