A Read-Scale Availability Group is a Clusterless Availability Group. It’s sole purpose and design is to scale out a read workload. More importantly is what it is not. It is NOT a High Availability or Disaster Recovery solution. Since this design has no cluster under it, you lose things like automatic failover and database level health detection. For example, You have reports that run for customers that are in your DMZ that is fire-walled off from your internal network. Opening up ports for Active Directory so that you can have a cluster means opening a ton of ephemeral ports and ports with high attack vectors. Remember the Slammer worm? This solution removes those dependencies.
Click through for the setup scripts as well as a video Ryan created of him putting it all together. As long as you recognize the trade-offs involved, this can be a nice solution to certain problems.