Just another nail in the coffin for Plex. Supporting this skill would take absolutely nothing for Plex as one developer with AI is now more productive then a team of 10 developers previously.
The good thing is that the software moat that big companies like Plex used to have are now gone because of AI.
Expect open-source alternatives like Jellyfin to get a lot better over the next year because of AI. At that point there will be free alternatives that will be way better than Plex.
This extremely disappointing considering how much time and money I’ve invested into my Plex media server. Using it with Alexa is something I do on a routine basis. It makes me seriously consider going to another service. 
I hope the Plex team is at least willing to share some of the code used to build the Alexa skill for some of us to tinker with. It seems beneficial for all to have some access to the api and logic around the feature that many use. Perhaps the “low use” was due to lack of awareness, but I use it in my daily routine.
It sounds pretty simple. I had an AI write me a copy of the skill. It seems that it heavily uses the Plex server API. Your skill must get a certificate to access your server, and it suggests using “Lambdas” to search your library for the requested song and retrieve the URI to the song. Then another to play the song from that URI in the local Alexa device’s audio player.
I think someone here linked to a github for their own implementation. It sounds a bit more involved than my own experiment, I’d have to compare the two at some time…
