What I'm afraid of in this space is that you're going to have a mix of (a) unaffordable good ones and (b) cheap, less capable people.
It depends on location and opportunities but half-decent freelancers tend to be over $100/hour where I am (New York). People will work at a lower rate for good-citizen reasons (and because they care about the technologies, want to meet new people) but that component is limited and will never get to on-demand service.
That's exactly right. The good-citizen effect works against you, because even the most talented developers with the best intentions to help others won't be as committed or available as the middling developer who actually needs the money. Quality control became an issue almost immediately for us, and managing that component can become a FT job rather quickly. I think this can work, but probably has a better chance as a Kumon-type model where you have much tighter QA over your mentors and can provide a consistent experience to learners.
But teaching will only make the middling developer better. Trust me, this won't be the first time students labor under less-than-perfect teachers. At least these will be motivated.
In our experience, and with companies that have large teams of teachers training is the most important piece of all this.
Pointing out freelancer rates in major cities is also not representative: Lower cost of living cities have lower market rates for freelancers. When you're delivering the service over the web as we do with http://www.thinkful.com/ each party (the freelancer and the hirer) can split the cost savings.
