Stanford Offers 13 classes, including AI, for free | Hacker News Reader

Stanford Offers 13 classes, including AI, for free

see.stanford.edu

517 pointscpfohl55 comments

Just to clarify: These aren't the same thing as the ai-class (http://www.ai-class.com) that has gotten all the buzz. This program has been around for a while and only(1) offers downloadable materials (videos of lectures + pdfs of assignments/handouts etc.)

(1) "only" seems a terrible way of putting it. The materials/lecturers are amazing even if you never have your assignments graded.

thanks, I totally didn't notice that!

The Machine Learning, and Databases classes are like the AI class on the new system, with quizzes and a statement from the professor about how you did at the end. At least the intro videos for the ML class stated as such and I got that impression from the database class as well.

Am I the only one yearning for the convex optimization I & II classes to be offered in the new style?

Indeed, best courses on SEE by quite a wide margin.

I took those via HCP (which is like SEE but for 3x the usual grad student rate) and he's a great teacher. He's not one to let his smarts get in the way of teaching, which is not true of all the professors at Stanford.

I would love to take a second shot at a convex optimization course. I took one from a former PhD student of Boyd's during my junior year of college and I was not at the right point in my education for a class like that.

It touched on some really great real world optimization problems, including an interesting couple of lectures on the Netflix prize.

For those interested:

Machine Learning - Professor Andrew Ng - http://www.ml-class.org

Introduction to Databases - Professor Jennifer Windom - http://www.db-class.org

Introduction to Artificial Intelligence - Professor Sebastian Thrun and Dr. Peter Norvig - http://www.ai-class.com

2. How is SEE being funded? SEE’s pilot program was funded by Sequoia Capital, the Silicon Valley venture capital firm that helped launch Cisco, Google, Yahoo, NVidia and many other successful technology startups.

They deserve a shoutout.

As universities offer more and more curriculum online, would you hire someone who had self taught himself through a CS degree? Or would you want to see the diploma?

I've found little correlation between diplomas and effectiveness as a programmer in my hiring. So, that includes people who are self taught. I may be in the minority, and I think that people who have diplomas sometimes have the opinion that anyone who didn't go thru the pain they went thru is missing something.

ditto for me. I've actually found after years of interviewing that on average, self-taught programmers are almost always better. Not always true of course, but usually true.

Interesting. I wonder how much that skews hiring, including in other fields.

I'm an old dog so there were no CS classes at my middle school when I started programming. Naturally I have a bias towards those who took matters into their own hands and are self-taught.

I have a huge bias towards self-learning. Most people I know get a degree/diploma to get a job. The people who self-teach are doing it more because they're curious cats. I'll take the curious cat.

I'm self taught and I'm pretty terrible. The kids who went to school with me but majored in C.S. are pretty rockstar and work at Google et al.

Damn, that's a good anecdote to counter my argument. So I'm going to have to provide my own. I took electrical engineering. In third year I was doing a lab, and the guy next to me didn't know that capacitors are used in analog filters. I would have asked him how he got to 3rd year knowing so little, but I was too depressed.

What about people who do both at the same time? Personally I find this model to really work for me, Uni opens up my eyes to what exists out there, and I dig deep into the specifics of what piques my interest on my own time.

I find that most programmers who are only self taught are missing breadth. They're great with what they encountered and had to try, but usually suck at even knowing of the existence of what lies just beyond the edge.

Or at least the ones I've had the pleasure of talking to :)

> They're great with what they encountered and had to try, but usually suck at even knowing of the existence of what lies just beyond the edge.

That is one area where the internet could use improvement.

Google can teach you anything... if you know what you're looking for; but there doesn't seem to be a good system for exploration beyond that. If you follow the right discussion/link services and keep a general watchful eye out, the stuff out beyond the edge will eventually flow to you, but there is certainly room for improvement.

I believe it is a solvable problem and the company that gets it right will have amazing growth potential.

Wikipedia's List of Algorithms page is a start. The category pages are often useful too.

http://en.wikipedia.org/wiki/List_of_algorithms

You can also look up what textbooks are used at major universities, then look up their tables of contents on Amazon or Google Books.

It really depends on the person. Self-taught people are proven to be curious and like to learn, but you can also find people with degrees that are like that. They start with the a degree but never stop the learning process.

The problem is the start. Self taught folks usually find it hard to break into the mainstream, unless they have already done stuff on their own (startup) or are known in the open source community. The CS degree guarantees a channel to enter a company. I was reading a related article from Prof. Vivek Wadhwa about this (his take on students dropping out to do startups), and he encourages going to school (any school). Though, there is some truth I think in school providing a bigger picture/research that is hard to for some self taught folks to get to.

I self-taught myself programming for a while, discovered I liked it, and decided to go back to school in CS so that I could learn even more about it. (The first time through school, I was interested in different things and ended up with a philosophy degree.) While some of people I encountered in the CS department were there to line up a job, school can just be an efficient, focused learning method for the curious.

In general, one nice thing about hiring CS students as opposed to purely self-taught programmers is that, in my experience, they're more likely to have a good understanding algorithms and data structures. Self-taught coders often only learn as much as they think they need to know about the topic to complete the task in front of them. (Obviously, for those interested, it's as straightforward to teach yourself analysis of algorithms as any other topic.) For some types of programming positions, having a rigorous background in this area is crucial, though in others not so much.

