So I can't help but point out some more things...(sorry about being late to the party, I was doing grading for my non-online courses ;^)
1. Thrun is running a for-profit outfit now. He's in it for the money (sad because he did such a good job with Stanford's robotics program)
2. In my opinion/experience, for-profit "universities" generally don't give a rats ass if you learn anything or if you graduate. They, like all for-profit entities, are out to satisfy their shareholders/investors, not do good. (Nope, not even Google, despite their motto) That's why outfits like the University of Phoenix have students with extremely high default rates on student loans.
3. If you look at the numbers from the article, we get...
Stanford AI course - 160,000 enrolled, 23,000 finished (that's only 14.375% - or an attrition rate of 85+%), and 248 or 1% of those who completed the course got 100%.
Machine Learning - 13,000 out of 104,000 completed or 12.5% completed, an 87.5% attrition rate.
Databases - 7,000 out of 92,000 completed or 7.6%. a 92.4% attrition rate.
If I had attrition numbers like that in my courses I'd be out of a job pretty quickly.
4. As a liberal arts professor I have to make the argument that this type of transfer of knowledge is not education but training. Evans from Virginia (another sad story as he's got a really nice textbook on security out) says as much. Training has its place, but sorry, it doesn't make you better able to cope as the world changes around you. (IMHO).
