Note: I wrote this in September 2021 somewhere else and I just got around to posting it here.

Remember 10 years ago when everyone was saying that MOOCs would massively disrupt higher education? They obviously haven’t - why? They provide access to lectures and homework from “top tier” universities to anybody with an internet connection, no matter where they live. Many of them provide it for free, but the ones that don’t are vastly cheaper than the ~$5-10K price tag you’d have to pay for a single class at the universities themselves. They have the potential for scaling far, far beyond universities, giving them better economics as well. So why do people keep going to universities? Why are MOOCs still seen as a supplement to institutional education, where most of their successful participants already have degrees or are enrolled in colleges?

What about bootcamps? Programs like General Assembly and Lambda School have been able to train software engineers and place them into high-paying, prestigious companies like Google and Facebook. Their ISA-driven business model - taking a cut of the first year’s salary of their graduates - incentivizes them to actually help their students get jobs, unlike colleges, who simply collect tuition and manage their hedge funds endowments. So why do the vast majority of young people still opt for a Computer Science / Engineering degree instead even though it’s 8x longer and costs 10x more?

If you're in the crypto space, you've probably heard a lot of hype surrounding the educational applications of crypto, especially the "earn as you learn" model. Projects like Earn and RabbitHole originally set out to try to do what MOOCs and coding bootcamps couldn't. However, Earn ended up being bought by Coinbase and was essentially turned into a user acquisition funnel and neither are teaching any of the topics or skills we still rely on universities for (math, computer science, economics, etc). Why?

Why each alternative fails

The way each of the three contestants fail is different. Each failure is indicative of one or more missing pieces, giving us some insight into the question we’re really trying to answer - why are universities still decisively better than all of the internet-age alternatives so far, even after the internet massively disrupted almost every other industry? Even after COVID?

MOOCs

Let’s start with MOOCs. The lectures and homework for most MOOCs are oftentimes verbatim from top-tier programs at top-tier universities, so MOOCs are great at instruction and have the best materials you can have. They can scale beyond belief, so scalability isn’t a problem either. And the economics is great too - they’re oftentimes free, and when they’re not, they’re usually only a few hundred dollars tops.

The failure becomes obvious when you look at the abysmal completion rates - 3.13% across all MOOCs from 2017-18. To put things in perspective, Harvard’s admission rate was 4.6% in 2020. Why are the completion rates so low? Two reasons.

Firstly, in a MOOC, you don’t have office hours or recitations where students have the ability to get direct help from course staff. Having just graduated from college, I can say from experience that the vast majority of the people who were able to get in depend heavily on office hours, recitations, and supplemental instruction / practice sessions to succeed. If they depend on extra interactions with TA’s and Professors outside of lectures, then you can sure as hell bet that the people who aren’t at the university - the very people taking MOOCs - will depend on extra instruction at least as much.

Secondly, as a sort of corollary to the first point, MOOCs lack the social atmosphere that makes people think about the course more often and take it more seriously. Are you more likely to complete an online course if it’s just you and you don’t know anybody else in it, or if you and all of your best friends are taking it, puzzling over problem sets, and struggling through the projects together? Chances are it’s the latter by a longshot.

(It's worth noting that universities also give students a financial incentive to succeed, and that might be a stronger incentive than the social one. But the "I can't waste literally all of my family's savings, I'm already under a mountain of debt" situation is one of the thing's we're trying to avoid anyways.)

A separate but similarly important issue with MOOCs is they can’t enforce the quality and rigor of assignments in the same way as their university counterparts due to their sheer scale - at college, someone actually inspects and grades your homework, allowing them to answer more open-ended, challenging questions on exams (proofs, short answer, essays, etc) than what MOOCs are able to (mostly limited to multiple choice, fill-in-the-blank, and auto-gradable coding questions). Employers know this, so they typically weigh completion certificates from MOOCs below that of actual universities or bootcamps. Thus quality of learning can really only be enforced on an individual level, and the vast majority of people are not disciplined or deliberate enough to do the hard work it takes to really really learn something. On top of that, this kind of learning experience is not very conducive to building practical skills.

In summary, we can collate this into a list of pros and cons:

Pros

• Highly scalable
• Highly accessible
• Very good economics (very cheap per student)
• Can teach almost any topic
• Very good content

Cons

• Very low engagement
• Little to no opportunity to ask questions or for instruction outside lectures / text
• Poor learning quality for most students
• Model is bad at teaching skills

Bootcamps

What about bootcamps? Bootcamps are led by industry experts, and they (the good ones at least) send to be rigorous and project-based. So content quality doesn’t seem to be an issue, nor does learning rigor, unlike MOOCs.

Next, they typically operate in “Cohorts”, where many students all go through the course at the same time and they actively encourage students to get to know one another and help each other, so engagement isn’t nearly as much of a problem as it is for MOOCs.

Finally, bootcamps are typically fully-staffed with people who can answer questions and provide detailed feedback, so human interaction with course staff isn’t a problem either, unlike MOOCs.

There are three reasons they haven’t been as disruptive as we’d like. Firstly, they tend to focus so much on hands-on skills and projects that graduates lack the theory they need to do more “advanced” things that megacorps typically promote people for - thus they tend to struggle a lot more when it comes to upward mobility.

