Now will be some rant and justification for the question, the question itself is in the last paragraph.
At the time (2012-2013) I was very hyped and enthusiastic, imagining new courses of the same quality and kind will be appearing at the same rate in future. By kind I mean university level subjects, fundamental topics that have high return of investment.
But it seems to me it went quite differently. I don't really see any new courses being published and recommended on the lists among good oldies. My own experience with stuff that came later also wasn't as pleasant, I started dropping courses much more often.
The platforms themselves changed significantly: 1. Switch from fundamental to hands-on subjects like technologies and frameworks. 2. Switching to paid model (not that I have anything against it, although being poor in a poor country I avoid paying as much as I can and haven't paid for a single certificate). This leads to less people checking out the course and giving it a "media coverage". 3. Switching from strict start-end date to "take anytime you want". Because of that the social element for me has been essentially lost, the forums are half-dead. 4. Increased amount of courses. Together with (1) this makes it hard to find something by random exploration. When I open edx computer science section I have to go through pages and pages of microsoft courses about their technologies.
This made me gradually lose interest in MOOCs and switch to books and self-learning. I occasionally go back to MOOC platforms or MOOC aggregators, thinking maybe I just missed something or something interesting came out recently.
Did you take any good (valuable, mind-expanding, long-term rewarding, intellectually stimulating) courses that aren't famous and have a high chance of being overlooked?
It's not necessarily a MOOC in the exact same vein, but given that the knowledge you gain from a MOOC is the valuable part (as opposed to any completion certificate or check mark), it's still extremely valuable and a great opportunity to learn.
My first time doing online-learning, and registering for the university, but now I've done it I'm gonna work my way through a few more of the courses.
(You don't need to register, you can just start reading/completing the exercises, and use the telegraph chat-group for help. Mixture of English and Finnish, if you ask questions in English people will use that happily. You only need register if you want official credits, which might be useful in the future for me, but probably not people outside the country)
Plusses:
* neat environment to work with: there's a drone and plane simulator based on Unity3D's Unity Player and NVIDIA PhysX that you connect to from Python, as well as another full simulator implemented with OpenGL (in C++).
* Considerable ground covered: representation, path planning, drone control, estimation, even basics of sensor fusing. Optionally, you can implement a full autopilot for a fixed wing plane.
Neutral:
* All the boilerplate code is provided (Python, C++), and you're asked to fill in the blanks here and there. So, one avoids a lot of hassle. On the other hand, you know, the instructions say "Compute the new variance matrix after the transition using the formula \Sigma_t = G \Sigma_{t-1} GˆT + Q", and then you plug in the line
ekfCov = G * ekfCov * G.transpose() + Q;
Minuses:
* quite some mistakes. The course seems to get little love and few updates or bug fixes. If I had paid for it, I'd be miffed.
* not too much going on in the chat and discussion forum.
As to your points:
1. more "vocational" courses: Yes. Some of it seems more or less designed to quickly train the next generation of code monkeys.
2. paid model: Absolutely. Without the COVID-19 free month, this would have cost EUR 400 per month. Steep.
3. anytime you want: Agreed, the social element was pretty thin. I had the one month deadline before the billing would kick in, so that was useful for self-motivation.
4. more variety: yes, hard to identify good ones, particularly in combination with #2 (pay to study).
I was commenting to a friend the other day that it's insane that we hire 1000s of teachers to teach roughly the same math class 2x per year for 40 years. Instead I'd rather a fraction of those make a truly fantastic video course and then the remaining fraction could be hands on helpers/instructors for those who need help. This is kind of khan academy, but I'd like to see that happen in the public school system.
I really enjoyed this class. It provided me with several very useful mathematical tools to think more clearly about important and frequent problems. Scott is a great teacher.
I mostly use MIT open courseware for self learning. A lot of course websites are also online and if I hear about an interesting course at some University I'll just google the course number. Stanford online and MIT ocw have a lot of lectures on their YouTube channels and I've found some interesting lectures just searching YouTube for a topic. When I wanted assignments and lectures for Russell and Norvig's AI textbook, I googled "Russell AI course" and stumbled on Berkeley's AI course. I've been enjoying it.
How has the social aspect + taking the course with others helped you in the past? I never found it that important. But I have always preferred self learning even in college.
I think it's quite solid and valuable, has a good amount of exercises and I really like the frameworks/technologies they teach (React/Express/TypeScript/GraphQL).
Also trying to expand on my own the topics and concepts they introduce, so I think at the end it's gonna be a very good long-time investment.
For more specific stuff like Algorithms and Data Structures or Programming Languages knowledge, I usually prefer books over MOOCs.
I have let go of edx and moocs and now fish for online university courses which are put up on their websites. Notable ones are MIT, UCB, UC (davis/san diego i don't remember, but one of them puts up all videos as podcasts), CMU (some videos), Brown, etc. Then you have top profs putting up stuff online, like Sedgewick, Pavlo, etc. Personally I feel this is the way forward, pick and choose courses online directly from universities, instead of mooc platforms, for each subject. There is this github repo - https://github.com/prakhar1989/awesome-courses/blob/master/R... which lists out a lot of courses. Occasionally, if you browse enough, you might find some rare links like these courses on database systems - https://bigdata.uni-saarland.de/datenbankenlernen/ and https://www.youtube.com/channel/UCDIJAkbAr53I4fggNsbzdrA/pla...
https://www.executeprogram.com/
Also, Dan Boneh had a YouTube thing up a few months ago where he said that Cryptography II would be coming to Coursera this year, so that will be interesting too.
But i am doing 'mountains 101' during the pandemic https://www.coursera.org/learn/mountains-101
This answer doesn't answer the question that was prompted, but I volunteered for Stanford's Code in Place (introductory CS) and it was kind of amazing getting to be on the other side of the MOOC experience as a TA. I lead 5 assigned class sessions with a small group of people (never more than 10) and held 5 other office hour style sessions as well.
It was a really rewarding experience and I think a lot of people who can volunteer for it should consider doing so if Code in Place continues.
CS Video courses: https://github.com/Developer-Y/cs-video-courses
Math/Science video courses: https://github.com/Developer-Y/math-science-video-lectures
Some of Electrical/Mechanical engineering courses: https://github.com/Developer-Y/engineering-video-courses
If you don't have problem understanding Indian accent, India's NPTEL publishes University level courses on this site: https://nptel.ac.in/course.html
You can take Indian IIT courses online with assignments (and proctored exams if you are in India) at following site: https://swayam.gov.in/