It's using existing mechanisms in YouTube's JSON API to find suggestions and play them. All the true "magic" is provided by YouTube's engineers. Not to take it away from someone who writes the right lines of JavaScript at the right time, but you don't have to place the guy on a pedestal.
The real lesson here is for yourself and others here who think like you -- if you tell yourself that things like this take superhuman efforts, how will you ever be able to produce something creative yourself?
Nothing is more enjoyable than seeing posts on the internet by programmers who look at an idea, then scoff at its simplicity while making a list of all the idea's component parts, pointing out how easy each component is.
But this is an easy idea, and it's been done before. The only thing that's really different here is the timing of the announcement. And the "omg chad hurley reads hacker news" angle.
(Of course, everyone thinking that this is even remotely the same thing as Google Instant is also completely missing the real innovations in that product.)
And I’m sure they'll give him time to work it out further, but is this the innovation YouTube is looking for? Hard to believe their UX people and dev teams haven’t already toyed with something like this.
He threw something together that people could get a feel for. I'll concede that it was hardly usable given the performance. But it was only a demo, enough of one to show just how usable the real thing could be.
