9 Popular Online Courses That Are Gone Forever… And How You Can Still Find Them

Online learning has grown tremendously over the past decade, with millions of learners from all over the world participating in free and low-cost courses offered by top universities. Since 2011, over 900 universities have launched more than 13,500 Massive Open Online Courses (MOOCs) on platforms like Coursera, edX, and Udacity.

However, not all of these courses are still available today. According to data from Class Central, the leading MOOC search engine and review site, about 16% of courses (2,200+) have been discontinued and removed from their respective platforms. Some of these lost courses were extremely popular and well-regarded, making their disappearance a significant loss for learners.

Perhaps the most notable example is the Natural Language Processing course offered by Stanford University on Coursera back in 2012. Taught by esteemed computer science professors Dan Jurafsky and Christopher Manning, this course was one of the first MOOCs ever launched, and it quickly became a sensation, with hundreds of thousands of enrollments.

The Natural Language Processing course

The Stanford NLP course provided a comprehensive introduction to the field of natural language processing, covering a wide range of key concepts and techniques used to teach computers to understand and generate human language. Some of the major topics included:

• N-grams and language modeling
• Part-of-speech tagging
• Parsing and generating syntactic structures
• Language models for machine translation
• Information extraction and named entity recognition
• Sentiment analysis and opinion mining

Alongside the video lectures, students completed practical programming assignments in Python that gave them hands-on experience implementing various NLP algorithms and models. The assignments covered real-world tasks like building a part-of-speech tagger, an information extraction system, and a sentiment analyzer.

The course made a significant impact by bringing NLP to a broad audience and making it accessible to anyone with an interest in computer science and AI. It received glowing reviews from learners who praised the engaging teaching style of the instructors and the well-designed curriculum that balanced theory and practice.

Unfortunately, after being offered for a few runs, the Stanford NLP course was discontinued and all the materials were removed from Coursera. For learners who missed the chance to take the course, this was a huge disappointment. An often-cited reason for the course‘s discontinuation was the challenge of providing support and updating the course over time, which became difficult for the instructors to sustain.

However, although the NLP course is lost on Coursera, the core materials are still available online for free. The full set of lecture videos and slides can be found on YouTube and Academic Torrents:

• YouTube playlist: https://www.youtube.com/watch?v=nfoudtpBV68&list=PL6397E4B26D00A269
• Academic Torrents: http://academictorrents.com/details/d2c8f8f1651740520b7dfab23438d89bc8c0c0ab

So while it‘s no longer possible to earn a certificate or participate in the forums, the core educational content of the Natural Language Processing course lives on in the public domain.

Other examples of lost courses

The Stanford NLP course is just one of many once-popular MOOCs that have disappeared over the years. Some other notable examples include:

• A Beginner‘s Guide to Irrational Behavior: Taught by bestselling author and behavioral economist Dan Ariely, this Duke University course explored the irrational side of human decision-making. It was offered from 2013-2014 on Coursera and attracted over 200,000 enrollments.

• CS188.1x: Artificial Intelligence: This highly-acclaimed AI course from UC Berkeley ran on edX from 2012 to 2017. Taught by computer science professor Pieter Abbeel, it dove deep into machine learning, computer vision, and robotics. Many learners described it as the best online course ever made.

• Startup Engineering: Another popular Stanford course, taught by Balaji Srinivasan (then of Counsyl and a16z, now CEO of 21.co). It covered full-stack development and startup engineering concepts. The course only ran once on Coursera in 2013 but had over 100,000 students.

In each case, these courses made a big impact in their respective fields, introducing key ideas to large audiences and pushing the boundaries of what was possible with online education. But due to a variety of factors, including the instructors moving on to other projects, and the challenges of sustaining free courses over many years, they eventually vanished from their original MOOC platforms.

