AWS CloudFront is already very cheap and easy to get setup, probably less than 5 mins. No reason to go through a 3rd party. CloudFlare is even easier, better and completely free.
I don't see why this is worth the risk (or IMO, the trouble).
And if it doesnt stay cheap for you, then why would it for this guy? He even says to keep the traffic small in which case paying for your own CloudFront distribution with the security, performance and control it brings is far better.
Ease of use and price I don't think are the most important issues when you the most important issue is reliability and how do I know you won't shut this down because too many people use your service .
So why give props and kudos? It a very bald idea but (and I am talking from my own experience) when offering free services like this; when they really get picked up in no time you can not afford to keep them up unless you start to implement a paid service. In this case that would mean just reselling CloudFront one-on-one. So, cool for now but not really maintainable if you ask me.
In theory there is an arbitrage opportunity here, given that with enough users on the service Jare could be paying for most of the bandwidth at higher tiers ($0.06 - $0.02) and solo the users would likely be in the lower tiers ($0.085 - $0.08).
However, I suspect without an AWS Partner agreement in place, they will be skating on thin ice.
I don't see why this is worth the risk (or IMO, the trouble).