Massive open online courses turned fifteen this year. Coursera launched in 2012, Udacity the same year, edX in 2013. Back then, the rhetoric was that they would democratize higher education and put pressure on tuition by offering Stanford, MIT, and Harvard lectures to anyone with a laptop. A few years later the reporting turned to completion rates, and the verdict was harsh. Most learners signed up and quit. The headline number that traveled was that 5 to 15 percent of enrolled students finished a given course.
We never thought that was the right metric, and we still don’t. Most people who registered for a Stanford machine learning course in 2013 weren’t pursuing a credential. They were curious, or hunting for one specific lecture, or testing out whether the topic was for them. Judging that population by college-style graduation rates was a category error from the start.
What’s interesting fifteen years in is which corners of the original promise actually delivered, and which collapsed. Worth taking stock.
Where MOOCs found a real audience
Two pieces of the original vision held up. The first is technical skill credentials with a clear job pipeline. Google’s IT certificate on Coursera, the Meta and IBM tracks, the AWS courses on platforms like A Cloud Guru and Coursera, the Microsoft Learn material. These are now part of how working adults change tracks. They don’t replace a CS degree, but they aren’t trying to. They credential a specific competency for a specific employer-recognized purpose, and the format (video plus assessments plus a verifiable certificate) suits that goal cleanly.
The second is professional continuing education that used to be done by expensive in-person workshops. Project management, data analysis, UX, the whole adjacent-skill ladder, moved online and didn’t move back. Lots of accountants, marketers, and engineers are now picking up the next thing through self-paced video plus a forum, paid for by their employer or written off on their taxes.
Where it didn’t
The free-or-cheap liberal arts vision largely petered out. The Stanford intro philosophy course you could take in 2013 isn’t really there anymore in the same form, partly because the platforms moved upmarket toward credentials people pay for, and partly because the audience for a free philosophy lecture turns out to be small and disinclined to grind through a video plus quizzes structure. YouTube ate that audience, with shorter formats, better creators, and no enrollment funnel.
The “MOOC will replace college” version of the prediction was always going to be wrong. College is partly content, but it’s mostly a sorting mechanism, a credential, and a four-year residential social structure. None of those translate to free Coursera lectures.
What actually replaced the rest of the promise
A few things, none of them branded as MOOCs.
Bootcamps. Twelve to twenty-four-week intensive programs in software engineering, data science, and UX, often with a job placement guarantee and a tuition share or income share agreement. Lambda School (now Bloom Tech) had a bumpy ride, but the format itself is alive. The good ones place graduates into engineering jobs at decent rates, and employers in 2026 will hire from them more readily than they did in 2018.
Niche skill platforms. Maven for cohort-based courses run by individual practitioners. Newline and Frontend Masters for working frontend engineers. Brilliant for math and science. Each of these is a smaller, sharper version of what generic Coursera tried to be.
Employer-delivered learning. The biggest shift is that companies have built or licensed serious internal learning systems. Workday, LinkedIn Learning at scale, and a long tail of internal academies. For most working adults, the way they pick up a new tool isn’t to enroll in a course on the open web; it’s to watch the internal video that HR pushed to them.
YouTube. We keep coming back to it. As of 2026, YouTube is the largest open educational resource on the planet by a wide margin, and it isn’t close. The signal-to-noise ratio is messy, but there is a creator on YouTube for almost any subject at almost any level, and the audience for the long format has gotten more patient, not less.
What this means if you’re picking a platform now
Three rough rules of thumb we’d offer.
If you want a credential that an employer will recognize, look at the partnerships first. Google, Meta, IBM, AWS, and a handful of universities running master’s-level Coursera programs have built actual hiring pipelines. Most other certificates are decorative.
If you want to learn a specific skill quickly, look for a small platform run by people who do the thing for a living. Maven cohorts, Frontend Masters tracks, the better Udemy instructors. The brand of the platform matters less than the credibility of the individual teacher.
If you want to learn for its own sake, give yourself permission to use YouTube and library books and not feel like you should be paying for it. The value of free learning material is genuinely higher than the marketing of paid platforms suggests.
The MOOC bet of 2012 wasn’t quite the revolution it was sold as. But the world it pointed toward, with adults reskilling fluidly, credentials decoupled from four-year programs, a long tail of teachers reaching learners they never would have reached, has mostly arrived. It just looks different than the press releases promised.