I think that the general idea of making it easier to learn something by presenting the user of the service with on-topic posts regularly is useful. Often, when you read things on the internet, you read something really useful too early (before you understand the prerequisites) or too late (when you already learned things another way). Regarding programming, I made the experience that there’s a lot of useful content in stack overflow posts (but also a lot of junk that I’d rather not have wasted time on).
What you propose sounds to me like a take at curated content - and I do believe there’s still lots of room for good services that provide curated content that is actually relevant. Pinterest does a good job for products, apparently, but when I search articles on a given topic with the intent to learn, I still just plain use a search engine.
From the description I’m not sure if your proposition is about curating the new articles that crop up every week or whether it’s about indexing old articles as well (I personally would find the latter more interesting because as a reader I’d like to read the very best articles on a topic - the age of the article matters little to me as long as the content is still relevant today).
I think the challenges with curated content for the purposes of learning might be the following
- ranking the quality and relevance of an article
- discovering relations between articles and categorizing articles (understanding-this-is-prerequisite-for-that, equivalent-to-reading-this) so that you present things in the order that the majority of users finds easier to understand them in. That is, build a graph of excellent content that is ordered in such a way that learning is easier (whether a single “best” such ordering exists is debatable, there might be multiple good ones instead that appeal to the learning styles of different people).
I think that implementing a system that provides people with relevant curated content that is spot-on is, in the general case, a difficult problem. The more you move towards a niche, the more manageable it should become.
http://rightrelevance.com/ tries to solve the problem of finding current relevant content by data-mining Twitter (they seem to rank users on Twitter for a lot of individual topics). They’re a very new service. They don’t address any specific learning-related aspects. They do seem to be trying to address the “find new relevant content as it crops up”-issue, though.