I'm feeling a bit stuck in a paradox here. While I can speed up my development work with tools like AI coding assistants, making the build process incredibly efficient, I find that I'm still having to manually create and maintain Playwright or Detox tests for quality assurance. It feels like I'm leaping into the future of development, but testing is lagging behind. Is there a modern method to streamline end-to-end (E2E) testing? I'm looking to simplify things by using an autonomous agent that can verify critical features like vendor registration and checkout just by stating it in plain English. How are other solo developers and small teams handling this disconnect? Are you using custom scripts, autonomous QA tools like AutoPlaywright, or are we just wrapping the same old manual tests in an AI-friendly package?
2 Answers
Honestly, I've yet to see anyone really achieve a verified 10x speedup using AI. The research shows only about a 5-10% improvement at best. Many claims seem to come from folks who underestimate the time spent just getting the AI set up for their projects. More time is often wasted battling hallucinations and errors rather than just coding from scratch.
I think it's important to note that QA is usually part of the development team. So, while you feel like you're doing the work of the entire team, QA is still crucial. Who’s handling the code reviews in this scenario? That’s something to consider!
Exactly! If there's no code review, you're just churning out potentially buggy software. It might be faster, but it won't be a solid product.