The Cloud Hasn't Been Won

I agree. I think cloudformation could be there, but it would need a lot more work and support. The documentation is sort of poor, and iterating and developing on it is a huge pain. Terraform is better, but I still find developing on it a really poor experience with AWS. There's a lot of barely documented features, and the feedback loop tends to be slow as well. I also find the IDE support for all these tools to be terrible.

I think there's room for something closer to a compiled language with strong types for deterministic infrastructure setup. It would be nice to catch most if not all issues at "compile time". Bonus too if there's something like Docker Desktop shipping with kubernetes, where I can stand up a replica of my cloud on my machine to debug locally.