The difference is that Khan would be underqualified if he were trying to do something innovative, but he's not. He's just making a bridge technology, a collection of mostly mediocre content that's designed to be moderately useful until something better comes along. There's no way a for-profit could get funded if this were their goal, at least not if they were upfront about it. Plus even if it's not the greatest site ever, at least it serves a purpose and he's not actively trying to scam people.

The lectures are fairly standard, the exercise scheduling system he's attempting to build is not. While I've criticized him for not being ambitious enough (Khan often stresses he doesn't want to replace teachers), he is building far more than just video lectures.

Also, he had several funding offers from VC's, so your assertion about what VC's might fund is incorrect.

John Resig and Craig Silverstein didn't join Khan Academy to generate "mostly mediocre", "moderately useful" content until something better comes along.

I think it's important, especially with startups, to look to the future rather than to sit around judging the present.

In the beginning, Microsoft were making a limited Altair BASIC that only appealed to a handful of hardcore geeks, but later they had a computer in every office and home. You don't get there if you try to solve the world in step 1.