The scale of the MOOC preservation problem

To put the disappearance of courses like Stanford NLP in context, it‘s worth looking at some data on the overall MOOC landscape. According to Class Central‘s analysis:

• The total number of MOOCs has grown from a handful in 2011 to over 13,500 by 2020
• In 2020 alone, over 2,800 new courses were launched – a 70x increase from 2012
• However, the percentage of courses that are no longer running has also steadily increased, from 12% in 2015 to 16% in 2020
• On major platforms like Coursera and edX, the discontinuation rate is even higher, at 25% and 21% respectively

So while MOOC platforms are still rapidly adding new courses, a significant portion are being lost to time. Extrapolating these trends, thousands of courses may be at risk of disappearing forever in the coming years.

The technical challenges of MOOC preservation

As a full-stack developer who has worked on online learning platforms, I can attest to the many challenges involved in keeping MOOCs running sustainably for the long term. It‘s not as simple as posting a folder of videos and PDFs – modern online courses are dynamic, interactive experiences with many moving parts, including:

• Streaming video lectures with closed captioning
• Interactive quizzes and assignments, often with server-side grading
• Discussion forums for peer learning
• Live video sessions and virtual office hours
• Downloadable course materials like slides, code, and datasets
• Learning analytics to track student progress and engagement

Properly archiving all of these assets requires thinking carefully about file formats, hosting solutions, and future compatibility. Videos may need to be converted and stored in open formats, separate from a platform‘s proprietary players. Quizzes and assignments need to be either ported to new platforms or rebuilt with open tools. Communication features like forums are especially difficult to archive in a meaningful way.

Moreover, even if all the raw course materials are saved, there is still the challenge of preserving the actual learner experience as technologies change over time. Platforms are constantly updating their codebases and designs, so a course built five years ago likely won‘t run properly without ongoing maintenance and compatibility updates.

Building online courses for the long haul

Based on my experience as a developer in the MOOC space, I believe there are several best practices that universities and professors should follow to make their online courses more sustainable and resilient to platform changes:

1. Use open, interoperable standards and formats rather than proprietary, vendor-locked file types. For example, use MP4 for videos, Markdown for text content, and standard Javascript libraries rather than custom code. This makes it easier to migrate content between platforms.

2. Maintain an off-platform copy of all course materials, including videos, text content, assignments, and code. Keep these in version-controlled repositories. Don‘t rely on the MOOC platform to be your system of record.

3. Treat the platform as an interface layer for presenting the course, not as the source of truth for content. Build courses in a modular way so that components can be swapped out or rebuilt on new platforms.

4. Have a plan for long-term maintenance and updates, even if the original instructor moves on. This might involve training a TA or finding a co-instructor who can keep things running.

5. Advocate for more open, sustainable practices from the MOOC platforms themselves. This could include providing better tools for exporting course content, offering open APIs, and collaborating on shared content repositories.

By taking a proactive, platform-agnostic approach and planning for the long term, educators can help ensure that their courses remain accessible and impactful online even as the MOOC landscape shifts.

Preserving the MOOC promise

The loss of groundbreaking online courses like Stanford‘s Natural Language Processing is a reminder that, even as digital education has grown massively, there is still much work to do to safeguard the amazing learning resources that have been created.

MOOCs have the potential to make high-quality education from top universities accessible to anyone in the world, regardless of background or means. Losing these courses threatens to undermine the promise and impact of the MOOC movement.

Luckily, in many cases, determined learners can still hunt down the core materials from discontinued courses through channels like YouTube and BitTorrent. For example, the Stanford NLP videos and slides are still out there for anyone who wishes to learn. But this is hardly an ideal solution, as the materials may be hard to find, poorly organized, and missing key features like quizzes and forums.

Going forward, it‘s critical that universities and MOOC platforms invest more in preservation and treat online courses as long-term, sustainable resources rather than ephemeral projects. By adopting sensible best practices and working to keep educational content open, we can fight link rot and ensure that learners continue to benefit from MOOCs for years to come. After all, the whole point of the MOOC movement is to make learning as accessible and flexible as possible.

Despite the challenges and some high-profile losses, the overall story of MOOCs is still an inspiring one. In less than a decade, the number of learners worldwide who have access to free, high-quality courses from renowned universities has grown from zero to many millions. I‘m hopeful that with some intentional effort, we can keep those learning opportunities alive and available forever.