-
What I Learned Using DeepSeek V4 Pro for a Real Coding Task
A practical retrospective from an AI-assisted software delivery experiment I recently tested DeepSeek V4 Pro on a real coding task from my AEIP — Agentic Execution Infrastructure Platform project. This was not a toy example. It was not a “generate a function” prompt. It was a production-style task involving a Rust/Axum backend, PostgreSQL, NATS JetStream,…
-
AI helped me build faster than I could handle, and that changed how I think about software delivery
There is a moment in AI-assisted product building that feels almost magical. You describe a feature, define a direction, set a few boundaries, and suddenly the system starts moving at a speed that would normally require a full senior team. Features appear quickly. Improvements stack on top of each other. The product starts taking shape…
-
3.5 Hours, 6 EPICs, and €13,138.75 of Team-Equivalent Delivery
On March 30, 2026, I used Codex inside my own governance framework, Virtual Company Kernel, to deliver and merge six substantial EPICs in roughly 3.5 hours. Based on the repo’s own role decomposition, that scope maps to about 457 hours of senior team-equivalent labor, or €13,138.75 at €28.75 per hour. There are a lot of…
-
The Missing OS for AI-Driven Development: From Hallucinations to High-Governance.
We’ve all been there: You open your favorite AI agent, paste a complex request, and hope for the best. Instead of a finished feature, you get a “hallucinated” mess that breaks your build. The problem isn’t the AI—it’s the lack of rules. If you want to move beyond hacking on weekend projects and start running…