The Two-Tier Future of Learning

Folu Ogunyeye (MBA '27) on the starkly stratified future of education, learning and reskilling

A few weeks ago, I sat in on a fireside chat with the CEO of Coursera, hosted by the HBS Education Club and the Harvard Project on Workforce. He knew his numbers: fifteen learners enroll in an AI course on Coursera every minute. If the world were 100 people, 59 of them will need retraining by 2030, according to the World Economic Forum. In response, Coursera has stepped up to the task by partnering with governments to upskill entire populations. Degrees are important, he said, but no longer sufficient to stand out in the age of AI automation. Micro-credentials are therefore the panacea in getting ahead of a monumental epidemic of redundancies and existential crises.

And then, as an aside: This is what I tell my son, who's in college.

That sidenote struck me. His son is in college, presumably on a 4-year course. The micro-credential is the supplement. Yet for much of the 59 out of 100 who need retraining in the near-future, the micro-credential is increasingly being pitched as a substitute. 

When it comes to telling stories about the future of learning, a quieter tale lurks in the shadows. It is the story of rapidly unfolding stratification in what education was, what it is becoming, and who experiences which version.

What it once was

The oldest surviving arguments about education, from Socrates to Aristotle to the Montessori classrooms now fashionable among well-to-do American parents, share a common premise. Education is not the transfer of information; it is the shaping of a person capable of judgment. This kind of education is slow and expensive to deliver. It requires small classrooms, patient teachers, and time spent on things that do not obviously pay off. Historically, there has never been enough of it to go around, and the question of who gets it has almost always reflected who holds power. Socratic dialogue was for noble men in Athens; Oxbridge tutorials were for aristocratic sons; Montessori kindergartens are for families who can afford the fees. What is new about our moment is that the problem of scarcity in quality schooling is now being dressed up as solvable through the easily accessible utopia of online learning. Readers of my earlier piece in The Harbus may recognize the underlying logic. I have written before about impact neutering: the sterilization of social impact into an abstraction that lets institutions feel virtuous without changing very much in terms of core power dynamics. 

In the most salient takeaway of the afternoon, the Coursera CEO remarked: "Information is not the same as knowledge, and knowledge is not the same as skill." He meant it as a pitch for Coursera's move toward learning-by-doing. I heard it as an acknowledgement that the three are genuinely different, yet the vast majority of edtech platforms that claim to democratize access to learning (Coursera included) are much better at the first than the third.

What it is becoming

A few weeks ago, the CEO of Cloudflare visited an RC classroom and offered a sobering diagnosis. Many of us, he suggested, belong to a lost generation: too old to be AI-native, too junior to have accumulated the seniority that would make us hard to replace. You could feel the room recalibrate their life choices in that moment.

I do not believe him. The underlying threat is real, but his diagnosis to a room of MBA candidates assumes each of us is facing it alone. We are not. In the weeks since that class, I have been to four student-led vibe coding and agentic AI workshops, each one led by classmates with solid tech track records, motivated by the desire to see the rest of us not fall behind. I picked up more in those evenings than I would have in a year of trying to teach myself in the nonprofit sector, where I have spent the majority of my career. The tweak I would make to the CEO’s foreshadowing is to underscore the power of having access to the rooms where people at the forefront of societal shifts are quietly sharing notes on what to do to get ahead.

The response from industry, governments, and education platforms has been to build out a rapidly-expanding reskilling economy designed to train people more efficiently and affordably than a traditional degree ever could. Bootcamps, online certificates and other goods on sale at the ever-growing market in stackable credentials are all pitching themselves as the modern, democratic alternative to a four-year degree. In theory, this is progress.

The trouble is what these ‘future-forward’ interventions are replacing, and for whom.

For a long time, a quality education has been one of the most reliable engines of economic mobility we have. Not a perfect one, but the data on lifetime earnings and access to professional networks has been consistent enough that “get your kids into a good school” has functioned as folk wisdom across almost every community I know as a Nigerian-British-American. Education worked as a mobility engine because it did two things at once: it trained people for jobs while shaping them into the kind of people who could navigate institutions and advocate for themselves. 

What I hear in the current conversation about the future of learning is a separation of those two functions. The shaping part is becoming a luxury good – think small classrooms, residential experience, seminars, the slow formation of judgment. The training part is being democratised on the frontend while priced to scale in the backend boardrooms. 

