🎦 Education should be like Hollywood

Imagine: A movie studio has $200m to make a movie. But due to high distribution costs, any film production can only ever be played exactly once and only to a small audience.

So the studio employs a million filmmakers, each of whom is given $200 to make her own version of the movie. Each filmmaker does her own script, directing, acting (all the characters), effects, marketing, distribution, etc. Her film is played exactly once to a small audience.

Among the million filmmakers, a rare few may — either through sheer dumb luck or extraordinary ability — create a decent movie out of the miserable $200 budget. But most of the filmmakers would make a very shoddy product.

The production model of film-making I’ve just described is obviously absurd. Yet it exactly describes the production model of education — a model we used at the dawn of history and continue to use, even today.

Consider for example Calculus 1 (probably the world’s most-enrolled university course). Each Calculus 1 teacher uses only limited resources, but collectively such resources add up.  Suppose 2,000 teachers teach it 3 times a year; each time spending 10 hours a week, for 12 weeks; and each teacher’s time is worth $20 per hour. Then the value of the teachers’ time — ignoring absolutely all other costs — is alone worth $14.4m per annum. At a discount rate of 5%, the present discounted value of teaching Calculus 1 this way forever is $288m.

The numbers in red are arbitrary. Tweak them as you please. The point is merely that the resources involved are staggering and, as I argue, mostly squandered.

If only distribution costs were low. Then we would do the sensible thing: Don’t dissipate the $288m into the production of myriad lousy, disposable products; instead, use it to make a single, excellent Calculus 1 course, which can then be mass distributed.

The Game Has Already Changed

Skeptics point to the failure of promised revolutions in the past: moving pictures, radio, TV, VHS, etc. The skeptic’s favorite source of amusement is Edison’s (1913) mistaken prediction:

Books will soon be obsolete in the public schools. Scholars will be instructed through the eye. It is possible to teach every branch of human knowledge with the motion picture. Our school system will be completely changed inside of ten years.

But the skeptic would do well to pause and consider: Why is it that even a mere 10 years ago, nobody — much less teenagers and young adults sitting at home — could have reached millions (and make millions), just by talking about makeup, playing FIFA, explaining the difference between the UK and Great Britain, or simply screaming a lot while playing video games?

Thanks to the maturation of the internet, the game has already changed. The technical problem of high distribution costs has already been dramatically mitigated.

And distribution costs will only continue to fall, as internet access and speeds continue to grow, and as innovations continue to be made at the interstices of education and technology (or ‘EdTech’).

But even if technology were frozen at today’s levels and distribution costs fall no further, we can — nay, we should — already be doing much better. Today. Current aspirations are far too modest, compared with the possibilities within our grasp. Some think we can only ever hit the Olive Garden ‘good enough’ standard. And that’s a damned shame.

Raise the Bar Now: The YouTube Standard

Ever heard of a professor bemoaning that to do all the work involved for his 4-page publication in Nature, it cost him a whopping $70,000 in NSF funding and 200-300 hours of work? Me neither. Yet when a MOOC production makes use of such meager resources, the Cornell Daily Sun considers it ‘costly’. That’s how depressingly low the bar is, when it comes to teaching. Let’s raise the bar now.

Online education should already be hitting the YouTube standard: “If it ain’t at least as good as CrashCourse, CGP Grey, Vsauce, or any educational YouTube video that you can make sitting at home in your pyjamas, then it ain’t good enough to put online.”

Consider for example a 5-second delay. Even a modest educational video today can easily get 100,000 views. In which case, one such delay costs the world 6 man-days. Yet there are still MOOC videos with these astonishingly unprofessional delays. Watch any of the top YouTubers and you will find nary a frame (or even a pixel) that is boring, irrelevant, or superfluous. Much less a 5-second delay.

In the future when educational videos routinely get a billion or more views, a 5-second delay will not merely be unprofessional. It will be criminal: a 5-second delay inflicted upon an audience of a billion amounts to two counts of manslaughter (5b sec ≈ 158 years).

But even today, when audiences number in the “mere” hundreds of thousands (or sometimes millions), perfection is already worth seeking. It already makes sense to spend scores of man-hours to make a 6-minute educational video even 1% better. It already makes sense to do a bit of work cutting out the “uhs” and “ums”.