Secondly and more obviously, they’re only focused on programmers. In no way could they have possibly attempted to contend with universities, a one-stop shop for learning anything for anything else - were they meant to. They don’t really teach math, and they definitely don’t teach the physical sciences, social sciences or even liberal arts for that matter.

Just like we did for MOOCs, let’s condense this into a list of pros and cons:

Pros

• Much Better economics than universities
• Good engagement
• Good content
• Good at teaching skills (assuming it’s a good bootcamp)
• Lots of opportunity to ask questions or for instruction outside lectures / text

Cons

• Limited to a very narrow set of topics or skills
• Aren't meant to be an alternative to universities
• Doesn’t teach theory
• Doesn’t Scale

Crypto-Native Education

This category is really broad, but overall, crypto-native education systems tend to have the same properties. We'll look at three examples here - Coinbase Earn, RabbitHole, and 1729.

Coinbase Earn is composed of videos that you have to watch and short quizzes at the end. Pass the quiz, and they put a small amount of crypto into your coinbase account. RabbitHole, in contrast, is a gamified collection of “quests”. For each quest you complete, you gain XP, crypto, and sometimes NFTs as well. 1729 publishes “Contests” during which users are challenged to perform a certain task that’s designed to help them learn something. Then, at the end of a contest, winners are selected and paid crypto.

(It's worth noting that 1729 was founded by Balaji Srinivasan, the same person who founded Earn before it was bought by Coinbase. Hence, a big part of 1729's mission is to make sure that what happened to Earn doesn't happen to 1729. To that end, they've adopted the "contest" model described above. It seems that prize winners tend to be people who already know their stuff, which is a potential problem they'll have to deal with. Nonetheless they're innovating and they're focused on their mission. It's gonna be interesting to see what they can come up with.)

The commonality between all of these platforms is they try to incentivize and gamify the self-learning experience that already exists on the internet. Even if they're not written by MIT professors, there's a virtually endless supply of excellent blogs, books, videos, tutorials, etc on the internet. It's definitely true that anyone with nothing but discipline, motivation, and an internet connection can teach themselves (almost) anything. The reason it doesn't happen often is most people aren't self-motivated enough to do it - hence gamification and incentivization.

If a crypto-native education platform’s content source is the entire internet, then, assuming good curation, in theory its content is on-par with MOOCs - maybe even better. Crypto-native education can also scale just as well, if not better than MOOCs. Since they’re gamified and task based, they can teach skills just as well as theory. And the industry already knows tokens are a great gathering point for a community, so a platform token can help build a strong community with strong engagement. A strong community then becomes a great way to ask questions and coordinate supplemental instruction.

So the crypto-native approach (in theory at least) seems by far the most promising alternative we’ve considered. But the current state-of-the-art has a few challenges they need to solve before they can actually compete with universities.

First, typical completion rewards aren’t worth it for tasks that require significant effort, and the more valuable skills and topics tend to fall into this category.

Second, platforms must decide when a task's completion is satisfactory for a reward. Too selective and the platform excludes people, too permissive and you stop incentivizing rigor. For most skills and topics this decision making process can’t be automated.

Third, Financing the crypto to give to learners is a hard problem. As we mentioned above, RabbitHole and Coinbase Earn ended up pivoting to customer acquisition from education because they needed to make money somehow, and that was the first thing they found.

All in all, here’s our pros and cons list:

Pros

• Highly scalable
• Highly accessible
• Very good content
• Can teach almost any topic
• Good for teaching skills
• Extraordinary economics (students don’t pay to learn, they make money from it)
• Good engagement
• Lost of opportunity for questions and supplemental instruction from humans

Cons

• Rewards alone aren’t enough incentive for people to do the hard work high-quality learning requires
• Deciding what qualifies for a reward (enforcing rigor) is a very difficult unsolved problem.
• Financing the rewards is a very difficult unsolved problem
• Doesn't actually work yet

Comparison

We can collate the pros and cons of all each model into a table:

Now that we can clearly see the differences between each system, the answer to the question “Why don’t we have a good alternative to universities yet?” is clear: all of the alternatives are deficient in a crucial, necessary aspect of education that Universities aren't (or are at least less deficient).

MOOCs struggle with rigor, engagement, and access to course staff - all of which are necessary for a student to become an expert in a difficult field.

Bootcamps don’t teach anything other than coding - they’re not even meant to be an alternative to universities outside web dev, mobile dev, or whatever the bootcamp is for.

And crypto-native models still haven’t found a way to enforce rigor and finance themselves in a sustainable, scalable way, so they aren't even "production ready" yet.

That said, crypto-native systems seem to be the only model with potential to compete with universities without sacrificing too much of its benefits. MOOCs probably can’t improve rigor without losing scalability and consequently accessibility. And bootcamps don’t have breadth by design.

At the same time, the problems with the crypto-native models are, I surmise, solvable with the right combination of mechanism design and community building. If someone can come up with an effective, scalable, and sustainable solution, then that might actually become the university-slayer we've been looking for. But it’s really hard and It’ll probably be a while before anyone manages to pull it off.