Hacker News new | past | comments | ask | show | jobs | submit login
A Contrarian View of MITx: What Are We Doing? (mit.edu)
115 points by r_singh on Aug 3, 2014 | hide | past | web | favorite | 40 comments



MIT had $100M to define new tools, new content, and new directions for education, leveraging all the advantages of current technology.

They chose an alternative: record some lectures, and put a slew of course syllabi and homework assignments online. There's nothing necessary bad about what they did, except that it was completely uninspired and non-transformative. From a school that has had such a critical role in pushing the frontiers of technology, it was a disappointment.

A research professor lecturing to a large room of students is an effective way to teach to a room full of students (well, somewhat effective, as many research professors are terrible lecturers). Is it the most effective way to convey concepts in a digital medium? Of course not. It's just a minimal effort, cost-effective approach.


Disappointment is what you get when you do not try. Comparing solutions designed by people that had a benefit of knowing MIT's mistakes is simply wrong. In the end what MIT did was put free knowledge on the internet with their trademark scale and they had done that way before anyone else. I think that inspired a lot of people and without that first step I doubt we would have anything as good as we have right now.


I took (and thanks to how great the course materials were, nearly aced) the first MITx course, 6.002x (circuits). Each lecture sequence had computer-graded exercises, and they had an online schematic editor and circuit simulator that was integrated into some of the homework and tests.

I thought that was an adequate use of digital technology to enhance learning, and I learned significantly more than I did from my failed attempts at taking introductory electronics in a live lecture setting. The ability to pause, rewind, and speed up the lectures was priceless.

Have subsequent MITx and edX courses been less impressive?


I am currently taking the edX Intro to Linux course. I don't actually need the course, but since I am self taught I wanted to see if I missed some things. Popd/pushd come to mind as a feature I've never used.

The course though is rough. I have found frequent spelling and grammar problems. The English is plain broken in the recorded examples. The interactive portions don't attempt to emulate a terminal, it is just straight string matching. Plus, it is amazingly easy even if you are unfamiliar with Linux.

To contrast this Udacity has made some great interactive learning experiences. Their python based courses were very good. The only disappointing this is that they were fairly watered down.

Coursera is in the middle. Some courses are great in the content aspect. Tim Roughgarden's algorithms course comes to mind. But at the same time they are usually a rough port of the classroom material with a slight amount of interactive window dressing.

As I try to pull myself from being a crap programmer to being a good software/hardware developer I am begging for difficult content. Everyone has a bunch of lower division courses in various subject available. I want hard courses that shape my thinking as I try to solve them rather than easy courses that stroke my ego about how "smart" I am.

Also I really wish Udacity and GIT would have opened their 7000$ masters up to people that never finished university. Though I probably wouldn't attempt all the coursework quite yet, I feel shut out of that option. The whole point of pervasive online learning was to remove barriers and friction to learning. By having a arbitrary degree requirement it just takes a step back to the old system that it shouldn't have.


With respect to the Intro to Linux course, I think you probably overestimate the intended audience. I'm not sure the typical Windows user who has seldom or never touched a command line would find it "amazingly easy." That said, the course covers well-plowed land in a pretty pedestrian way. Working through any of dozens of books on the market would probably suffice just as well for the motivated learner.

I agree that most MOOCs seem to be rather watered down relative to what I expect their parent university classes are like. I suspect part of this is that a lot of both problem sets/projects/etc. and outside reading tend to be removed. Some of this don't well with the MOOC format. I also suspect there's a limited appetite for MOOCs requiring many hours of "homework" every week. They do exist but they seem to be in the minority based on what I've seen.


You're thinking of Opencourseware, which is a different project that launched in 2002. Totally unrelated.


I'm honestly not sure what Prof. Flower's view is here. I feel like I'm missing a lot of assumed history, context and perspective when I read this.

I've completed several courses through MITx, perhaps half a dozen total through edX and many more between edX, Coursera and Udacity.

The best content was an all around better experience than anything I've experienced in my time doing online programs as a adult student at an accredited university.

The best MOOC courses I've done are highly produced and thoughtful, particularly relative to the 'Read chapters 8,9,10 and submit these exercises directly from the painfully written $147 textbook.' accredited courses which I'm paying thousands each year for.

Relative to the state of the typical community college or state university online experience edX is revolutionary.

Opting for paid certificates of completion via MOOC wherever possible, I've spent the equivalent of perhaps 1.5 university credits and gained more than the 30 I've accrued over the past year.

MOOCs as a better and cheaper alternative to the current options for non-traditional students would seem like a great thing.

Obviously, there's one big issue, the gap between being educated (what you know) and having an education (what degree you hold).

