That said, I can imagine a good number of times it could come in handy, even for the techiest of us.
It might have simply not been the lowest hanging fruit, and the one thing I keep noticing that with AWS is that there are only two types of features they roll out -- new SKUs (e.g., something new that they can charge money for) and services that they should have written a long time ago, but probably didn't, because they were too busy launching new SKUs.
You can react predictively (or predict reactively, whichever way you want to say it.)
Set up a cascade control system, training it as it runs on (process load x cost-of-scale.) It will begin to "see the signs" of load being about to occur, and adjust accordingly to reduce it. You know, just like any modern thermostat.
Even easier, you could just say "Scale it up, we're demoing this to 3,000 users at PyCon", and do that proactively as well -- that said, it's likely done at the expense of efficiency or cost.
As a matter of fact this is why new EC2 instances don't get swap -- running into swap can cost a pretty penny.
You can swap to local (ephemeral storage) at no charge. Or you can created a provisioned IOPS EBS volume and swap all you'd like. There's no separate I/O charge for these volumes.
I assume this is a reference to the charge per 1M I/O requests on EBS.