Online education today should already be hitting this easily-attainable YouTube standard. It is inexcusable that serious online educators are not at least as good as teenagers sitting in their parents’ homes. Imperfections — especially easily rectifiable ones — are intolerable.

Make a Few Excellent Products, not a Stream of Unremarkable Ones

In the medium to long term, each MOOC budget will run to hundreds of millions.

But even today, nobody should be complaining if each MOOC costs millions, not only because this is what the world’s learners deserve (it is), but because this is what already makes perfect economic sense.

The reader is probably scoffing at my naïveté. Where are these millions going to come from? Who will pay for it?

Well, right now online education industry is already sinking very many millions of dollars annually. But for this, we are getting a return on investment that is only somewhat better than the traditional educational model’s. All we are getting is a digital replication of what the better teachers do in their classrooms. This is merely a marginal and unremarkable improvement on the status quo.

Consider for example Khan Academy, where “shipping beats perfection”. By any of the usual standards that we’re accustomed to, Mr. Salman Khan is a very excellent teacher indeed. But each of his videos is a single-take, unscripted, and unedited affair. It is clear where he stands on the “quantity versus quality” tradeoff. And it shows: Despite the hype and the nearly 3m subscribers, each video garners only several thousand views. (The subscribers-to-views ratio must be some sort of a record.)

CrashCourse has shown that if one is willing to slide a little up along the “quantity versus quality” tradeoff, it is possible to do much better. But CrashCourse can do still better, were they to strive for the levels of perfection exemplified by the content of a CGP Grey or the visual effects of a Kurzgesagt (or In A Nutshell).

The Medium Term: The Hollywood Standard

The point made above is that in the short term (i.e. today), we should already aim for the YouTube standard. But a little down the road, in the medium term, we should aspire to the Hollywood standard: “If it ain’t good enough for a Hollywood hit, it ain’t good enough for a MOOC.”

By the ‘medium term’, I mean what is already technically possible with today’s technology and today’s resources. Getting there will not be easy — it will require introducing more vigorous forces of competition into the teaching profession and destroying the belief that education is worthless except as a signal. (I will discuss the problems of competition and signaling in future posts.)

Some see a future with a handful of superstar teachers commanding global audiences. But such a vision is an unimaginative extrapolation of the past: “Individual teachers have always done their own thing; so the future will simply be more of the same, just on a grander scale.”

I see something different: Each MOOC will be produced by a professional MOOC production team, each being composed of thousands of full-time experts.

Fantastic? Not really — Hollywood has already been doing this for a century.   A superstar actor is paid only to do what he does best — deliver lines in a compelling fashion. He is not paid to do scriptwriting, directing, marketing, special effects, props, makeup, or anything else. Chances are he’s not very good at any of that. After all, each person can have only one area of comparative advantage.

A lone superstar actor does not a Hollywood hit make. Perhaps MOOCs of the future will likewise have a Jennifer Lawrence or Robert Downey Jr. as the star of the show. However, not unlike Hollywood, the hardest and most important work will be done behind the scenes. The massive backstage team of superstar teachers and technical crew will obsess over every last detail and guarantee that the MOOC is perfect in every way. Most of all, they’ll obsess over what’s most important: the story, the content, the pedagogy.

Education can also take a page out of the stand-up comic’s production model. A comic spends a year or two honing to perfection his material in front of small audiences at comedy clubs and shopping malls. Upon finally attaining nirvana (and 60 minutes’ worth of material), he then films a special, to be released on Netflix, HBO, and DVD. Perhaps future production models of education will resemble the stand-up comic’s.

Each MOOC will have its own StackExchange-style site, where for every point of confusion, there will be an ever-growing and permanent list of the very best explanations. (“What’s the difference between demand and quantity demanded?” “Why isn’t (x + y)^2 = x^2 + y^2?” “What’s the difference between evolution and natural selection?”) We will no longer have generation after generation of students struggling over the same, damned points of confusion.

Around each MOOC will grow a fiercely competitive mini-industry of dedicated tutors from around the world. At present it is only the education-obsessed, Confucian societies of East Asia where superstar, millionaire tutors can thrive. In the future, they will span the globe.

In 30 years, MOOCs will command the same level of resources, effort, talent, and dedication as today’s Hollywood mega-productions. The economics of the matter means that it already makes sense to start working towards this vision. Today.