For those of you not familiar with what “vibe coding” is: https://en.wikipedia.org/wiki/Vibe_coding
if you're working on a personal or hobby project, do whatever you want. let your cat sleep on the keyboard, feed that into DeepSeek, and pipe the output directly into `git commit -am 'lol vibes' && git push --force` if you want.
if you're working professionally, by all means use whatever tools you find useful. the only thing that matters is the finished product.
however, if I'm your coworker, that finished product is a pull request that has your name on it. if I'm reviewing it, I expect you to be able to answer questions about the details of the code. after it's merged, it's getting deployed to production, and I'm in an on-call rotation that can wake me up at 2am if it's broken.
if I ask a question about some detail in your PR and your answer is "I dunno why it's that way, the AI wrote the code based on vibes and I ignored the details" I am going to have to have an uncomfortable chat with you and/or your manager.
details matter. ignore them at your peril. the fanciest tape measure in the world does not relieve a carpenter of the need to "measure twice, cut once". an LLM is likewise just a tool you can use. it does not remove the need to think about the details.