Udacity launching paid course enrollment | Hacker News Reader

Udacity launching paid course enrollment

udacity.com

76 pointsbcjordan52 comments

There are a few things about Udacity that make it different from yet another TeamTreehouse/Tuts+/Pluralsight/Lynda where you pay to get course content:

1) It appears they're sticking with free if you do-it-yourself and charging for coaching and certification.

2) Their pedagogical approach, at least in the few courses that I took about a year ago, has a tightly integrated learn-do loop. And, going beyond just pausing for a mini-quiz, in the programming lessons you stopped to write actual code and attempted to "make this pass the unit tests."

2b) Being presented with a conceptual framework, given a problem that framework could solve, coding your own, then seeing how Peter Norvig would have coded it was one of the best programming learning experiences I've ever had.

3) Their content has, historically, never dipped below first-rate.

The transition to more career oriented skills makes me wonder if #3 might go away, but even if you relax #3, #1 and #2 are better than anything else I'm aware of.

I try to consume as much elearning content as possible (ideally 2x the hours I watch tv), and udacity is in my opinion far and away the best producer.

I have the same concern about #3 and even the viability of the certificate/coaching tier. They need to make a pretty compelling case for the certificates as a valid path to employment.

Though, I'm happy that they managed to find a strategy to monetization without blocking access to existing course material.

Certificates solve a huge issue I've had as an e-learner - getting good credentials for the learning I do electronically.

Prior to this it didn't matter how much you learned on Udacity, there wasn't a good way to signal that achievement/knowledge/learning.

I'm fine overall because I have traditional credentials that speak well enough and open enough doors for me. A udacity certificate would just be icing to add to my resume.

But, the learning Udacity provides is probably most valuable to people who don't have traditional credentials. And, Udacity as an institution is most valuable to the economy/society if we can use it as signals of competence for people without traditional credentials. So I find a lot of reasons to get really excited about this move.

I imagine there's probably a budget/frugal hack that would work where you take the course for free, finish it, then enroll in the credential/coaching for a single month. I wouldn't be at all surprised if one-month-cost post-hoc credentialing is an intended feature of their platform.

yep, it seems built to encourage learning and then pursuing the credentials to guarantee that enrollment stats are good.

What do you mean "guarantee that enrollment stats are good"?

One thing that continues to bother me about Coursera (haven't tried Udacity) is why courses are offered only for a small window of time with stringent due date requirement. They tout themselves as this revolutionary force in education. Then why are they so insistent on continuing traditional university course model?

I wish Coursera would just keep the courses around for people to watch lectures, do exercises and HW.

There's more or less flexibility depending upon the course (speaking about Coursera). Some of the reasons given for having a schedule are: - Provides a structure that a lot of people find useful - Is necessary for any sort of peer grading - Keeps discussion on discussion board in some degree of sync with the class.

If you signup for a class, at least the videos are kept around for some length of time for viewing. That may not be true with grading infrastructures and obviously not peer grading.

It's a tradeoff. I came into Coursera thinking that I'd like more flexibility. But I now see the benefits of a more structured approach so long as deadlines aren't too rigid. After all, if you want something totally unstructured, there is lots of video and other material on most topics on the Internet

> It's a tradeoff. I came into Coursera thinking that I'd like more flexibility. But I now see the benefits of a more structured approach so long as deadlines aren't too rigid. After all, if you want something totally unstructured, there is lots of video and other material on most topics on the Internet.

This is key. I still find myself wishing the deadlines were different here and there (or that they were better distributed... taking 10 courses with everything due in a six hour window Sunday night is annoying and takes being at least a week ahead to avoid). Still, I know that I do better because of it. The peer assessments aren't effective in every course but for some they're absolutely vital.

What I'm really looking forward to is seeing how some courses settle, probably after receiving a little more institutional support, into better defined roles. Instead of just splitting up the sequence in terms of "Part I" and "Part II" they could offer exclusively peer reviewed components for people who wish to write papers and a quiz/homework component which then give different certificates (it would be encouraged to earn both). At the moment EdX seems to be the strongest supporter of human review (either peers or TAs) with Harvard courses such as Copyright Law and National Security, but they can do it because the classes are limited to a few hundred people and everyone else must audit.

