Agreed. His demos are always impressive, but my biggest worry for him is that he's stuck in demo-land. What's sorely missing is guidance (or code) that can help empower more people to create things like this.

But the reality is probably simpler in the sense that complex things are just that: complex, and modeling them is at least as complex, so it will always be hard, and it's only the meta-ideas that can have lasting applicable value. And he's pretty great at disseminating his meta-ideas.

As a non-academic and a web developer, if I were to publish some work of mine, I have always known that I would do it in the form of a webpage. That way, I can make it as dynamic as I need it to be.