This makes multi turn interactions much faster and is an awesome feature. Ahmed Bouzid has a great video demo here:
Voice is all about speed and this mode helps make the interactions smoother. For the interim, Amazon is making you enable it on every device but given how much in can improve the interaction, it really should be default.
It’s fantastic that Amazon continues to put out these tools. Another really cool set of tools is the Alexa Skills Audio Library — a collection of pre-made sounds for Skills. These can use the SSML tags (and the Ubi Portal Skill is great for this).