Anthropic offered Blender 240.000€ a year in ongoing funding. Four programmer jobs. The community revolted, and three days later the Foundation folded. Here's why it bothers me.

A couple of weeks ago I decided to try to build an RSS reader. I grabbed my notebook, listed every feature I wanted, and started imagining the UI. The code is coming together. The name, however, is proving much harder than I expected.

Last week the platforms I count on for client work failed one after another. Not in some dramatic collapse, but at the exact moments I needed them most.

Anthropic announced a model too capable to release publicly. Project Glasswing is the response, and it cuts right to the heart of who controls the tools the whole internet runs on.

John O'Nolan built a CLI for Ghost on a whim, used it for an hour, and found it hard to go back to the browser. I was thinking about a version of the same problem, but from a different angle.

I've been reading through RSS feeds since 2007. Every platform built on top of the format has disappointed me, but the format itself never has.

There is a version of this story that is clean and motivating. This is not that version, this is the one with imposter syndrome, slow progress, and the strange feeling of being experienced and a complete beginner at the same time.

Or at least, the one that's holding it together right now.

AI coding tools are genuinely impressive. But between Cursor, Claude Code, Codex, and whatever Google launches next, the subscriptions stack up fast. Here's why I'm heading in the opposite direction.

Two weeks of client work, a JavaScript course I've barely started, and nothing to show for it. This is what building in public actually looks like most of the time.