Robert Reich has three very good questions about Sebastian Thrun’s new online university, Udacity, which I wrote about last week. I spoke to Thrun yesterday, so I took the opportunity to clear them up.
1. Why did Thrun need to quit Stanford? Why not pursue the project under the umbrella of Stanford, with its enormous and global reputation? Indeed, hadn’t he already carried out a demonstration proof of the concept with his Artificial Intelligence class at Stanford? Why not just continue with that in expanded form at Stanford?
As Thrun says on his homepage, he quit Stanford on April 1, 2011 — before offering the free class in artificial intelligence — “primarily to continue my employment with Google”. Thrun is a very senior and successful Google employee, and he had somehow been combining that, before 2011, with being a fully tenured professor at Stanford. That was done through something called “leave time” — but leave time is finite, and eventually Thrun ran out of it. When that happened, Thrun told me, “I said I’m not ready to leave Google just yet”. So he gave up his tenure at Stanford, and became an unpaid Stanford research professor.
This helps to answer another of Reich’s questions. “If Thrun developed [the AI] class as a faculty member at Stanford,” he asks, “then doesn’t Stanford have some claim on at least the course content?” The answer, it seems, is that Thrun developed the class after he gave up his position at Stanford. And as a result, the course content belongs not to Stanford but rather to KnowLabs, Thrun’s company.
What’s more, the online version of the course, which was not hosted at Stanford’s website, was very careful with its Stanford branding. Yes, the online course made no secret of the fact that it was basically exactly the same course that Thrun was teaching a group of Stanford undergraduates. But the final certification made no mention of Stanford.
Leaving Stanford, Thrun told me, “was the only way I could pull this off. The statement that we could let the students take exams and compare themselves to Stanford students, is something I don’t think the university would have approved.”
When NPR interviewed Thrun for a story about Udacity, they also got a none-too-enthusiastic statement from Stanford:
Thrun’s colleague Andrew Ng taught a free, online machine learning class that ultimately attracted more than 100,000 students. When I ask Ng how Stanford’s administration reacted to their proposition, he’s silent for a second. “Oh boy,” he says, “I think there was a strong sense that we were all suddenly in a brave new world.”
Ng says there were long conversations about whether or not to give online students a certificate bearing the university’s name. But Stanford balked and ultimately the school settled on giving students a letter of accomplishment from the professors that did not mention the university’s name.
“We are still having conversations about that,” says James Plummer, dean of Stanford’s School of Engineering. “I think it will actually be a long time — maybe never — when actual Stanford degrees would be given for fully online work by anyone who wishes to register for the courses.”
Stanford, then, has managed to come out of this story smelling reasonably good: it helped give Thrun the launching pad for Udacity, and didn’t visibly complain about his course material appearing online for free. But it didn’t really help him in an active way, and probably, if Thrun had been affiliated with Stanford going forwards, would have ended up hindering what Thrun wanted to do.
Thrun isn’t upset about that: he understands why his dream isn’t really compatible with what Stanford does. “If you grade external people without checking identity,” he says, “you open the floodgate to fraud. Stanford’s position was very well justified. They weren’t being anal, they were concerned. But I felt we should just teach students online for free, and that the engagement we get from this exceeds the uncertainty that comes from these new certificates.”
Indeed, the sign-up rate for Udacity’s new courses seems to be exactly the same as the sign-up rate for the AI class which was co-branded with Stanford: the lack of a Stanford branding doesn’t seem to have hindered Udacity much. And similarly, the incredible success of Khan Academy has taken place without any co-branding with a venerable legacy institution.
There’s no doubt that Stanford is extremely good at doing what Stanford does — which is to hire great research professors, select a tiny minority of the students who want a Stanford degree, and then bring the two groups together in a productive manner, while building up a world-class reputation and acting as the beating heart of Silicon Valley, pumping knowledge out into its surrounding neighborhoods and creating what is arguably the single most innovative region in the world.
But both Stanford and Silicon Valley more broadly are elite institutions, home to the ultra-productive 1%. Thrun’s ambitions are more demotic, and in that sense cut against what Stanford stands for.
Udacity is very much a teaching institution rather than a research institution. “At Stanford, priority is your research career,” says Thrun. “That is counter to teaching 100,000 students, who generate 100,000 emails.” Looked at from a 30,000-foot view, Stanford is the institution being disrupted here, it’s not the institution doing the disrupting.
And that also helps explain why Thrun isn’t doing Udacity under the auspices of Google. He says that Udacity does fit quite easily into Google’s mission of making the world’s information available for free, but that at the same time Google doesn’t need a dog in this particular fight. “Having a clean slate is a better way to start,” says Thrun. “The last thing I want is people asking whether Google is disrupting education. Better to ask if Sebastian is trying to disrupt education.”
