It's unfortunate that the solution to cloud pricing complexity that all providers are adopting is – add even more complexity on top.
The number you see on your bill is increasingly calculated by running some black box algorithm on top of the billing events your resources generate. Was it accidental or not? What is a "weird" deployment vs a normal deployment? By what factor should the spikes on your billing graph be smoothened? None of this can be deterministically calculated from the pricing page. And there's no way for you to do these checks yourself before deployment because you have no idea what this logic even is. So you are entirely at the mercy of the provider for "forgiveness".
Who wants to bet that some provider is going to launch "AI cloud billing" within the next year?
The number you see on your bill is increasingly calculated by running some black box algorithm on top of the billing events your resources generate. Was it accidental or not? What is a "weird" deployment vs a normal deployment? By what factor should the spikes on your billing graph be smoothened? None of this can be deterministically calculated from the pricing page. And there's no way for you to do these checks yourself before deployment because you have no idea what this logic even is. So you are entirely at the mercy of the provider for "forgiveness".
Who wants to bet that some provider is going to launch "AI cloud billing" within the next year?