Sascha Heyer's Weblog

A Journey into AI-Driven Engineering

August 28, 2025

I write only about 5% of the code myself.

Just three months ago, I asked myself a question: could I be an engineer without writing a single line of code?

Today, my engineering life has been completely transformed. I now use AI agents for almost everything, and I write only about 5% of the code myself.

We are moving from a world of asking, "What code do I write?" to a world of asking, "What problem do I need to solve?" Think about that. You will no longer be limited by the syntax you know, the documentation you read, or the libraries you've memorized. You'll be limited only by your imagination.

The Experiment

To put this new paradigm to the test, I decided to conduct an experiment. I wanted to visualize my own AI-driven engineering process by using an AI to create charts based on my coding activity.

With the help of my AI-driven engineering workflow, we added a series of charts to this blog that track the AI's activity, including:

  • Code Contributions: AI vs. Human: A provocative view of who is doing the work.
  • Top 5 Most Used Tools: A look at the most frequently used tools in my AI's workflow.
  • Tool Call Reliability: A health metric to see how reliable the tools are.
  • Token Usage by Category: An insight into the "cost" of different operations.
  • Most Frequently Edited File Types: A visualization of what kinds of tasks the AI is used for the most.

While I track token usage, it's worth noting that my Gemini Code Assist subscription offers unlimited tokens and is instead limited by the number of requests per 24 hours. Therefore, the most relevant metric for me is the number of requests, which is also visualized in the charts.

All of this data is collected using a custom OpenTelemetry service that I built. You can find the code for it on GitHub:

Loading...

This experiment is a living example of the power of AI-driven engineering. As I continue to work with my AI-driven engineering process, these charts will update in real-time, providing a fascinating look into our collaboration.

I Challenge You

This is more than just a new way to code. It's a new way to think, a new way to create, and a new way to solve problems. The future of engineering is here, and it's a conversation.