(That said, Thrun’s friend and boss Sergei Brin does feature prominently in the launch video for Udacity’s first course; it’s clear that Udacity has Brin’s strong support. And of course Google is also a big supporter of Khan Academy, having gifted it some $2 million.)
So it’s pretty easy to see how Thrun, if he wanted to create an organization which could grow incredibly fast around the world, might want to do so without having to get Stanford’s sign-off on everything first. But that doesn’t answer Reich’s second question:
2. Why is Udacity a for-profit company? Thrun said that Udacity courses would be free to students, and Thrun cited Salman Khan and Khan academy as inspiration and model for what he’s doing. But Khan Academy is non-profit. Stanford University is a non-profit. Thrun says he wants to democratize higher education, offering knowledge to the world for free. How does this mission fit with his for-profit online university?
I asked Thrun about this, too, and he replied by saying that “for profit is not forced to make profit. I needed to get people together really fast, and it’s much easier to do that under the ways of a Silicon Valley company.”
Certainly the speed with which Udacity launched, complete with a high-quality staff, is testament to the natural velocity with which things get done in Silicon Valley. Driving the launch was seed funding from Charles River Ventures, while the site’s jobs page proudly offers “Competitive salary, benefits, and Series A stock options” to anybody thinking about working at Udacity.
This is an interesting model, and it’s not necessarily the one I would have chosen. Salman Khan, for instance, is quite vocal about why it’s good that he’s a non-profit, and the way in which the dreams of venture capitalists who have approached him conflict with his fundamental vision. What’s more, Khan did end up receiving $5 million from Irish venture capitalist Sean O’Sullivan — just as a philanthropic grant, rather than as an equity investment. And certainly Thrun has no desire to join the ranks of America’s for-profit colleges, which make their money from tuition fees.
But still, online education is young enough that it’s worth trying many different models to see which ones work. Udacity seems to be built on the standard VC model of get scale first, worry about monetizing it later. And if Udacity does end up with millions of students, I should imagine that there are quite a lot of companies which would pay Udacity to be able to reach those students. Simply charging technology companies to put job opportunities in front of students with given grades and qualifications would probably generate quite hefty fees. So long as the education itself remains free, I don’t think that being a for-profit is in and of itself a bad thing.
“We should try many different things,” says Thrun. “I believe in the educational revolution that Salman started. I believe that education can change the world. So why not try a hundred of those things.”
This seems reasonable to me. A large part of the success of both Khan’s courses and Thrun’s is the way that they’re presented and executed, rather than any business model behind them. Khan, in particular, is a hugely gifted natural educator. And what both of them aspire to doing is to build what Thrun calls “magic” into the way that they teach. Thrun wants to add another element, too — community. His courses have a start date and an end date and deadlines, with thousands of students all taking the same class at the same time; that makes them inherently social in a way that Khan’s YouTube videos aren’t.
Finally, asks Reich,
3. What to make of Thrun’s apparent pleasure at the fact that 170 of the 200 Stanford students who had enrolled in the real, not online, version of the Stanford AI class stopped coming to class, preferring the online Thrun to the flesh-and-blood Thrun?
That pleasure, I’m quite sure, is genuine. I think that Khan and Thrun are at the forefront of a new, more personal way of teaching — think of them as having screen-actor skills in a world which has historically rewarded stage-actor skills. When you teach online, you’re teaching in a conversational manner, in a one-on-one space. And it turns out that many students — quite possibly most students — prefer being taught that way, as opposed to the old-fashioned model where a lecturer stands up in front of a crowded classroom and declaims to many people at once. Most students are naturally shy; they don’t like speaking up in class and saying that they don’t understand something. Online, they can just rewind and replay, or pause and look it up on Wikipedia.
And then of course there’s the fact that the incentives for the teacher are so much greater online, if like most teachers you’re driven by the opportunity to impart knowledge to students. “This is the best thing I can do in my life,” says Thrun. “I empowered more students in 2 months than in my entire life before. On that scale, I was off the charts in the last quarter.” And of course Thrun is barely on the charts if you compare him to the number of students that Khan has reached.
What Khan and Thrun and others are creating is a new educational paradigm, which promises not only much greater scalability than anything we’ve had until now, but also higher-quality education. That’s the real lesson of Thrun’s Stanford students taking his class online: it means that the online model really can have its cake (reach millions of people) while eating it too (be better for students than the courses offered at elite institutions).
The trick is intimacy, in a way which takes full advantage of the lean-forward nature of computer screens. I’m in England right now, where the Open University has been around for over 40 years. The OU has historically reached students through the lean-back mediums of TV and radio, which in turn encouraged its lecturers to behave as though they were trying to reach a large audience. When you see Salman Khan or Sebastian Thrun drawing pictures on the computer screen in front of you, while listening to them talk to you through headphones you’re wearing, the experience is very different — it’s a much more immersive and intimate experience. Blow that YouTube video up to full screen, and jump down the rabbit hole. You might just learn something.