Hacker News Reader

Comment by mark_l_watson | Hacker News Reader

I have a question perhaps someone knows the answer to: is Udacity concentrating on short, relatively easy classes for mass appeal and Coursera concentrating on more difficult and rigorous classes?

I am currently taking two Coursera classes (one is on probabilistic graphical models and the other on NLP) and so far at least, I am impressed by the depth of material covered and general quality. I will probably need to spend 10 to 15 hours a week on each class.

There are some subjects that I am interested in that I would like just a lighter weight overview, get a lay of the land - I think that this is what Udacity is for.

View on HN

I'm not with Udacity (though I did talk to the team last year before they launched). The difference with the Coursera courses I'm seeing is they're more conventionally lecture-heavy with a few questions mixed in, while the Udacity ones try to exploit the new medium for more interaction and to encourage people to think more on their own. I see no reason that style can't work for harder courses; I don't know if they intend to go that way. (I can say the CEO had lots more in mind to do, and what we see in the courses now is just a start.)

I'm taking the PGM class too and in fact right before I switched to this tab I was stuck on the d-separation question because I hadn't had enough practice on the active-path stuff, apparently. So I'd like to see Coursera work more towards the Udacity interactivity, while keeping their greater polish.

In my experience, Udacity isn't concentrating on short, easy classes. They've simply chosen to start with courses that have larger appeal.

Sebastian Thrun, the professor for CS373, has routinely stated that the caliber of the coursework and lecture material is equal to or beyond what he'd expect from his Stanford classes at the same level.

The teaching model they use is phenomenal though. If you haven't tried it, you should. They do brief 2 to 5 minute videos, followed by questions and coding exercises testing what you just learned. You learn things in very gradual increments, and are immediately forced to apply that knowledge. By the end of the lecture you're left thinking "Of course it works that way."

Udacity is definitely focussed on shorter classes. I wouldn't call them easy though - accessible is a better word.

Anyone is able to take their CS 101 class - no prereqs are required. But it's fairly substantive. There are always forum complaints about how hard some of the homework is.

I think they've picked a good balance of accessibility vs. rigor. But I expect that as they add more classes, people will be able to use Udacity to acquire real skills.

Put another way: The fact that 50% of their current classes are for novices (i.e. one course) doesn't mean the situation will stay that way for long.

And they are very, very good at teaching and online delivery.

As a programming noob, I feel like I learned as much in a day on Udacity as I learned in a month left to my own devices using Khan Academy, The New Boston, and other online materials.

The fact that they show the immediate application of what you've learned goes miles into helping you remember it. To answer the question, they cover a lot in a short span of time and don't waste time teaching concepts that don't have immediate application.

The algorithm course for Coursera looks like it's fairly general in its scope.

I wouldn't call it "light", though, and it's not that good in my personal experience FWIW. Courses will vary, of course, but the attention to presentation in the Norvig video is mind-blowing compared to Coursera's algorithm course whose videos gloss over everything with hastily-scribbled, illegible black on white.

"is Udacity concentrating on short, relatively easy classes for mass appeal and Coursera concentrating on more difficult and rigorous classes?"

Absolutely, your inference hit the nail on the head. Coursera is trying to fill a gap. Very few people across the globe have access to high quality higher education.

Udacity on the other hand, though noble in in their cause, have considerably watered down lectures/assignments for greater accessibility. I can understand the watered down programming class for greater accessibility, since even limited knowledge of programming is something which can greatly empower people who had no knowledge of it before. But advanced classes like robotics are pointless if they don't go into depth of the topic.

I think the sample size is too small to tell (there are more higher level courses from Coursera, different topics, professors), but from what I've seen of the NLP and Machine Learning Coursera courses versus Probabilistic Robotics, I think that Udacity has the better teaching model: short videos immediately followed by a quiz question, rather than Coursera's more traditional PowerPoint approach with maybe 1 question / 6 minutes or so. Thus hours/week is probably a poor metric (FWIW, I feel like I have spent alot of time on the programming assignments in CS373).

I started in the CS373, and having a not so good experience in programming, I found the course very difficult. I mean, the style of teaching is very good. When you program step by step, you learn more and feel more confident.

I think the difference between Coursera and Udacity, is the first one is more focused in lectures and homeworks, and the latter is more focused in practice all the time.

It depends on the course. NLP and PGM are some fo the more hardcore Coursera courses.

Reply on news.ycombinator.com