My guess is that NMS will appeal to people who like exploration. There are infinite worlds to discover and you can tell and show people what you found. Anyone who is hoping for depth in any other aspect will likely be disappointed.
It will be a really cool game if the worlds are unique and complex, so that exploration doesn't become monotonous. Ruins with coherent archaeology would be nice. Maybe some galaxy wide mystery to solve? Where the more planets you explore the more pieces of the puzzle you can put together.
In the videos I've seen, they haven't said much about what you can actually do, or what meaningful goals you could have.
'Things to do' can be very subtle from one game to the next, and game A may do everything right, while game B with almost similar 'activities' fails miserably. The devil is in the details.
To continue on the subject of ED: it hás all the activities, and the actual way these activities work in the game, just doesn't really stick for repeated play. I think Kursah is referring to the same thing; you really need to be able to keep thinking up your own story to make it work. The activities themselves are actually not immersive and interesting enough to stay entertaining on themselves. This ties a lot into the way you 'feel' progression in ED, and it ties a lot into the amount of (or lack of) persistence in the world.
See, I like exploring and the first couple of hours in ED, exploring was awesome. The skyboxes, the sense of size and expansiveness, the incredible number of place to visit... it adds a great deal to immersion. But when you are past that, there has to be something to keep a player in the game. ED lacks that. Exploring is the best example of how ED lacks that depth that is required to keep playing. You can explore places, you can make pictures (this is an out-of-game feature, you make screenshots and post them, so it's an example of 'players creating content to do', not one of game design) and if you are lucky, you find some place that no one has ever been to, you make a surface scan, and you tie your name to the discovery.
.... and that was exploring. Apart from the occasional 'first', exploring offers nothing in terms of actual gameplay. It also doesn't reward the player properly / relative to the time invested. I can't drop a probe or anything to 'do' something with my discovery later on. I can't leave my mark on any place that has already been visited by someone else. I can't share my knowledge with others. My knowledge doesn't 'earn' me anything in terms of mastery of progression in the game.
This example is what plagues ED and procedural games in general. The lack of persistence, lack of progression, lack of a main activity to follow (or story to progress). I like to compare it with LEGO. LEGO is a 'procedurally generated toybox'. You've got building blocks, and every time you use them, you can build and expand on your world, using the very same blocks. But for a new build, I need to tear down what I had built before, unless I have an infinite number of blocks.
Open universe/world procedural games that can offer an infinite number of blocks and allow me to revisit what I've done before, automatically introduce these concepts in a good way: persistence can exist, progression can be made visible.