Journal
What we learn building software with AI in the loop, written down so it's useful to someone else.
-
Shipping on-device AI with Apple's Foundation Models
Practical lessons from building journeybot's packing suggestions with Apple Intelligence, on-device. Availability, structured output, prompt size and fallbacks.
-
How an AI-native software house actually works
What changes when AI agents write much of the code, what doesn't, and why senior judgment matters more than ever. A look at how Loam delivers software.
-
Accessibility audits that end in fixes
How to run a WCAG accessibility audit on an e-commerce site so that it leads to shipped fixes, not a spreadsheet nobody opens. With the European Accessibility Act in mind.
Follow along with RSS.