Sorry for the slow response on this one! I've just flown down to New Zealand for a conference, so I was out of contact there for a while. It looks like the load balancer I'm using in front of my instances was erroneously thinking there were no healthy instances for significant periods of time. I've swapped out the load balancer, but that involves a DNS change so it may take a few hours to be visible to everyone. If you need something sooner, http://ec2-23-20-224-56.compute-1.amazonaws.com
should work, though there's no guarantee that will be a valid address after this weekend.