First Post

So it begins. It’s Friday the 13th and I’m starting to blog.

Good times

Last year Anthropic released Claude Code. And things took off. Software development got much faster. It feels like it happened overnight.

In good hands, coding agents are really productive. In ways I couldn’t have imagined a year ago. AI engineering is here, and it’s going to stay. And I’m here to embrace it.

Do I deliver faster? No. Am I more productive? No, not yet. But I iterate and explore new ideas much quicker. I learn faster.

Working together with AI on software feels great. I don’t need to spend hours on naming things1. I don’t need to endlessly browse Google, StackOverflow, or docs (I can and still do though). I can focus on the part of the development process I like the most: staring at the code, talking about it with someone, and being productive.

I’m having a good time.

The problem

It’s not for free.

I don’t need to challenge myself that often. And since everything is fast now, I can do many things in parallel. I can literally work on three projects while keeping an eye on the Austrian school of AI2 on X.

My attention span is getting out of control. I tend to take shortcuts. I don’t need to go as deep into technical problems anymore3.

I want to fix that. I want to do better. Writing could help me. It could also help me explain complex things more clearly, both to people and to AI.

How

I plan to write here by hand. Like a caveman. I will use AI for spell checks, but that’s it.

I’ve spent a couple of nights building my custom Hugo theme, setting up a GitHub Actions pipeline, and tweaking this page to have a proper (Atom) feed and be accessible to AIs.

I’m already invested. And I do have a couple of things I want to write about.


  1. Martin Fowler: Two Hard Things↩︎

  2. Austrian school of AI: Armin, Peter, and Mario↩︎

  3. Judy H. Shen & Alex Tamkin: How AI Impacts Skill Formation↩︎