The hard part is finding something that people might actually want to do. Geocities worked for this because you actually wanted a Geocities website. It would be difficult to start a company with the assumption that you were going to create some new "thing" that people are going to actually want just for the side-effect of their learning to code.

On the other hand, arguably Wordpress, Scratch[1] and even the Chrome, Apple, etc. app stores have already filled this niche in different domains (PHP, kids, browser plugins, apps).

[1] http://scratch.mit.edu/