Udacity Founder and CEO — and famed inventor of self-driving cars and wearable technologies at Google — came on the Structure Show podcast this week to talk about the promise, limitations and future of online education. Here’s what he had to say.

Sebastian Thrun was instrumental in building Google’s self-driving car and Glass projects, and helped launch the company’s Google X wing to spearhead “moonshot” projects like Project Loon. But, as he explained on this week’s Structure Show podcast, Thrun’s latest endeavor, online-education startup Udacity, might have an even bigger impact on the world than his futuristic technologies.

The interview covers the promises and limitations of online education — from the misconception that massive open online courses, or MOOCs, should be considered replacements to traditional university educations, to the importance of training all students in the ways of data. Here are the highlights, but anyone interested in the evolution of online education will want to hear the whole thing.

Download this episode

Subscribe in iTunes

The Structure Show RSS Feed

Udacity was never about replacing Stanford

“Part of the problem with the public dialogue is that some people have very quickly moved forward to tout what we do at Udacity and our partner companies as kind of a replacement for college,” Thrun said. That kind of positioning, he added, has “unnecessarily polarized the field.”

Of all people, Thrun — who has taught for years and still teaches at Stanford — should understand the differences between the two media. In his mind, platforms like Udacity should be complementary to colleges and aim to reach people who don’t have access to schools (for any number of reasons) or who need to learn specific skills. That’s part of the reason why technology companies, rather professors, teach so many of Udacity’s courses.

“In the end of the day, a lot of us are really in it for innovation, for really changing the access that people have. Most of our students right now would never ever be able to go to a physical campus, and now we can reach them with education at home, and also the ability to be very contemporary,” Thrun explained.

While college professors might teach generally the same material year after year, he added, “In this new medium, we can get companies into the driver’s seat who really take bleeding-edge material that might only be fresh for three years, but it’s very easy to replace it.”


Yes, there are limitations

“It turns out online is not for everybody, for some people face-to-face works better,” Thrun acknowledged. “And some of the data that we returned showed lower success rates as you’d find in face to face.” Although, he added, the company has improved success rates over time by analyzing data around students’ performance and engagement.

But other challenges remain — like reaching every student who might benefit from this type of education platform.  Udacity (and its peers like Coursera and EdX) do reach broad broad demographic ranges but, Thrun said, “[T]ruth-telling is that we reach young professionals, age 24 to 34, more than other segments. We reach very few 80-year-olds at this point, partly because you have to be computer literate to really engage.”

You also need a broadband connection, which eliminates two-thirds of the world, and, in most cases, must be able to speak English.

But when MOOCs work, they work

Thrun said he recently calculated scores for a group of about 200 Stanford students who took a particular course, along with some 23,000 who completed it on Udacity, and found that the top 412 were all online students, with the best Stanford student coming in at No. 413.

“And I think it speaks a lot for the potential of this medium,” he said. “It also speaks for the fact that our existing universities, as beautiful as they are — they’re certainly the world’s best system — by virtue of being somewhat exclusive, fail in identifying every single talented person on the planet.”

Want proof? Udacity’s introductory computer science class teaches students how to build a search engine, and, Thrun said, “Some of the best students in the project space came back with search engines that would blow your mind. … [Two students recently] came up with ways to present video as the result of a search that I believe is significantly better, and I showed it to a few friends at a major search engine company and they actually said, ‘Wow, that’s quite amazing.'”

Another course, in which Google engineers taught students how to build online games in HTML5, has spawned a few startups, Thrun added.

Learning Python in Udacity.

Learning Python in Udacity.

Training workers, not shaping minds

So really, Thrun sees Udacity, at least, as less of a cheaper, online version of Stanford and more of a place where Stanford grads — and anyone, really — can learn the technological skills they haven’t yet learned.

“I will be satisfied with what we do if it can make a material impact on the skills gap,” he said. “… In this country alone, we have 2 to 3 million open positions in technology and we can’t fill them. Every company that needs technologists complains about they can’t find skilled labor, even today. At the same time, we have people who are currently unemployed and can’t even get into jobs. There we see the biggest opportunity for us at this point.”

