I spent this evening writing a Java-based Alexa Skill that lets me check the state of my apartment building's laundry room. So now I can say, "Alexa, open Laundry View" and it responds "There are X washers and Y dryers available."
Overall, the process of going from a simple Java app that prints to the console to an AWS-based Speechlet that can run on my local Echo device was surprisingly painless and the APIs are well-documented.
The AWS / Alexa Skill setup documentation provided with Amazon's sample apps, however, is basically "take this UUID on this AWS page, copy it here, compile again, copy another incomprehensible number and put it on some unrelated page, then click these buttons in this order." I have no idea what I just did, but the skill showed up in the Alexa app so I'm a happy camper.