[languagewars]

I think they ran into the problem that running the programming competition sites costs a lot less and has little maintenance compared to building and updating the classes. It was doubly difficult since they needed mostly applied classes to have students that were interesting to employers and finished in under 4 years.

[wolfgke]

> think they ran into the problem that running the programming competition sites costs a lot less and has little maintenance

The skills you need for programming competitions are in my opinion quite different from the skills that you need for a "typical" programming job. Programming competitions "usually" mean hacking together a barely working program using ugly tricks that any project manager would strongly frown upon. Also one does not care about maintainability, understandability etc.

Also there a programmers (like me) who love to learn new things (such as in MOOCs), but hate the time pressure and competitiveness of programming competitions and thus never participate.

In other words: I would be very careful to hire people by looking at programming competitions.

> It was doubly difficult since they needed mostly applied classes

On the other hand: This could have been used to bring in some of the cost: If there are employers who would love to hire graduates with knowledge in CurrentHotTechnology, they could sponsor the development of courses for that.

[languagewars]

I don't disagree that they had some students that would be otherwise difficult to reach as recruits, just that they would have had trouble getting enough money per student given that most of the market has low/no costs in providing anything to future employees. The need to provide personal tutoring for inclusiveness was really the nail in the grave for profitably and had nothing to do with that market of recruits.

Reasonably, they had a max of 5k per matched employee which would only be fraction of students that complete a program and even want a new job in one of a handful of places in the US. Competing recruiters would collectively take more than half those students, especially when the students were already in industry and not completing whole tracks.

I also think the sponsored courses backfire as these things are out of date immediately and/or completely irrelevant in most of the market; they probably had to turn most down to prevent alienating students that trust udacity to recommend learning paths.