Working in tech it's easy to forget that the field is pretty unique in valuing the former on par with if not completely in lieu of the latter. The rest of the world doesn't operate that way.

As long as MOOC courses are not accredited or otherwise valued by employers they aren't terribly useful to people who need education most.


> I'm honestly not sure what Prof. Flower's view is here ... The best MOOC courses I've done are highly produced and thoughtful

Keep in mind, this article is 2.5 years old. While there were some good MOOCs at that point, a lot of MITx material was just PDFs of syllabi, homework assignments, and lecture notes posted online.

And to this day, most MOOCs are just a one-to-one mapping of how the professors teach their live class. I think Flowers' view is that MITx had the funding and talent to completely re-think digital education, develop new tools, and raise the bar. They did not do that. What they did wasn't bad, but it wasn't transformative either.


>'I think Flowers' view is that MITx had the funding and talent to completely re-think digital education, develop new tools, and raise the bar. They did not do that. What they did wasn't bad, but it wasn't transformative either.'

I don't disagree with that.

I don't know if making that sort of criticism with the benefit of hindsight is entirely fair though?

I'm thinking that to 'completely re-think' anything is always going to be something of a gamble - with success depending not only on the merits of the solution but specific timing as well.

It's easy to imagine the people behind OCW and later MITx opting for a 'safe' approach as opposed to an all or nothing moonshot. Create measurable results in terms of courses online, visitors and students and iterate from there.

Also, holding up Khan Academy (2006) as what OCW could have or should have been is questionable considering what most of the web looked like 2001/02 when OCW launched. Those intervening years were a pretty big deal in web time.


> Obviously, there's one big issue, the gap between being educated (what you know) and having an education (what degree you hold).

> As long as MOOC courses are not accredited or otherwise valued by employers they aren't terribly useful to people who need education most.

If you watch Chris Dixon's Startup School talk, he talks about how MOOCs are meant to replace, only one of the functions a tradition College is responsible for.

I think Startups should formulate solutions (easier said than done though) to revolutionize other functions of a College as well, like: credentials, networking & collaboration with classmates, on-campus job placements.... while MOOCs could continue to be the Course Material for the education. Something like blended learning with flipped classrooms. This could potentially be a very powerful change in the Education System.


> The best MOOC courses I've done are highly produced and thoughtful...

Great comment. What were the courses that you really liked?


My favorite course so far has been MITx - 15.071x The Analytics Edge on edX. Everything I've seen from MITx has been above average. Andrew Ng's Machine Learning course on Coursera is quite nice as well.


The problem with most MOOCs is that they've taken the part of school that sucks—lectures and homework—and put it online. Inspiration and community are the best parts of school, and they are nowhere to be found.


> Inspiration and community are the best parts of school, and they are nowhere to be found.

Yes. Other things not found:

* Program structure

* Incentive for passing your courses (i.e. getting kicked out of school)

* Brand recognition associated with a diploma, to help you land a good job

* Keg parties

* Networking with classmates, alumni, and, if you're lucky, professors, to help you land a good job

For these reasons, and others, MOOCs in their current form don't pose an immediate threat to colleges.

Still, I think technology could be leveraged to make traditional programs more efficient.

For instance, university professors serve two functions: they spend most of their time and effort doing research, then they also have to deliver calculus/biology/whatever lectures to undergrads. In practice, the overlap between great researchers and great lecturers is often small. Why not supplement undergraduate education using canonical lectures from the very best speakers? I imagine some students are already doing this, by watching Khan Academy or pulling the lectures from MOOCs that feature better speakers than their professors.

Also, as tablet/e-reader screens continue to improve, digital textbooks will likely become more prevalent. There is so much room for improvement of digital textbooks by abandoning the one-to-one mapping from paper textbooks; specifically, by using multimedia more effectively and providing interactive examples. Further, adaptive or interactive problem sets would provide a rich set of data to identify where students get lost. A/B testing could identify which presentations are most effective. There's a lot to be done here.


Getting kicked out of school is something we need to replicate? Heck no.

You either have the motovation to learn or you don't, but the point of online learning is to let people figure that out themselves. We absolutely do not need some ridiculous system of gating access when we won't need it.


I agree that a huge value of a school is learning from your peers, but I personally think homework is one of the most important parts of school also. That's honestly where I learned most of what I did in college.


I agree at least up to a point depending on the course. And, having said that, it's a problem for most MOOCs:

The fact that student are mostly casual and at least shouldn't expect any sort of meaningful certification means that it's hard to expect the sort of time commitment that I put into many of my university courses.