For the average global citizen, the second half is being pitched as the future of education. And in fairness, it often is more accessible than what came before. But we should be honest about what is being democratised. It is access to employability for a specific point in time, not access to an education. Most importantly, the two sides of the crossroads do not lead to the same long-term outcomes.

This matters because the employability-oriented credential is a bet on a specific skill holding its value in a labour market that is, by the admission of the people selling the credentials, changing faster than anyone can predict. The Coursera CEO was candid about the retraining imperative. He did not dwell on the fact that the people needing to be retrained the most are also the people with the least room to absorb the cost of being wrong about which skill to bet on. A four-year residential education is, among other things, insurance against obsolescence and against the narrowing of options that follows from being trained for one thing only. 

This is where the credential model starts to look less like insurance and more like a subscription. How many courses will the average worker have to take to stay ahead of the next wave of layoffs? Think about the lifetime process flow of a worker under this model. The micro-credentials they are stacking now will not necessarily hold their value two years from now, let alone five or ten. So they will have to keep buying. What the credential economy quietly constructs is a steady stream of returning customers from the very demographics least able to absorb the cost of paying. It is an elegant piece of business design – in a Marketing class we’d call it ‘stickiness’ or ‘low churn’. It is also a rather damning piece of social architecture: those who can afford a one-off, four-year investment get a lifetime return on it, while those who cannot are charged, over and over again, for the privilege of remaining employable.

A closer look, closer to home

This phenomenon is unfolding inside Harvard Business School too.

Last semester, an anonymous writer in The Harbus published a sharp piece titled The Prestige Paradox of HBS Online, arguing that HBS Online dilutes the brand equity of the residential MBA by issuing credentials at scale and at a fraction of the price. The numbers are striking: in 2020 alone, HBS Online issued roughly forty thousand credentials against nine hundred MBAs, with acceptance rates near forty percent against the MBA's single digits. The writer's recommendation was to protect the crest. If HBS truly wants to democratise knowledge, drop the credential and make the content free.

I agreed with every observation in that piece yet almost none of the conclusions. Once again, a question lurks in the shadows: of the forty thousand people who pay to take an HBS Online course, how many would have chosen the residential MBA if it had been available to them? Not all, but some meaningful fraction: people with jobs they cannot leave, families they cannot uproot, savings that cannot absorb two years without income, or passports that make jumping through visa hoops impossible. For them, HBS Online is not a different product. It is the only version of HBS they can access. And the version they can access is, by design, the thinner one: asynchronous modules, no familiar whiff of a classroom, no professors who will remember their name in ten years, no network that converts into a career.

HBS Online makes visible what was always true — that the highest ROI version of this education is rationed, and that the rationing sorts people into two categories. One group gets shaped into leaders with the kind of skills and connections that are harder for AI to replace. The other is being fitted, efficiently and at scale, into the shape of a cog.

What it should be

It feels like everyone is in agreement that the age of AI has made critical thinking more valuable, not less.

If critical thinking is the thing that matters, then any education system worthy of the name has to be in the business of producing it. A critically literate person is someone who can prompt an LLM effectively. They are also someone who can cross-examine the CEO of the company that built it. The two capacities draw from the same well. If we are going to hand an entire generation the first, while withholding the second, we should at least be honest that this is what we are doing. Otherwise, we sell ‘democratisation’ to the world while robbing democratic society of its core.

I do not have a tidy policy answer for what a more equitable future of learning looks like (but I look forward to the prospect of cross-registering at Harvard Kennedy School and the Graduate School of Education next semester in search of one!). However, I am certain of the measure against which we should assess every credential or workforce development intervention. The measure is not how many people a platform can reach, or how cheaply it can deliver a credential. It is whether the person on the other end of it gets to be shaped into somebody, or fitted into something.

Folu Ogunyeye (MBA ‘27) grew up in Milton Keynes in the UK. She graduated from the University of Cambridge with a degree in Human, Social and Political Sciences. Prior to HBS, Folu worked in HR at BlackRock in Atlanta and in social impact consulting between New York and London. Beyond work, Folu loves to write, explore art galleries or catch a live jazz gig.