Peer review can be pretty good for a number of things when the students doing the reviewing are reasonably competent in the material. They get more problematic as the diversity of the class increases. In my experience, they're also pretty reasonable at grading according to a well-defined and explained rubric (though in a large enough population there will be frustrating outliers). They're much less good at providing useful feedback--a point that seems to be ignored far to often in discussions about peer review.

I'm actually taking a Harvard edX course at the moment that takes an interesting approach. For part of the content (labs) what they do instead of peer grading is a self-assessment with detailed evaluation criteria. This works quite well in this instance. If you don't concern yourself with certification it makes a lot of sense to focus on what provides the best feedback.

Have you taken the Dan Grossman's Programming Languages on Coursera? It gives a very detailed rubric– you're required to complete 3 peer reviews and a self evaluation with the same rubric. That being said the peer review is only worth 10% of your homework grade, the rest is done by the autograder. I haven't gotten a ton of feedback so far– just reviewing other people's solutions is probably more helpful. Still, it's better than some of the other courses. It would be great if there was an incentive for providing meaningful feedback– I don't know if rating your feedback would be taking that too far but it might be more effective than a simple wordcount.

My experience on EdX is limited to MIT and Delft– I'm looking forward to a time when I can try out one of the Harvard classes.

I haven't taken that particular course. I suspect that peer grading could be made better through incentives but, at some point, you're limited by what your peers are capable of in terms of providing feedback. That said, even making the peer review process more explicitly about feedback than grades would be a be a nice framing in the right direction.

I'm taking a HarvardX course on Science and Cooking at the moment which is very well done. Not exactly programming though :-)

I found that the fact of reviewing others is what actually was the most interesting. Not even getting to see good stuff I wouldn't have come up with myself, but also trying to figure out why people have issues understanding the subject at hand.

For more direct and "valuable" feedback, the forums are great. In every course I took there were fantastic discussions on the forums.

(I'm extrapolating most of this negative rant.)

Coursera is founded by bored professors who were tired of seeing their students get rich, so they decided to do a land grab in an "online education 2.0" space. They don't want to revolutionize education. They don't want you to have the option of running open ended courses. They believe in credentials, requirements, sacrifice, and the holy university cram session. They want to keep repeating what they already do make just make it more wide spread (because wide spread translates into money these days).

Have you looked at running your own course on Coursera? They require you have a PhD to even be considered. It probably makes perfect sense in their world, but in the real world, it's just condescending.

Coursera offers some phenomenal courses. If you don't like them, you can teach on Udemy, or skillshare, or a dozen other platforms. Don't b*tch just because they don't fulfill your every requirement.

Right now I'm taking a Coursera class taught by Thaddeus Hogarth. His bio does not list a PhD.

If you are not famous or educated enough to get a spot in the Coursera line up (neither am I), that is nobody's fault. Different strokes. But MOOCs are here to stay, and to revolutionize education.

Can we at least change the name?

STEM (science, tech, engineering, math) started out as SMET, but they rearranged it because SMET sounds like a disease.

MOOCs sound like a cavalcade gullible people with brain damage screaming slurred almost-words from rooftops.

I hear what you're saying. It's better than Large Online Course Offerings, or Classes Over Internet Telephony Usually Superb.

If anyone is unaware, I'd like to note that Udacity does not enforce a course schedule in their free courses. You can begin them at any time and traverse them at any pace you want.

I took Roughgarden Algorithm courses. Due date made me committed. Without due dates, I would just be lazy and might have never finished the courses.. to me it made sense to have students thinking about the assignments and then discussing it on the forums in the allotted time.

Absolutely agree! Without the deadlines, I would not have finished. That said, a 5 weeks course is a lot easier to commit to than a 10 weeks course, since it is very intense during the course. It is simply hard to find the time for it.

I like that Courser split the algo course by Roughgarden into two parts. Made it a lot easier to find the time for it then. Some other Coursera courses are 10 weeks or longer, which so far has kept me from signing up.

I wrote about my experience of both courses: http://henrikwarne.com/2012/05/08/coursera-algorithms-course... and http://henrikwarne.com/2013/02/18/coursera-algorithms-course...

