I haven’t had time to blog in a while — weekends have been very busy. February is also known as VABuary in Sweden: kids get sick, then I get sick, and the cycle continues.
But even through the fog of tissues and cough drops, things have been moving. Fast.
The Openclaw Experiment
Since my last post, I found an old computer of mine and started experimenting with Openclaw. It’s not connected to the Moltbook or anything like that, and I haven’t really given it permissions beyond Telegram and its own GitHub access.
That’s enabled me to do so much, though. Where do I even begin?
Building the Toolbox
After seeing what was possible with Event Modeling and building software with AI, I decided to create some tools to help me with future projects.
With the help of Claude, I was able to first improve my file-based event store so it supported sessions. I experimented a bit with skills too — since converting the event-modeled slices to code was really just four different patterns, I thought it would be good to have four skills specific to implementation. While it succeeded at creating the skills, I found that creating .NET templates was actually pretty easy and probably more accurate. Not to mention more efficient, just thinking about token consumption.
Again with the help of Claude, I created a starter kit using the event store, along with a sample app that had commands, queries, and automations to start with. The next step was to stress-test the starter kit with a new app idea — from event model to production.
Picking ChoreMonkey Back Up
After all that infrastructure building, I wanted the payoff of shipping something again. So I picked ChoreMonkey back up.
I had already implemented some base slices, so next I wanted to see how much feature development I could actually do with just Telegram.
In the beginning, I had it fork my repository and make updates there, but I thought it was unnecessary and added its account as a contributor to my repo directly. Together we set up the deployment so that whenever things are merged to main, I can immediately see the changes.
The Feature Factory
Once the CI/CD was done, we started with the features. Part of the initial setup was having integration tests in place, so with new features I would describe the feature and ask it to create integration tests for the desired behavior. Then it would implement it, deploy the branch, and if all looks good, we merge to main.
I have to say, things have been moving so fast that we also set up a blog for ChoreMonkey. I asked it to regularly update the blog with the latest things it’s been working on.
Bed-Ridden Development
This weekend it was my turn to get sick, so I really got to test out “bed-ridden development.”
I do have to say, it was a pretty productive early morning. We pushed through a list of updates before I even got out of bed. Just me, a fever, and Telegram open on my phone — describing features, reviewing test results, and merging to main. All from under the covers.
And that’s when it really hit me. While AI is a tremendous help as it is, the fact that the system was built with event sourcing is what enabled adding features so quickly. With a traditional CRUD application, every new feature risks touching the data model — migrations, schema changes, the whole dance. With event sourcing, there’s none of that. Just new events or new projections.
The implementation is also very AI-friendly. A state change is a new event. A read model is reading all events so far and translating them. The patterns are so clean and consistent that describing a new feature to Openclaw and having it implement it is almost frictionless.
So I kept going. The entire morning, I was coming up with features just to see how much I could push it before something actually broke. We eventually did — it was a profile status with a null value that started breaking newly created households. It wasn’t that difficult to find the issue, though. And even that felt like a win — the system was predictable enough that debugging was straightforward.
If you want to follow along, here are some links:
The Only Limit Is Imagination
All in all, I have to say, it’s been such a fun time just thinking about a new feature and minutes later being able to see it deployed live.
I love how the technology democratizes app creation. The only real limit now is one’s imagination.
being in tech 
Leave a comment