While his central point was more about developers asking this question to see whether they were creating something because they could or whether it actually served a real need,I was struck by something else he said. In his post he notes that developers need to watch how people are actually using the site.
The reason for this is that in a great many cases what you thought would be the killer aspect that everyone would want to use the site for turns out to be exactly the opposite. Instead they find some other way that - as Mark says - resonates with them.
The best example of this of course has to be Twitter. Twitter was never intended to be what it is today. Where the company saw a mini-CMS application, the users discovered a dead simple real time communication tool and this almost killed Twitter.
The only thing that really saved them was that what they - or rather what the users saw was something totally new and they were willing to put up with all the problems that Twitter had as it made what turned out to be a massive transition. I don't think that something like that will be repeated again as users are being geared to the idea that companies can basically turn on a dime and provide what the users want and have it work.
I also saw this same sort of thing happen on FriendFeed but they had the benefit of the Twitter Experience and very quickly added; or improved, features that the users wanted as they began using the service for something beyond what FriendFeed may have envisioned. Where FriendFeed had thought that a well featured personal lifestream aggregator with a very simple way for people to be able to leave comments on other members posted items, the users found something much more important to them.
What the users discovered was that FriendFeed was another communication community. While FriendFeed had thought the simple comment system would be enough the users suddenly began using the service as a meeting place where they could all hang out with friends and chat up a storm.
This wasn't something I believe that the FriendFeed team believed that their service would be used for; but to their credit when they saw it happening they very quickly started reacting to feature requests that the users were making in order to make the service an even better communication hub.
When the iPod first came on the market it was meant as strictly a music player and iTunes was to be its supplier of those music files. However at some point podcast producers figured that this would be an excellent way to get their podcasts out to a wider audience and the users liked the ease by which they could keep up to date with their favourite podcasts.