Professor Resigns From Stanford To Launch Online Education Project
mikejuk writes "Professor Sebastian Thrun has given up his Stanford position to start Udacity — an online educational venture. Udacity's first two free courses are Building a Search Engine and Programming a Robotic Car. In a moving speech at the Digital Life Design conference, he explained that after presenting the online AI course to thousands of students he could no longer teach at Stanford: 'Now that I saw the true power of education, there is no turning back. It's like a drug. I won't be able to teach 200 students again, in a conventional classroom setting.' Let's hope Udacity works out; Stanford is a tough act to follow."
Hope it turns out better than his class did. The other classes were far better managed than the AI course.
how will it be monetized, and I don't mean that in a negative way. (also, bad first link in summary)
I read TFA and all I got was this lousy cookie
This headline is like writing "Man shot in Ford's Theatre"
Funny - I thought I had ads blocked on here. What's so unique about online education?
Of course he did, he teamed with Khan and relied on him for some parts of the AI course teaching prerequisite maths and probabilistic theory.
Achille Talon
Hop!
good on you sebastian, That ai class was amazing.
What is their business model? I checked out their page, they've got a couple employees, they're offering stock options to new hires... I'm seeing this as one of those /. jokes:
1. teach a free class at a profitable school
2. quit and teach a free class at a startup
3. ???
4. Profit!!!!!
Is the plan to operate on donations, or ... ?
"Science flies us to the moon. Religion flies us into buildings." - Victor Stenger
I wonder when he will hold office hours?
When Fascism comes to America, it will call itself Anti-Fascism, and tell you to give up your guns.
That was unexpected. But then, his automatic driving work had already moved to Google.
He turned around the Stanford CS department, which was embarrassingly bad for years. (I have a degree from there; I know.) It was being run by the mathematical logic people, who were trying to make AI work through predicate calculus and expert systems. That turned out to be a dead end, but the existing faculty didn't want to admit it. Thrun reoriented the department towards statistical methods for AI, and things got moving again.
When lectures can be saved to a video format on the Internet, why pay the teacher to deliver the same lecture every year?
When books can be copied for free, why pay 200$ for a physical version of the book?
I think the only thing we'll have in terms of live people will be live tutors you can ask questions via advanced IM
The cool thing about this is that it is the opposite of the "No child gets ahead act", if a kid is motivated, they can watch hundreds of supplemental optional videos related to their course. Or with proper understanding of the subject at hand, they can move ahead to the new videos. Also this is all available for free or nearly free, so the cost of an education is simply 100$ or less for a laptop. This means people across the world who couldn't have access to quality education will. If you're in a 3rd world country with nothing to do all day, maybe you'll devote your life to getting a grand education. We might find new Einsteins popping up and at younger and younger ages.
God spoke to me
I hope the class search-engine projects turn out better than the Udacity.com website did.
Does anyone know where the Google-Car code resides?
Mr Thrun was with the GoogleCar...
I am too lazy to look, so it seems like an opportunistic time to seek "the code".
cellurl...
In the first universities anyone could stop in and listen to a lecturer for free. If they were interested in perusing individual education they would work out a fee between the professor and the student. There wasn't any strict curriculum or degrees. The professors paid the university a cut similar to the way a barber shop works today.
The business model should be the same. Free to watch the lectures and pay for individual attention.
I love Jesus, except for his foreign policy.
are in dire need of writing their own search engines and programming robotic cars.
To quote the phrase: Don't quit your day (or night ) job.
Yours In Minsk,
K. Trout, C.I.O.
D'oh! Posted in the wrong story! Mod down. Offtopic.
File under 'M' for 'Manic ranting'
Now that I saw the true power of education, there is no turning back. It's like a drug. I won't be able to teach 200 students again, in a conventional classroom setting
I had the exact same feeling of elation when my Chicks with Dicks site really took off. How could I go back to being just another carny running a ring toss game after that?
SJW: Someone who has run out of real oppression, and has to fake it.
I was one of those who took the online "Intro to AI" class Advanced Track. While I didn't do as well as I'd hoped (screwed the pooch on the final and ended up with a 78% total score), I do have to say it was a really great class and that he and Prof Peter Norvig deserve a ton of credit for this. It was both engaging, well presented and mind-stretching. I did find myself consulting the Khan lectures on Probability and Linear Algebra quite a bit.
Is this related to the delays in the start date of the other free stanford clases? (http://www.class-central.com/ for example)
Thrun is (I think) the first tenured Professor at a major University to stand down in order to try to bring learning online. Unlike the offerings from Stanford, MIT, Berkeley etc etc, Udacity wont be under the same "Don't damage the university's business model" constraint, so they are truly free to go for broke.
There has been a lot of criticism of the AI course - most of it by people who didn't attend beyond the first couple of weeks. I finished the course and had a good time doing it. It wasn't without flaws, but I have no doubt that with the necessary financial backing, they can make the necessary improvements and push on to create some remarkable content.
If they can solve the question of certification, they, and those who will inevitably follow, might just revolutionise the educational landscape.
And if it all goes wrong, Google wont kick him out of bed.
The Edison Project was private for-profit K12 schools combining modern business management and high technology. Possibly a good idea, but got little traction. The wiki site said it was had to get the "education establishment" to buy in and build many of these.
Around where I am now there is a flourishing charter school ecology. Some are to escape the low-expectation public schools. Others have religious slant. And still others advocate challenging education like Chinese language immersion or computers. Even the high end schools worry some parents whether these are too fringe for their kids to get into a top college.
"Khan, you bloodsucker! You're going to have to do your own dirty work now! Do you hear me? Do you?"
Start with the basics (beat Google) and build up over time to something really difficult.
Deleted
I find it interesting though, that Sebastian Thrun gets so much attention, and Andrew Ng for example, gets no mention. I think that Ng poured in a tremendous amount of effort to teach an absolutely outstanding class with far more structured and well-developed content.
Don't get me wrong, Thrun is an enthusiastic and obviously knowledgeable individual, but having followed both AI and ML classes, I was of the opinion that Andrew Ng was the better teacher. Thrun needs to improve his teaching skills, so that he can impart his great store of knowledge better to students. Although that is my personal opinion, I think you might find that it is backed by some evidence, if you were to trawl through the comments on the respective forums of the AI and ML classes. Overall, both of them + Peter Norvig and the rest of their teams, made fantastic contributions, and that should be recognized equally, whenever possible!
I have no doubt that these guys are brilliant and are providing us with remarkable educational opportunities ( thanks), but I would not want to learn web development from them based on my experience of their web page. I keep Noscript turned on at all times unless there is something specific I need to do or see. With Javascript turned off the only links on the home page that work are in the footer.
Maybe it's just me, but I don't think so. I expect more (and better) from -- well -- everyone, but to not see it here ( on a CS educational web site), it kind of bums me out.
/absolutely agree.
The Open University have been doing this stuff literally for decades. I fondly remember as a child watching lectures from teachers with excruciatingly bad 1970s hair styles and clothes.
http://www.open.ac.uk/
Note the fees are the standard (substantial) UK university fees, so it isn't free by any means.
Deleted
What's the problem with their "Enroll" Button? This isn't exactly a good start for a course called CS101.
I felt the other way. I really enjoyed Sebastian's almost-whimsical adventure through the field of AI. I really think Ng is a good professor but I often found myself having trouble paying full attention for the entirety of the course.
Slashdot is proof that Sturgeon's Law applies to mankind.
I took the advanced class too, got 73% (not very good) BUT, I didn't have a lot of time as I took the ML and DB course (got 100% and 95%), work full time, and have a wife and three kids. My problem, like yourself by the sounds of it, was Probability and I just didn't have the time to study/learn it so it really brought me down. Over the Christmas break I went looking for a course/book on Probability but didn't have much success. I got one general Linear Algebra book that has a chapter on Bayes Theorem and found an online lecture series (from Harvard) about probability as well but it was so boring I couldn't get through it :(
Hopefully I will get back into that Harvard one and/or get another (newer) online course on Probability soon. Any suggestions for a good self-paced book or course for Probability would be greatly appreciated :)
(BTW, I did University 20+ years ago and calculus is a faint memory (surprisingly, of the Rum smells wafting from my Profs office) and I don't remember ever doing a course on probabilities but did do a Statistics and Linear Algebra course)
But what about research grants?
What makes some professors interesting is the part of their time they spend 'pushing the envelope' doing research. The online stuff can (if used properly) allow them to reach more students with less demand on their time. So, more research. I'm fine with that.
What made Sebastian's class interesting was some insight into his (award winning) work on the DARPA challenges and other robotic car stuff.
Have gnu, will travel.
I registered for a couple of free online Stanford classes that were supposed to start yesterday (Machine learning, NLP, game theory). All were delayed for an indefinite time (couple of days to couple of weeks, no exact number was given). Might this be the reason?
The modern lecture format originated in medieval Northern Italy, and hasn't changed significantly. The rationale for the lecture as a method of transmitting knowledge and skill was that books were extremely costly, due to the cost of scribes.
Since Gutenberg the rationale for lectures has disappeared.
Rather than moronically scaling up lectures in a TV-like way, we need some R&D done on better methods of teaching. This has finally been realised and academics are - with great trepidation - starting to measure themselves and experiment with different methods. I expect that this century will see the death of the lecture.
I took Thrun's computer vision course at Stanford. The man is probably a visionary researcher, but he's not a very good teacher, in my opinion.
You only need to record a lecture one to three times with possible editing. Maybe update it now and then. It's the equivalent of a electronic textbook. Only thing left is answering questions and office hours per session. If he does this it just seems to me he will just be the director of a tutoring group. Are professors/teachers really needed anymore?
Oh yes. I took all three classes (AI, DB and ML), and AI was plagued with ambiguity, handwaving and mistakes caused by sloppiness. Some homework questions had so many errors that they were unsolvable and had to be amended several times shortly before the deadline, of course without telling the students who had already submitted their answers. Check out the old discussions on aiqus.com, it was horrible. The AI class was also the one with no practical exercises at all, except for a tacked-on codebreaking exercise after the final exam that was neither graded nor properly discussed afterwards. The AI class server software did not include video streaming (they relied on Youtube instead) nor the promised forum. In contrast, both Andrew Ng's Machine Learning class and Jennifer Widom's Database class were hands on and thoroughly prepared. I learned a lot more there. I think the problem with Thrun and Norvig was their attitude, probably not unrelated to being employed by both Stanford and Google. They expected to be venerated as superstars. They seemed to think they could pull this off without much effort.
The name "Udacity" seems a bit too cute if part of the goal is to have someone want to put successful completion of this education from this institution on their resume/CV and have it be taken seriously.
Udacity seems to have been over-whelmed as of the time of this post: 9:11p EST (maybe apropos?).
I doubt Thrun intends to offer a few courses and stop there. I think he'll offer an entire CS curriculum within maybe 3 years, and offer some soft of CS degree program soon thereafter.
It seems like you could offer other degrees using this same technology -- probably all engineerings, physics, probably math and statistics, maybe biology (but without labs).
Not only would the degrees be FREE (a huge thing for the poor in the third world and BRIC countries), but they'd be FAST. By excluding all the non-essentials, the equivalent of a BS in CS could be completed three times faster, in no more than 1.5 years.
Based on what I've seen from Thrun so far, I bet the degree will be widely respected, and frankly, better than 3/4 of today's CS degrees.
Universities beware. You're about to run smack into The Innovator''s Dilemma. And in my humble opinion, it's about damned time.
Govt held the Darpa Autonomous Car Challenge that motivated Thrun.
I think Andrew Ng had an easier job: machine learning is a course with a curriculum that's better defined than the almost all-encompassing AI class. That's why I had the idea the AI course sometimes jumped from subject to subject, while the ML class was more building up to something. I agree the ML class was easier to follow, I don't think it's because of the teacher, though: swap Andrew and Sebastian/Peter and I think you'd gotten the same result. Aside from that, I immensely enjoyed both courses (enough to enroll in one of Udacities courses as soon as it was announced), so they both did an excellent job.
I doubt Thrun intends to offer a few courses and stop there. I think he'll offer an entire CS curriculum within maybe 3 years, and offer some soft of CS degree program soon thereafter.
That's a heck of a lot of work to create. I know from experience that a half-semester part-time course requires a massive amount of work to create (though not so much to maintain afterwards, to be fair) and you need a lot of those to build a proper curriculum since at degree level you can't just teach everyone the same thing.
Not only would the degrees be FREE (a huge thing for the poor in the third world and BRIC countries), but they'd be FAST. By excluding all the non-essentials, the equivalent of a BS in CS could be completed three times faster, in no more than 1.5 years.
All that effort to make it has got to be paid for somehow. Yes, it doesn't have to be by the students, but it's too much work for it to be reasonable to expect it to just spontaneously happen.
That said, I wonder at the timescales you mention. A nearly 5 year undergraduate degree? WTF? That's absurd. I did mine in 3. I suppose you could have shortened it further by getting rid of vacation time (but that'll likely give most students a mental breakdown of one kind or another; you need breaks from learning) and ditching all those annoying exercises and projects. Which would make the degree totally worthless as it would leave someone only able to vomit up what they'd been told instead of being able to apply it (or their brains).
Nah. A BS in 18 months is an idea put forward by someone without a fucking clue about education or people or the subject. You know, a classic Dilbertian pointy-haired manager.
"Little does he know, but there is no 'I' in 'Idiot'!"
This is why I left academia. Nothing is ever good enough. Thrun delivered an amazing course but 'there's always room for improvement'.
They may hit a snag on the more practical side of things when they find that their full paid course is in competition with other examining bodies.
Certainly the model wouldn't extend to my field -- language teaching (and I mean human languages). The language teaching industry is split into two main segments -- the teaching industry, and the exam mill. There are a few big national institutions that handle a few exams offered universally. In UK English, you need to have the Cambridge certificate to be taken seriously. In Spanish, it's the Instituto Cervantes. So you can't teach for free and make people pay for the certificate, because they want someone else's certificate anyway.
In order to keep the customers paying, Udacity need to keep the courses very unique and individual, which isn't going to appeal as much to employers unless they are in that exact specialism. For more general skills (eg a particular programming language), you want to know that they have general competence beyond the confines of the specific course, so it's actually of benefit to have the teaching and assessment from independent bodies. (Maybe a second wind for the likes of Brainbench.)
In essence, the problem is that these courses are modeled on university-style modules, but don't lead to a degree. Employers like degrees, and always will. But they'll want any additional certificates to be business-relevant, not academic in nature.
Got them moderator blues I blieve I walk out the do', With these mod-points I been gettin', I 'most never post no mo'
Wow there are some fussy fuckers on slashdot.
To have a right to do a thing is not at all the same as to be right in doing it
Not only would the degrees be FREE
No they won't -- just like Coursera and MITx, Thrun's business model involves free training and paid-for certification. The classes are free, but the qualification isn't.
By excluding all the non-essentials, the equivalent of a BS in CS could be completed three times faster, in no more than 1.5 years.
Based on what I've seen from Thrun so far, I bet the degree will be widely respected, and frankly, better than 3/4 of today's CS degrees.
There's still the issue of accreditation, and one of the stipulations of an accredited degree programme is the amount of study time. If this methodology really is much more time-efficient than traditional universities, then what we would expect to get out of it (and I would hope we would want this, too) is a higher quality education, not a shorter one.
European universities are already making learning much more flexible -- the European Credit Transfer System makes it much easier to switch universities or to spend a year abroad. At several universities in Scotland, you can leave your degree programme at the end of any year and have a qualification to show for it. 1st year: Certificate of Higher Education, 2nd year: Diploma of Higher Education, 3rd year: Bachelor's degree (ordinary), 4th year: Bachelor's degree (with honours).
If you could teach the equivalent of an honours degree in 2 years, it would still only be a diploma, not a degree. But then you could transfer to a slower university and take lots of interesting modules to expand your breadth of knowledge and get a high pass at the end of it....
Got them moderator blues I blieve I walk out the do', With these mod-points I been gettin', I 'most never post no mo'
I found both the AI and ML classes had pronunciation errors and/or sloppiness that took away from the course. In a real live setting, you could ask the teacher to repeat what they said. Online, you can only rewind the video and hope that another listen helps (it often doesn't). Closed captioning would really help in this respect. Not all of the quiz problems were well defined, either. In the AI class they didn't take away from your grade, but getting answers wrong due to either an ill-defined problem or not being taught the material yet is psychologically draining. Anyway - main point is there is room for improvement and I'm sure they will.
I took Thrun's AI class, and passed. It was fun, I learned some things, but it was _not_ an university class, by which I mean that while the material was clearly university-level material, it was skimmed over like we would do in high school. I understand it's impractical to test an implementation of AI algorithms for 100,000 students, but well, that's one of the differences between "real class" and "online class".
I don't know if a biology degree without labs, or an engineering degree without labs, would be worth anything. Sure, you would learn basic theory, understand the principle... but you know, that's not enough. It's like saying you understand high-level physics without doing the math around it, it's just not the same level of mastery.
My personal opinion, is that he could replace some technical training schools and teach you CS/IT basics, but I really don't like when people think that learning physics or math is useless in CS/Engineering. These are things that shape your mind and give you a thought-process that is hard to acquire otherwise. And the social skills that universities teach you, through face-to-face group projects and labs? Invaluable, and unlikely to be acquired online.
"If you're in a 3rd world country with nothing to do all day"
Maybe they are interested in eating at least one meal each day. That may require more work than opening a refrigerator or going down to the corner market.
I only look human.
My mother is a halfling and my dad is an ogre, so that makes me an Ogreling
If your professor goes from teaching 100 people to 100,000 people without much increase in workload, then either your professor sucked when teaching 100, or you are missing the point of attending his class entirely.
It's not just a matter of reading the notes and listening to the lecture. The professor needs to respond to problems the students are having, adjust pacing based upon performance, and be able to handle individual questions from students. That's why good professors hate classes that large (100). It's a lot of work.
I think blaming their attitude is going a bit too far, but the course definitely lacked a lot of what made the other two so great.
I started taking all three but dropped AI after a couple of weeks. I could have gotten over the software problems but the quality of the content made it difficult to keep going. The worst thing for me was the quiz questions. Rather than having one or two review questions per 20min video, the AI class seemed to have tedious homework style questions every two minutes. It was more of an annoying interruption than a useful addition.
Still, it's great that this exists, and I hope his experience with AI-class leads to improved content in these new classes.
There's also going to be more online classes offered by Stanford/Berkley starting soon, e.g. http://crypto-class.org./
The Airspace did a very comprehensive profile on Thrun and his experience with the AI class http://theairspace.net/events/robotic-mastermind/
It's amazing how emotional and significant this is. Very disruptive.