There aren't good mechanisms to evaluate and provide quality feedback for in-depth homework. (And peer review most certainly doesn't count.)


> The “Holy Grail” of true education is exemplified by the professor-doctoral student interaction

I have only known a few people who were at all enamored by the PhD process. Even those who had an overall positive experience speak quite poorly of most of their student-teacher relationships, which seem to range from adversarial to abusive. The best professors are heavily overworked and seem to have little time left for close student interactions. But then all of the people I've worked with who had PhDs left academia, so perhaps it is a case of selection bias.


I had the total opposite experience - I loved my PhD. Getting a PhD is one of the last true apprenticeships. If you put a lot into it and are careful/lucky in your choice of professor it can be a unique opportunity to spend ~5 years digging into an open-ended project at the edge of what people understand while being personally trained by a great scientist. Highly recommend a PhD as an end in itself - it can make you a better, more thoughtful person. Of course YMMV, that was my experience.


I dropped out of academia and my advisor was a self-absorbed prick, but: for concentrated learning there really is no comparison.


I think I may have lucked out, my advisor was absolutely awesome. But, he only has 2-3 doctoral students at a time. I know in large programs, some advisors end up taking on far more than that.


I talked to a graduate college dean about this at a conference we both attended. His assessment is that graduate faculty have abdicated much of their advising responsibilities to the “process”: course work and comps. He wants to see more pure research, zero-coursework PhD programs.

At that same conference, another professor mentioned that she was currently advising 18 students.


no-one 'advises' 18 students... this professor has a successful small business producing publications with their name on them - this is both sufficient, and increasingly all, that is required of them in academia. if their reputation is high enough you can springboard from here into your own academia-small-business start-up YMMV but you can be sure of the up-front costs!


Agreed. In the fields with which I am familiar, professors rarely supervise much more than that number of graduate students over an entire career.

But, in case of the aforementioned professor, she seemed to deplore having that many students, and it was her remark about the number of advisees that prompted the comment from the graduate school dean.


I've learnt a lot on MITx, and I've also benefitted form other MOOC providers. But MIT Prof. Flower's views on OCW & MITx are not wrong.

I would love to know what the people here think about MITx, edX (& other MOOC providers) and how they will affect the future of education.


I've gone through dozens of Coursera classes, 2 edX classes, and two full MIT OCWs. In my experience the OCWs were far superior in terms of actual learning.

A good textbook is actually quite a bit more efficient than Coursera or edX-style lecture videos. The classes on OCW are more rigorous, the prereqs are clear and it's all about the learning. There's no certificate, so the only reason to be there is to learn. Newer MOOCs, on the other hand, are often watered down. Automated graders on programming classes are great, though.


> A good textbook is actually quite a bit more efficient than Coursera or edX-style lecture videos.

I'm not sure they are really alternatives -- most Coursera and EdX courses I've seen have accompanying textbooks or or other assigned readings as part of the syllabus; very few are lecture-only though some list some or all of the readings as optional rather than required (particularly the case with non-free textbooks, though not all courses with non-free textbooks have the readings as optional.)

Most Coursera and EdX courses seem to be structured like college classes, but with the function of classroom lecture and discussion sections replaced with online video and instant-feedback miniquizzes, and the readings still present as readings. Sure, it'll be less valuable for learning if you just do the lecture and projects, just like it would be in a traditional college class if you did that and skipped the readings.


I agree, automated graders with thorough test suites are really great.

Would you mind me contacting you for some feedback/thoughts on a product I'm working on? I think you could help me out since the product has got something to do with programming textbooks, and you prefer textbooks to MOOCs (& have tried both.)

Is the email provided on logicmason yours?


It's probably worth mentioning as well that Woody Flowers ran the 2.70 course for many years at MIT (now 2.007) It was one of the key counterweights to an MIT Mechanical Engineering degree falling completely toward the theoretical side. So his perspective likely is influenced by the fact that academic theory courses only go so far.


What are we doing?

Creating a virtual marketplace for tech/business talent, much in the way the Bloomberg BAT [1] has created a virtual marketplace to connect individuals and firms in the finance world.

Create the marketplace and you're guaranteed both a share of the profits from market-related events (conferences, tuition fees) and an increased brand awareness. This will only work if you have a good reputation that will attract both sides of the market: the students (workers) and the firms (employers). MIT has a great reputation, therefore it can put out basically whatever content it wants, and it will profit.

Some more thoughts:

1. A $7,500/year corporate membership gets you: "-Free postings of company events on the MITX website Events Calendar -Free access to complete contact information of all members -Free company listing in MITX's comprehensive searchable online membership directory..."

