My primary question then is this: are these shortcomings enough that such a model should completely leave our consideration as an alternative?

My goal with this is to provide choice to employers and ambitious people, and the projects would be things the corporations want to achieve, don't mind sharing the results of with everyone (think more along the lines of a practical dissertation) and would normally be able to ahieve themselves (and possibly already have a rubric for grading results as these projects must be a normal part of the functioning of such businesses) but do not wish to invest more resources and miss out on discovering new talent simply for a more immediate, guaranteed return on investment.

Also, why not make projects interdisciplinary? The sort of rigorous documentation used for scientific studies could be adapted to the method by which students would be able to make notes and regular progress reports. Additionally, encouraging artists or multimedia focused individuals to make visual or audio documentation of their progress engages more fields in the process and encourages interdisciplinary networking.

I think this idea may be contingent on the development of a much more far-reaching change in the education or possibly corporate models in order to function in the real world, but there are many potential benefits I can see to this.

It seems like testing is the go-to idea among contributors thus far for determining whether a person has achieved the level of proficiency in a field that would be commensurate with earning a diploma from a reputable accredited university, but while I have no data to support the following conjecture, I wonder whether electonic testing is even a valid method of determining anything but memory recall at a specific point or set of points in time, if the testing involves multiple steps (i.e. midterms and final).

Why not use projects commisioned or suggested by interested corporations, involving the use of teamwork/teambuilding, leadership, logistics, creativity, and work ethic, while also providing opportunities for prospective employees - people who may have been using Khan Academy alone for years and have not developed the contacts and overall sense of common academic context college students develop over time - to develop those all-important working relationships. Additionally this would allow employers to have more control over the skillsets they actively seek out, and give self-teaching students an opportunity to understand the kind of skills that will actually get them where they want to go in their careers.

Corporations or individuals would use paid accounts to have the opportunity to work with our teams to determine the kind of project that would most help them find the talents they need, and also help determine the conditions of success.

Projects ideally would have practical applications and real-world effects, and any sucessful projects that end up turning their own profit would have predetermined payout models to distribute income between the patron, our company, and the actual prospects who worked on the project itself.

Students who wished to try for a project would could pay a one-time fee to have a lifetime account, and if possible this fee should be able to be covered by as many forms of reputable student financial aid as possible.

It's 2:00 in the morning and I just got back from Burning Man so I doubt my idea is actually coherent or worth pursuing, but on the off chance it is a good idea, I will just post this now and hope it is productive and promotes thoughtful discussion, if not actual support. That said, if there are any holes in the business model or logic that you post and no one else decides to address them, I will take another crack at it tomorrow.