Working with AI, written down honestly
I spent my career writing compilers. Now I'm retired, and I'm finding out what an AI assistant can really do, one real project at a time.
This site is a notebook. When I use AI for something real, like my home network, my backups, a website or a German lesson, I write down what happened: what worked, what surprised me, and what went wrong.
I'm not selling anything. I've been around computers long enough to distrust hype in both directions, so I'll tell you the parts that impressed me and the parts where I had to catch a mistake.
Posts
- The day my AI unplugged itself
Claude was going to finish switching my house to a new router. It missed one thing: it needed the old router to talk to me. Here's what happened, and why the second try worked.
- Two weeks, one network
My home network ran happily for eight years with little attention. Then I spent two weeks modernizing it with Claude. Here's what we did and what I learned.
- Mistakes AI made
Claude was right far more often than not. But it made mistakes, some of them stated with total confidence. Here's the list, and how each one got caught.
- Backups you can trust
A backup you've never restored is a hope, not a backup. Here's how we set up offsite backups for every machine in the house, and then proved they work, down to the last byte.