What sort of job are we talking about?

Certainly, being a self-taught programmer and self-teaching your way through a CS degree are different things. Likewise, Computer Science as an academic field and Programming as a career are different. Other responses have mixed this up... expecting someone trained as Computer Scientist and to be good at programming jobs is, I think, silly.

The CS department at my Ivy league school did not teach very much that would qualify the students (my peers) for work. In light of that, I consider myself a self-taught programmer, even though I have a diploma with a CS degree from my school.

So, I don't think that it's about having a degree or about (claiming) to be self taught. If you are applying for a programming job, it's about your programming portfolio, wherever and however it exists.

Given that my own background[1] is a weird mishmash of formal education at various levels, and auto-didactic learning, I would be totally down with hiring somebody without requiring that they have a particular degree. As long as they can convince me that they have the required knowledge / skills, I really don't care if they're self-taught, university educated, learned through an apprenticeship, learned at a community college, or some hybrid of all of the above. I'm only interested in the end result.

[1]: For what it's worth, I was a junior when I dropped out of my bachelor's degree in CS program, but over the years I've accumulated 3 associate degrees: General Education, Computer Programming, and High Performance Computing, and have studied all sorts of stuff on my own time... just because it's what I found interesting. And I still work on learning new stuff each and every day. <shrug />

I am more inclined, as a rule, to hire self-taught programmers who have built something interesting over programmers with flashy degrees and little experience to back it up.

HN readers may enjoy EE380, the Stanford EE Computer Systems Colloquium. It is listed in this article, but is easily overlooked.

The Colloquium meets w4:15-5:30 throughout the academic year and moves re-runs during summer quarter. Lectures can be seen live, viewed in real-time over the web, viewed on-demand over the web, and eventually find their way onto YouTube, iTunes, and elsewhere. http://ee380.stanford.edu.

Each class has been announced individually, thought I'd give the full list.

Thanks for this. Do you know why the courses aren't all hosted on Standford.edu, though?

Yeah, see the above comments about how these classes are different from the www.(ai|db|ml)-class.com classes. They're not quite the same thing (albeit, they're quite similar).

Thanks. I've known about the iPhone application programming course from previous semesters, so this looks promising that they may offer it again this fall. I didn't have a mac until recently, so it will be nice if they do since xcode has changed a lot from what the courses available to download now use.

All the HCI semesters I think back from 2007 have been available on itunes University forever.

http://itunes.apple.com/WebObjects/MZStore.woa/wa/viewPodcas...

This is the future of education. Stanford is just the tip of the iceberg.

I just saw that the Machine Learning course mentions homework assignments involving Matlab. Might there be a way to use some open source software instead? I have no experience with Matlab, and not being a student anymore it could be expensive to buy.

GNU Octave is a free Matlab clone (though I doubt the GUI is as nice). They actually mention this on http://www.stanford.edu/class/cs229/materials.html.

Thanks, missed that on the web site

Searching on iTunes U doesn't always return these courses. Thankfully there is an iTunes link available from the lectures link on Course page.

These materials have actually been released in late 2008 - the three new courses will follow a completely different arrangement.

This is amazing to me. It hints at a new business model for education. Give away the education and charge for accreditation.

Anyone have any idea if either the AI or ML course might help someone interested in building Watson Jr?

https://www.ibm.com/developerworks/mydeveloperworks/blogs/In...

This is the database one: http://www.db-class.org/

I don't understand this move. To study at a renowned US university, you need to cough up a forture. To learn exactly the same but not get a piece of paper that says you did it, you can just take the classes online.

Don't get me wrong, I'm all for free stuff, but I'd feel pretty cheated if I'd be attending Stanford on half my relatives' latest savings and they use my contribution to give it all to freeloaders.

Note, I'm not against it per se, I just don't get it. Isn't this very unfair to the people paying the bill? Can someone explain the reasoning behind it so I can stop complaining about this? :-)

67,026 people have signed up for ai-class

The largest class ever ...

Booo....the video is in Silverlight.I cant watch it on any of the devices I own (besides a desktop :( )

I'm sorry for that. We had to build it in Silverlight at the time since Microsoft was a significant contributor to Stanford Center for Professional Development. However, you can also find the same original material on YouTube, iTunes or BitTorrent.

Harvard: http://cs50.tv/2010/fall/

Step One: Learn Bayes' Theorem

The iPhone class is from Winter' 2010. Any updates for Fall'2011 version?

Has anyone taken these courses and is willing to offer a review?

how to watch this without itunes? and what is this itunes thing, by the way?

iTunes has iTunes University where a lot of educational videos can be found. If I'm not mistaken you can find the same videos on YouTube at Stanford's YouTube channel: http://www.youtube.com/user/StanfordUniversity

You can also view them on this site: http://www.cosmolearning.com/ in the computer science department.

I've found this site to be really helpful. Many of the universities' lectures that are available online have been submitted here. makes much easier to find them.And I occasionally run into some pretty interesting documentaries and stuff. ;)

HN study group?

haha that would totally be worth it, as well as provide a little extra motovation to finish.

wetgwew

Reply on news.ycombinator.com