I'm diving into whether generating infrastructure configurations—like Docker, Terraform, and Kubernetes—using natural language descriptions is still a hassle today. I developed a small prototype that creates full config files from scratch based on these descriptions. It's not a product launch; I'm just gathering insights on whether this approach is genuinely beneficial or whether existing tools and AI models do just fine. If you have a moment, I'd love your feedback on a few points:
- What works or doesn't?
- Does it save you time?
- What's missing or incorrect?
- Would you integrate something like this into your workflow? Any thoughts from DevOps, SRE, or cloud engineers would be super helpful for my research!
2 Answers
These config files are already quite abstract, simplifying the setup process. I really struggle to see how AI could improve that since you need to know the details to get a working setup.
Honestly, I don't think AI is going to replace DevOps anytime soon. Tools like this might feel like a time sink rather than a real solution. If you just need a basic AI to help debug things, sure, but otherwise, not really useful.
True, but AI can handle the boilerplate code for you. You can just verify the actual variables and check if all the parameters are right.