Reusing Payload and Cloudinary to publish these articles.

Comparing how eight local AI models describe the same photos.

Running a local vision model over my photos.

Laying out the same photos four ways, from a grid to a physics simulation.

Building a full-screen photo viewer with a five-mode gesture system.

Wiring up Payload, Neon and Cloudinary to upload photos without code.

Experimenting with my first GLSL shader to run a drifting fog behind the site.

Rebuilding an outdated GitHub Pages site with Next.js and Claude Code.