"Datalife streams", Stars and planets... This sounds more like Scientology than a product... I really don't understand all these things about "building a new internet", especially when they are building it on top of the existing one. If the existing internet is so broken what the hell would Urbit do to fix it?
All I can gather from their video is that they're trying the internet again with a different approach (built-in encryption, tools to far more easily build decentralized applications, your data is yours, and a point about API stuff that I don't quite grasp yet.)
Their goal is to be a self-governing digital republic, whatever that might mean.
As the guy says in the end, it's not up to them what urbit will become, it's up to us (the people who develop for urbit).
That's how I understood their video
Soooo... I'm not too sure why building (arguably?) better tools would result in a "better" internet. Additionally, I'm not too sure what the advantages are for end-users for decentralized applications.
So to conclude, it's up to the developers to find valid use-cases for urbit. Just like how it's up to developers to find valid use-cases for ethereum. Decentralized toolchains (like ethereum, other blockchain-things and urbit) are in their infancy, nobody really knows if they'll sprout useful use-cases or not.
Tl;DR Experimentation and uncertainty