Or, if I really wanted to blow up some (personal) cloud service bills, each instance could provision two new instances before the new instances tear down the previous instance 🙃 because what if we fork bomb the cloud?

I wonder how far it would get before someone at AWS calls me to say, "hey Sarai, cut it out"

And the real question here — which part of AWS would call first? My guess is that it would hit IAM limits before it hits compute or API limits, but if I designed it to create new accounts for each instance I could probably hit some AWS Organizations limits