I agree. The nature of my work is very cyclic. I would love to be able to work on courses during down time and put them on hold during busy periods. A coursera course that I was taking is just finishing up now and a few weeks ago work was busy and I feel behind. I'll end up 'failing' the course but at least I did manage to learn from it and hopefully the videos will stay up so I'll be able to finish it on my own schedule. There are also coursera and edx classes I've signed up for with no intention of participating so I have access to the material when I have time. To me, one of the biggest potential benefits of online learning isn't being utilized, everyone being able to work at their own pace.

Interesting to ponder this in conjunction with this Fast Company article on Sebastian Thrun http://www.fastcompany.com/3021473/udacity-sebastian-thrun-u...

It's hard not to see this move as Udacity essentially getting into low touch (but not lowest touch), presumably high quality self-paced adult education that seems to be mostly focused on practical Silicon Valley-type skills and shifting away from trying to "disrupt" education.

Thank you for that link. I was one of the 160,000+ students in Sebastian and Peter's AI class, and it was a blast. It motivated me to sign up for natural language processing at Coursera, another really fun class. The start of the NLP class was delayed, during which time I ended up taking Sebastian's robotic car class, so I'm somewhat familiar by now with his style and personality.

Maybe it's my perspective as a largely self-taught professional, but I always saw Udacity, Coursera, and others simply as continuing education. It was a little strange to see all of the coverage they've gotten as disruptive, threatening to upend higher education. Sebastian fed into this when he gave up tenure at Stanford. He's a type-A personality, a rock star. He can make you an expert in Bayesian probability in one week, write a particle filter "in just five lines of C code!", and truly believes that "now you too can program a robotic car!"

The magic of Udacity was that it bridged the gap between self study and filling a seat in a classroom. Besides the skillfully edited videos, there was value in the grades and the deadlines (in part because it kept the discussion forums in sync), and I think it was a mistake to make all of the classes self paced. Peter said in an interview, "In our online class, we give deadlines. Because if you can watch a lecture online anytime, you can watch it ... tomorrow." The paid enrollment is partly about making money, of course, but it's also about restoring a bit of the external motivation that grades and deadlines used to provide.

I think it's still disruptive if you can access all of the course materials just as you could before.

I'd argue though that MOOCs haven't really been disruptive. If you look at the demographics of a typical course, it's primarily people who already have degrees--and multiple degrees in many cases. So they're adult ed to a large degree. Nothing wrong with that at all. I certainly appreciate them in that vein. But it's not a disrupter of education.

I agree. So far this and others looks more like its disruptive solely within our echo chamber.

What will be disruptive though is if Udacity or one of the others can become a credible source for teaching. All of this means nothing on a larger scale unless it gives you skills and abilities others trust.

thatthatis:

I don't think Khan Academy is typically considered a MOOC--in part because it never really AFAIK considered certification as a goal as Udacity, etc. have. I'm also not sure that simply having a bunch of videos available on programming and related topics qualifies as disruptive. It's not a bad thing certainly! But what is it fundamentally changing? "Disruption" is a high bar.

Do you lump Kahn Academy in with MOOCs when you say they aren't disrupting education?

This is like the gym for the mind, i.e. A great place to get in shape, but many people will sign up with great intentions and quickly fall by the wayside. The cognitive dissonance of admitting failure to save some clams will prevent them from cancelling. Genius!

As you call it gym for the mind, it'd be interesting to see something like netflix for schooling.

What does that mean?

Something that you pay a monthly fee for and gives you access to ton of things. In this case, you have to pay per month per course, and the courses come and go.

The content is free, you pay monthly for the coaching and a certificate once you're done. I don't think the courses come and go. From what I've seen, they've only added to existing courses, and have yet to remove any of them.

extrapolating from my experience of netflix...something that you sit down at intending to do 'just one episode' and unexpectedly spend four hours on?

Isn't that the case for all the youtube education series (khan academy for example)?

This is a fantastic development, and the one-to-one model is perhaps the logical next step after establishing a one-to-many platform for online education. I have a couple of questions regarding the coaching aspect though:

1. Will the coaching group at Udacity be a walled garden, or is Udacity open to exploring a more liberal scheme by allowing experts from around the internet to help and earn some money on the side?

