Your experience with AI (vibe) engineering mirrors mine remarkably closely. The one thing I think you allude to without mentioning specifically is that I'm having remarkably good success when I tell the AI agent to use Test Driven Development methodology in a very strong way. With an existing test suite, and an approach where it writes tests first, runs them to who failure, and then implements the code to make the tests pass really allows the AI work on a harder problem and debug its own mistakes more before I have to review what it's done. I can review the tests before it starts too, which simplifies the acceptance work some at the end..
Shared this with my team and felt it really resonated with how we’re programming in our day to day now. We’re almost all using Claude and some Cursor, but have increased our emphasis on code review. It’s sped us up a bit, but not so much that we aren’t really confident in what we’re putting out!
Hi Simon,I put together a simple SaaS tool that lets anyone spin up their own community space in just a few clicks. It’s free for now, so feel free to give it a try and let me know what you think!
Love "Getting good results out of a coding agent feels uncomfortably close to getting good results out of a human collaborator". Amplified even more in the last 24 hours with Opus 4.6 in Claude Code.
Your experience with AI (vibe) engineering mirrors mine remarkably closely. The one thing I think you allude to without mentioning specifically is that I'm having remarkably good success when I tell the AI agent to use Test Driven Development methodology in a very strong way. With an existing test suite, and an approach where it writes tests first, runs them to who failure, and then implements the code to make the tests pass really allows the AI work on a harder problem and debug its own mistakes more before I have to review what it's done. I can review the tests before it starts too, which simplifies the acceptance work some at the end..
Shared this with my team and felt it really resonated with how we’re programming in our day to day now. We’re almost all using Claude and some Cursor, but have increased our emphasis on code review. It’s sped us up a bit, but not so much that we aren’t really confident in what we’re putting out!
Hi Simon,I put together a simple SaaS tool that lets anyone spin up their own community space in just a few clicks. It’s free for now, so feel free to give it a try and let me know what you think!
So far the best names imo are like ai-forward engineering. Disagree with vibe but worth tinkering on a name yes
Sound concept, not the best name (“vibe” con notates nonseriousness, not knowing what you're doing)
The faster we go, the more important safety measures become.
Love "Getting good results out of a coding agent feels uncomfortably close to getting good results out of a human collaborator". Amplified even more in the last 24 hours with Opus 4.6 in Claude Code.
What about "code-by-agent" like fly-by-wire in airplanes? Or simply "Agentic Engineering".
Interestingly purposeful