Depends heavily on what you're letting people do, and how long they're going to spend doing it. Games tend to have smaller numbers of "verbs", and longer engagement times - for a superb example of progressive teaching, play through the Portal developer explanation (available after you finish the game).
It's much, much harder to do this with an application where you can't boil the functionality down to a small set of verbs - line of business apps are the usual terrible example of this.
He's probably referring to the idea that much of the complexity around today is not particularly useful - in particular, that programming with values rather than variables makes understanding a program substantially easier.
Simple Made Easy is a talk by Rich Hickey, the author of the Clojure programming language which encodes a lot of these ideas into the language.
We geeks typically place too much importance on the fidelity of the experience: on the idea that the kindle itself feels cheap, or the music from the (standard) iPhone headphones is poor quality. Most people just don't care - they want to listen to the same music as their friends, or get on with reading lots and lots of books. The fidelity of the experience is totally unimportant in 90% of cases.
The "so-and-so read site X" thing will make this really transparent to people, really quickly. Its creepy how much "Friend X has read article Y on the Guardian" is happening.
Well, the fiddly part so far seems to be that's set to a single level (announce/don't announce) and every different app needs its own mute button.
Spotify, for example, will happily announce to the world whatever you happen to be listening to, and there's no simple way to hit "mute". So you're going to need to build a separate "mute" button into every app, on every platform, and have enough trust that it will work ...
While the app's a work in progress (and if you forgive the shameless shill), Splitsies (http://splitsies.net) does more or less the same thing and is freely available in the App Store.
Doesn't do groups yet (because I haven't worked out how to make them awesome yet) but its coming...
We're looking at it. I'll probably go for mobile web first - 1 developer and a backlog of essential work that needs to be done. But I hear the call for Android, its definitely on the urgent list.
(edit to respond to edit)
What do you mean "just does splits"? It doesn't (at the moment) do a good job of explaining that the IOUs balance each other out: if I buy you lunch, then you buy me lunch, we're probably pretty close to balanced out.
I'm working with a great designer at the moment to try and fix some of the programmer-isms in the UI.
Interesting - it looks really similar to my app Splitsies which launched late last year: http://splitsies.net/ - it'll be interesting to see how their dependence on Facebook connect works out for you.
* search: the built-in one takes ages compared to remote search. I understand that it makes the search behaviour the same both locally and remotely, but its still a hassle to have really slow search locally.
* labels with archived items aren't shown in the sidebar. I'd like a reminder that there's unread mail in the mailinglists.
The whole set (in reverse chronological order) is available at http://aphyr.com/tags/jepsen