2. How would they maintain a healthy balance between giving constructive feedback and spoon-feeding learners? To elaborate, while I think one-to-one feedback would be great occasionally, I suspect that students might misuse this freedom. Take IRC forums for instance. It can be quite annoying for experts to repeat answers that are just waiting to be read in the docs. From the learner's perspective, it might just be easier to ask someone rather than thinking through the problem and reading the material thoroughly. As a learner, this can be dangerous habit to develop in the long run. So, I'm curious how this works out in real.

But, congratulations and best wishes to Udacity. More and more people, me included, seem to really take to the byte-sized lectures interleaved with exercises. While, the exercises can sometimes tend to be easy (I can perhaps understand why), it's problably not too far away until we have more personalized learning experiences.

It's great that the instructor-student model is becoming increasingly available, but is this really what innovation in education has come to?

Have you taken courses from udacity and their competitors?

Even if this move just gives them enough money to keep producing more content, the world is better for it.

Udacity:Coursera::Community College:Harvard (and yes, I am qualified to make that comparison, I have taken classes from all four)

So what you're saying is:

Udacity:Coursera:Community College:Harvard:?:Profit ?

Oops, that should have read:

Udacity:Coursera::Harvard:Community College.

That's embarrassing.

Could just be me, but seems just another pay-to-learn type service. What differentiates Udacity from other similar educational services?

1) They're still offering free if you go-it-alone 2) Their pedagogical approach is a bit different, with a great learn-do loop 3) Historically their content has been "ivy-worthy"

Quality of courses. Udacity offers very high quality courses, just like Coursera, not like University of Phoenix.

Some platforms will let anyone teach. Skillshare, for example. Which is fine - you get what you pay for, and occasionally you get lucky.

Coursera and Udacity classes are (all?) taught by faculty from name-brand universities. That doesn't guarantee a great experience, but it improves your odds.

As I understand it, it's pay to have a mentor and get a certificate. I think the courses are exactly the same as they were before, and just as free.

Scaling an education venture is a hard problem. You work really hard to figure out how to distribute quality content cheaply to lots of people and it turns out that, at the end of the day, it seems that we still rely on labor/cost intensive methods to teach. [1]

Clearly, Udacity, edX, Khan, etc are successfully solving many problems (quality content, automated grading, gamification, etc) and that we're moving in the right direction.

I've very much looking forward to the day were we start to see education costs decline because we've gotten so much better at mass producing knowledge/love for learning/reasoning skills/etc.

[1] As a side note, it's not clear to me if actually reasoning and talking with another human is really required for learning. It certainly seems that way at the moment, but is this because the content isn't good enough or are our brains built in a way that learning requires human interaction?

> it's not clear to me if actually reasoning and talking with another human is really required for learning.

Dialog. Being able to ask questions, to get clarity on some angle that the teacher took for granted, uncovering assumptions / knowledge that the teacher has that the student doesn't.

Quite the opposite, deadlines made me cram and guess the multiple choice since 1/2 weeks per quiz was a ridiculously small amount of time to learn. I was taking both Coursera Computer Cryptography/Networks. If there was no deadline I could absorb the material at my own pace. In regards to marks, fuck 'em. I want to learn, not to get "marks" or a "paper".

Am I crazy or do they not actually mention anywhere what the fee is?

It varies by course, so you'll need to look through the catalog. The prices are monthly, and all courses are "at your own pace", but they give an estimate for each course of how long the course will take.

https://www.udacity.com/courses

Intro to Hadoop and MapReduce costs $150/month (or $105/month early enrollment) for a course that is estimated to take a month.

https://www.udacity.com/course/ud617

Exploratory Data Analysis costs $150/month (or $105/month early enrollment) and is estimated to take two months to complete.

https://www.udacity.com/course/ud651

Data Wrangling with MongoDB is $200/month ($140 early enrollment) for an estimated two months.

https://www.udacity.com/course/ud032

Just visit any of the paid course pages. It's $150/month.

I like Udacity, I did... If they want to monetize some of their content, good, but I wont be participating.

What? This is a pretty great way to monetize without locking down the content itself.

Reply on news.ycombinator.com