Slashdot Mirror
School's In For Summer At Udacity
theodp writes "Forget about his self-driving cars. CNN reports that Sebastian Thrun's Udacity — where you and 159,999 fellow classmates can take a free, Stanford-caliber online course together at the same time — just might be the future of higher education. Interestingly, of all the students taking Thrun's AI class globally and at Stanford, the top 410 students were online; the 411th top performer was a Stanford student. 'We just found over 400 people in the world who outperformed the top Stanford student,' Thrun said."
are they factoring in that the online students may have much, much, much... much more free time than a "brick and mortar" student?
Seriously consider the possibility that an in-person student may be taking many classes all at once, with attention diversified versus someone online who may only be taking one class.
As I said, I haven't read the article.
If the classes are good, who cares who's on top or not? The whole bit about other students doing better than an in person one doesn't matter a bit to me. Neither does the whole degree / not degree thing.
What matters is whether there's something really interesting/useful to learn. If you're looking to just get your degree and get out of school and forget learning, well... I suggest you get an MBA. This kind of thing is really great for those of us with a thirst for knowledge and learning that merely got its START when we were in college.
The Digital Sorceress
I took the class and was tied for rank 1, but I have an engineering degree from a Stanford-caliber school and the 411th guy was still an undergrad. It's an interesting statistic, but it's not an apples-to-apples comparison
I have a Stanford MSCS degree from the 1980s. Frankly, the teaching wasn't all that great. Other than Zohar Manna's class on mathematical logic, none of the lecturers had really good presentations. Having the chance to argue with John McCarthy was fun, though. I know things have improved since then. (CS was moved from Arts and Sciences to Engineering and given adult supervision. That helped.)
More recently, I've struggled through the original online Stanford machine learning course (pre-Udacity) starring Andrew Ng. Hacker Dojo offered it as a class, with meetings, two years ago. There he is, writing semi-legible math on a chalkboard (not even a whiteboard) for an hour at a time. The handouts don't quite match the videos, the motivation for much of the math is lacking, and the notation in the field is awful. (Sometimes a subscript is an exponent, and sometimes it's an index, depending on context. The precedence of operators is non-obvious and unstated. And everything, of course, is written with minimal parentheses.) Most of the concepts in that field have a geometrical interpretation, but there weren't enough pictures to give an intuitive understanding of what's the math is doing. What's actually going on is often not that complicated, but you don't get that impression from the lectures.
Some of the big-name universities work only because their students are so good they can make sense out of mediocre instruction. It's really the labs and the other students that make it worthwhile.