๐ฆ๐ถ๐ ๐ ๐ผ๐ป๐๐ต๐ ๐ข๐ณ ๐๐ ๐ช๐ฟ๐ถ๐๐ถ๐ป๐ด ๐ ๐ ๐ง๐ฒ๐๐๐
I stopped writing most of my test code by hand. I use AI now. My output grew. I shipped tools I never had time for before. But my job changed.
AI handles the boring parts. It reads tickets. It drafts plans. It makes Postman collections and Cypress specs.
I still review and run everything. I validate the AI work against the real system.
I built new tools:
- A Postman to Cypress converter.
- A flakiness dashboard for test stability.
- A company plugin marketplace.
- An agent for npm audit sweeps.
Reviewing code is harder than writing it. AI produces code fast. It looks clean. It looks right. Often, it is wrong.
It checks the wrong thing while looking correct. It drifts from project rules. It makes a mess if you stop paying attention.
AI amplifies your skill. Good judgment makes you faster. Poor judgment makes more problems.
The bottleneck moved. It is no longer about writing code. It is about knowing if the code is right.
If you are new, do not rely on AI alone. You need to break things first. You need to debug by hand. This builds the judgment you need to review AI work.
I am now a designer and verifier. I spend less time typing. I spend more time thinking.
Source: https://dev.to/adrianjiga/six-months-of-ai-writing-my-tests-what-got-better-what-got-worse-32li Optional learning community: https://t.me/GyaanSetuAi