For example, I can put a static web page up on an Amazon cloud IP and just hit that from my mirror. It will display static data and maybe some things from various APIs, if I code it cleverly enough. But that’s about it.

With the OS working behind the scenes, I can display anything behind my firewall (IP cameras, weather stations, temperature sensors, PIR sensors, etc.). I can also plug in a camera and/or microphone and use the data from those devices in my modules.

It also allows you to scale up if you need to. A Raspi is pretty slick for its size and cost. But if you put something like a NUC behind it, you can use that processing power for some really cool features. Voice recognition, facial recognition, complex animations, etc.

Also, you don’t necessarily have to throw any html or setup a web server. It’s all done automatically at the installation.
Yes, if you want to go forward there’s a lot to learn, but I have contributed / played with a lot of modules already and not written more than 3 lines of html (albeit lots of JS)

The fact that you can have a lot more background services then just with a regular webpage?

Just the fact that you can easily connect the gpio of a pie to a MM2 module.