Field notes from the part after the first prompt.
Hey, I'm John.
Software engineer and Vibejar editor · Sydney, Australia
I've worked as a software engineer on Insight Timer and Clove. I use coding agents every day, and most of the hard work starts after they have generated something plausible.
You still have to choose the right tool, find the real cause, test the path a user takes, and decide whether the result is safe to ship.
Vibejar is where I write down what actually helps with that work.
One useful decision at a time.
Name the decision
Start with a real debugging, testing, or tool-choice decision that keeps coming up while shipping software.
Check the evidence
Use the tool, read the primary sources, and date any price or capability that can change.
Write the useful answer
Explain where each option wins, where it loses, and what the reader should do next.
Keep it honest
Update or retire pages when the underlying facts stop being true. Search traffic is not a reason to bluff.
Vibejar does not sell a product today. The former capture app is retired. This publication and the email list are the whole thing for now.
I write the guides here.
The guides and comparisons on this site are written by me. They come from using the tools and from published primary sources. I don't invent benchmarks.
If this sounds familiar, I'd like to hear from you.
Send me the question you keep running into while building with an agent. It may become the next field note.