2. How much does it cost to record a class, create a couple of self-grading exercises, reuse old powerpoints and PDFS, and then throw the whole bundle up online, where it can be used and reused by millions of people? The cost/user is ridiculously small from MIT's standpoint. Creating in-depth, comprehensive OCW would raise costs substantially. Yes, the students would benefit, but why should MIT have vested interests in students who aren't directly enrolled in the college (besides the whole "education for the good of humanity" argument). MITx students probably won't provide any returns in the form of fame/alumni donations.

3. I don't think this will lower the "prestige once enjoyed by higher education". Going to Harvard summer camp while in high school really only signals the ability to pay for such a privilege. The institution doesn't suffer any loss of reputation and earns cash by licensing out its reputation. I believe it will be the same for MITx.

[1] http://about.bloomberginstitute.com/


>'MITx students probably won't provide any returns in the form of fame/alumni donations.'

That's an interesting thought.

If MITx were to ask me, I would be inclined to donate. As it is, edX does little/nothing in the way of building a sense of belonging among students - just an occasional email promoting upcoming courses.

I expect MITx students could be a resource, however small, but they're currently completely untapped. Every course I've taken has spawned post-course social groups and lists.

This is something the course operators could easily facilitate and use to some advantage.


> 2. How much does it cost to record a class, create a couple of self-grading exercises, reuse old powerpoints and PDFS, and then throw the whole bundle up online

From the article, "We have spent about $40 million over 10 years."

(I removed the part where I got confused by MITx and MITx.org, thanks keithwinstein for clearing that up!)


The $40M quote is actually related to OCW, not MOOCs directly. You could probably argue that MOOCs are, in part, a reaction to the limited impact of OCW.

(Jumpstarting discussions about broadening education, yes, but the actual impact of which was mostly releasing a bunch of course notes--no. Probably not surprising in retrospect. Was it really realistic to expect that teachers in Africa would take all this and create courses for their students?)

I suspect we're well on our way to some sort of crash of this first wave of MOOCs as well. Whether something substantial comes out of it other than a gradual petering out caused by the lack of meaningful returns and results remains to be seen.


Both of you seem to have gotten confused by mitx.org, the "Massachusetts Innovation & Technology Exchange."

That site is not related to MITx, the MIT MOOC offering. MITx was essentially renamed edX (after MIT partnered with Harvard) and is here: https://www.edx.org/

The MIT part (still called MITx) is here: https://www.edx.org/school/mitx

MITx does not have a "$7,500/year corporate membership," etc.


Ah well that certainly invalidates everything I mentioned above. Perhaps edX should offer a marketing and branding course.


I enjoyed this review from 2012, but in 2014 we are past a 2013 "MOOC backlash" and live in an era where lines are well drawn. There are still some educators committed to true distance learning, but some also who have circled the wagons and who decry such "solutionist" thinking.

The really sad thing is that the current battle isn't just about this or that "poor" offering in the present, it is against the whole effort. Do you have an idea for a better way to run distance learning? You too are a "solutionist" and have enemies in place.


Not a lot of specifics on how the author wants to improve the system. Spending $100 million per course is unrealistic and it's not clear what benefit you get from that. Higher resolution 3d graphics?


One specific example he provides is developing better digital textbooks. Specifically, digital content that is more than just a PDF version of a paper textbook. Some options:

* Use multimedia effective (video, animation, sound could all be integrated effectively with the text)

* Interactive examples (see this HN-featured page about Markov models for some nice examples: http://setosa.io/blog/2014/07/26/markov-chains/index.html)

* A/B testing to determine the most effective way to teach a concept

* Mine data from interactive problem sets to see where understanding tends to break down, then iterative on the text until you teach those concept effectively

Another specific example he touches on: better video content. Consider that a lot of MITx's video content was just a recording of a professor lecturing in a monotone at a blackboard. This is the least possible effort they could have made. For a specific example of a better way to lecture online, he references Khan Academy's video. I fully agree. And he points out that not only is this approach effective, it was also inexpensive to produce.


While I don't see how spending $100 million will make a course any better, the price tag is not unrealistic if you truly make the definitive course on a topic that doesn't change frequently. For $1 billion you could have the equivalent of a college major in a single subject. Over a few years, you can certainly find 10 million people willing to pay $100 for a college degree equivalent in mathematics, for example.


kaggle is probably "the MIT" of online learning: - Competition abounds - Those looking to passively absorb will be lost - To the winners go rewards and glory!

It's achilles heel is that it doesn't tap collaboration as well as it could. The future of (elite) online-education belongs to whoever can figure out how to blend competition and collaboration over the medium's limitations.




Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact

Search: