Hi👋I'm Vhen

The conversation always seems to be about whether AI will replace us. Sitting here, watching an agent write the bulk of my code, I understand why people ask. But they are looking at it the wrong way.

The real shift isn't about replacement—it's about responsibility. Accountability. As these tools get better, my job is no longer just about writing lines of code. It is about deciding what is actually worth building, checking the work and owning the impact. The agent can write, but it cannot dream. And when the system breaks, the agent doesn't answer for it. I do. I am accountable.

I learned that on a quiet day not too long ago. I had built an agent to watch over our servers. An observability agent. Suddenly, it raised a red alert, claiming three apps were down, then urgently recommended changing a production network setting. It was a drastic fix. Before I let it act, I checked the servers myself. Everything was perfectly fine. The bug wasn't in the infrastructure; it was in the observer itself. The agent had been mistaking successfully finished jobs for critical failures. I left the network alone and fixed the agent. A few days later, when it raised its next alert, it was telling the truth.

That experience solidified a second belief of mine: your tools should belong to you. If I am going to be accountable for what my AI does, I want it running where I can see it, control it, and trust it.

The projects below are what this belief looks like in practice.