Vibe Coding Building Production-Grade Software With GenAI, Chat, Agents, and Beyond

I read Steve’s new book on coding agents. I was interested because agents aren’t yet a net positive for me. Steve seems to have success with his “beads” local issue tracker, so that was promising. And he’s famous from the past. My baseline is AI code completion, which I find very easy to be productive with. I have also read a bunch already on agents, and tried them, with largely mediocre results in the Rust world I inhabit. For senior engineers who already have some framework of how to integrate AI into their routine, I do not recommend the book. I find it's more about visualizing productivity. It acknowledges the ways in which coding agents struggle (no surprise for me). The actual advice in the book is few and far between, maybe useful for me as a blog post, but not spread out over 384 pages. If you want to read such blog posts about the state of the art, may I suggest following Armin Ronacher's journey.