Agentic Engineering

Chapter 1. From Vibe Coding to Agentic Engineering

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 1st chapter of the final book.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

Not long ago, everyone wanted to know whether an AI agent could write working code. It can. The question that matters now sits somewhere less comfortable: once the agent writes the code, who owns everything that happens next?

This isn’t a theoretical question for me. On a typical day, I have between five and ten agents working in parallel, and I max out around five concurrent. Some tasks I delegate fully, because the stopping conditions and constraints are clear. Some I watch closely. Anything touching authentication, security, or money, I watch very closely. That calibration, when to delegate and when to hover, is the subject of this book, and the vocabulary people use for it is still a mess. So this chapter pins down what vibe coding is and isn’t, lays out the spectrum you already live on, explains why the pressure moved to what I call the outer loop, and names the discipline that replaces vibe coding when stakes go up.

One note on how to read it. I’ve tried to write the durable version of each idea, the ...