One of the most important skills people can learn, he noted, is how to understand and work with data. Indeed, as we’ll cover at our Structure Data conference in March with everyone from Foursquare’s Dennis Crowley to Vern Brownell, CEO of quantum computer pioneer D-Wave, data analysis is underpinning all sorts of new businesses and understanding of our world.

“I generally believe people should be data savvy and we should teach statistics in high school … and the reason is the world today really revolves around analytics,” Thrun said. “… So the ability to think data, the ability to speak data, to understand the power of data, I think everything citizen on this planet should understand, because it’s so powerful.”

Feature image courtesy of Flickr user HarshLight.

You’re subscribed! If you like, you can update your settings

  1. Michael Feldstein Sunday, January 26, 2014

    This is a little disingenuous from a guy who had previously said that there will only be 10 universities in the future and that Udacity has a shot at being one of them. I get that he’s evolved in his thinking about MOOCs, and that’s fine (admirable, even), but he should square what he’s saying now with his previous statements. And honestly, Derrick Harris should have called him on this and asked him to respond. It’s not like his “10 universities” statement is obscure and little-known. See Wired’s 2012 interview for the attribution.

  2. Thrun is so careless. Is this intentional, as every word he utters ends up at the top of the MOOC discussion. Let us take a sample from this article:

    “So really, Thrun sees Udacity, at least, as less of a cheaper, online version of Stanford and more of a place where Stanford grads — and anyone, really — can learn the technological skills they haven’t yet learned.

    “I will be satisfied with what we do if it can make a material impact on the skills gap,” he said. “… In this country alone, we have 2 to 3 million open positions in technology and we can’t fill them.”

    Stanford grads . . . and just anyone, really. Um, just what set is this, and if it is everyone, really, then what does “Stanford Grad” have to do with it?

    Is Thrun saying that Stanford Grads are contributing to the skills gap? Like, um, everyone else? As one who has taught at Stanford, I can report that I couldn’t find a B student in any lecture hall of 200 there.

    If Thrun is continuing to write off MOOCs as alternatives to college courses, could he at least come out and say that simply and clearly?

    1. I think the point is that you can’t learn everything you need to learn in college. A CS degree from Stanford from 1997 doesn’t mean you know the latest AI techniques, for example. Maybe you have an economics degree and an MBA but need to learn how to code.

  3. Thrun is actually a bit of a twat. Always over-hyping stuff then over-correcting. Just listen to his own lectures on Udacity – they’re peppered with statements like this:
    “… and now I must congratulate you – you’ve achieved something really, truly, absolutely amazing. I am so proud of you all and it’s a real privilege for me to be witnessing this. You have now understood how to (invert a matrix or whatever). This puts you in the elite of the elite of technology companies. You now understand 99% of what it takes to build a self-driving car.”


  4. Lynn McAllister Saturday, February 1, 2014

    I believe online is the way to go for education in our techno-savvy world; however, you mention the success rates and online not being for everyone. That is very true. In time, those seeking to advance their degree will meet up with the technology as a requirement and not an option even in the classroom, if it has not happened already. I think the trend is definitely with MOOCs and online degree education versus tradition and very expensive in the classroom learning. In our nursing program, we are given a 40% discount on textbooks if our students purchase ebooks. Using the technology for delivery saves money and this can be passed along to the student. I want to see this technology developed in order to bring education to the masses at a much cheaper rate.

  5. Claude Coulombe Thursday, February 13, 2014

    Watch out for Thrun, watch him carefully!

    He’s learning from his failure and taking a strategic retreat since he was in too hurry. He knows now that the MOOCs are not a sprint but a long-distance race, a marathon.

    Remember what he said, « the company has improved success rates over time by analyzing data around students’ performance and engagement ».

    In another interview to NPR.org, Thrun also said: « t’s certainly an iteration, » … « And the truth is, look, this is Silicon Valley. We try things out, we look at the data, and we learn from it. »

    The key is there… mining the data, learning from data. Thrun and his team are learning from the big data how to improve the « lousy product ».

    The glory can wait a little… let him just the time to learn a bit more.

Comments have been disabled for this post