After just talking about it for a while, I'm finally starting an online campaign to get Khan on the TED stage. He has an idea very much worth spreading. Hell, they even let me speak at TED, so please take a minute or two to nominate Salman Khan here on TED's speaker nomination form: http://www.ted.com/nominate/speaker
And if you've got another minute, nominate Paul Graham, too!
(edit: I'm an employee of Y Combinator, making PG my boss, but I'd nominate him regardless of that fact.)
1/. Videos are short, 10-15min digestible bits. Long enough to teach a concept, short enough not to lose my attention.
2/. He's a huge advocate of teaching intuition behind concepts instead of memorization. This makes learning less stuffy and more informal.
3/. Everyone has "holes" in their knowledge, but they are in different areas. Schools have a hard time to individually address weak points in each student's knowledge (particularly if you should have learned them in a prerequisite course). Videos organized by concepts allow you to address these weaknesses systematically.
4/. He makes mistakes! Those mistakes are quite insightful and give a glimpse into an art of solving problems that usually isn't presented in most lectures. You usually see a problem, and a solution thats a finished product.
5/. He doesn't have an ego. He's not breathing down your neck or thinking you're an idiot if you don't understand anything the first time. You can pause, google something, repeat, watch it anytime, etc.
Though I'm a fan of Feynman anecdotes, I've always disagreed with the point of the one used in this article. Names are important. You know what you get by knowing the name of the brown-throated thrush? You get an identifier by which to look up all the collected knowledge of humanity on the brown-throated thrush, and you gain the ability to talk about it with other people.
In the online SICP lectures there's a nice point about names and how they give you power over things, but I couldn't find it right away.
Feynman's brown-throated thrush anecdote isn't saying that names don't serve a valuable purpose. As you've shown in your examples names are a tool to help us reference knowledge, but they are not the knowledge itself. Unfortunately people don't always make the distinction. This way of thinking may lead them to collect many references but very little real knowledge. That's the point I think Feynman was trying to make.
I find this point the trivial one. The surprising point is that names really do matter. I had no idea how many sycamore trees there are in NYC before I had a name for them; having a name also meant having a set of characteristics that let me label a tree as a "sycamore" - being able to apply a name to a thing also means recognizing the thing, and being able to communicate about it.
(Actually, probably London Plane trees, but don't let my ignorance get in the way of my point.)
He wasn't making a comment that the names don't matter - he was trying to rationalize why he gained an interest and appreciation for the deep-knowledge of science - not just how to look up reference material.
Names obviously matter - language and communication matter if we're going to collaborate on anything - but when it comes to science, to make new discoveries, the ability to look something up in a book is useless without the ability to understand how things actually work and fit together... and sometimes, books, and even the majority of people, are actually incorrect...
It's disingenuous to take that Feynman quote out of context and argue that it's an argument against naming... it's certainly not.
Indeed. In Darwin's youth, he was obsessive about collecting and cataloging insects. I wouldn't exactly attribute his theory of evolution by natural selection to naming things, but certainly he would never have arrived at it if he was only able to name one kind of beetle. He had to be able to differentiate between thousands of beetles before he could see how they were connected.
I think Feynman's quote is meant that you should not learn only the name of things and pretend to understand what they are.
There's also another anecdote written by Feynman where he describes creating his own notation for trig, or maybe calc. At some point he had to abandon it because no one else understood him. The anecdote is not as quotable, but it correlates with what you said -- it gives you the ability to talk about it with other people, so Feynman is not ignorant of this.
The Khan Academy content is currently concentrated in areas such as math, science, and finance. Do you think the approach can be applied to other disciplines; e.g., programming, business (other than finance), other social sciences? And how about more advanced levels of skill development?
In short, are there boundary conditions to the Khan Academy approach? And if so, what are they?
Khan Academy currently heavily depends on examples. Anything that you can teach by example is probably fair game. For example (ha!), his history stuff right now, while perfectly fine, I find no more engaging/effective than reading about.
That doesn't mean it won't work, just it probably won't be as effective for some other stuff as it has been for math and such.
I agree. I think making it work beyond math and science will be a bit challenging, I think the style would need to change quite a bit, away from just the white board and more toward something like a real lecture.
mr. Khan and wikipedia are the two best things to come out of the web as far as I'm concerned. Project gutenberg third but at considerable distance.
Between Khan Academy and wikipedia education is pretty much limited by the amount of time you have to invest.
There was a character in a book that tried to learn by memorizing an encyclopedia, which as 'learning by rote' is not the most useful thing. But in wikipedia everything is interlinked, and you can dig down in to a subject as deep as you want, the references will give you the background. It's quite amazing how much you can learn about something just from there.
The Khan Academy is another part of education embodied in bits, a really good instructor that knows how to make subjects both interesting and understandable.
The Khan Academy is great. But, and I'm not really making any accusations at all, I hope amidst all this praise for what an admirable job he is doing and all the opportunities this affords for underpriviledged people that he stays on top of the quality and accuracy of the material. I just remember there being a few questions about some of his stuff on linear algebra. I also hope he will continue to add more advanced material to the various subjects, but I think that is his plan.
I think that's what's great about the whole concept, if you want to replace a lecture with a better one at a later date you actually can, so in the long term it can only get better, this just establishes an initial baseline, nothing will ever make it worse than this.
The success of the Khan Academy is by orthodox educational standards surprising. That it is surprising ought to tell us something: that the orthodox understanding of how education should be "delivered" is at least partly wrong.
We should be particularly suspect of the belief that higher production values are worth pursuing. Before KA, if you suggested that you could deliver unusually effective education through a series of ten-minute videos having truly horrible production values, the mainstream education industry would have ignored you. After KA, they can't.
What the success of the Khan Academy suggests is that production values are not important and, perhaps, that the pursuit of high production values is wasteful and even counterproductive.
This kind of surprises me. I love Khan's lecturing style, and he makes good use of a paint program to stand in for not having an overhead transparency to scribble on.
But as to his teaching style, the videos I've seen have caused me to react less like Patrick Mylund and more like this MetaFilter poster:
Perhaps there's an ordered curriculum you're meant to go through, but every time I've sampled Khan's videos at random they've been structured as "if these are the exercises you need to do right now, now you can see me solve one."
Sure he does a good job of narrating his steps, explaining details and side-effects, and making illustrative mistakes. I just feel like he's all about the how and never about the why.
I watched a couple of those videos after reading about it here and, while they were nice, they didn't stand out as something very special to me. Maybe it's because his style feels close to the way that I like to teach, or maybe you guys had way more sucky teachers than I did (though I don't think my math teachers were particularly good).
I don't want to disrespect the effort he's going through setting up all the online content, that's clearly something very useful to those that don't have access to real people to learn with. But to hold him up as as especially awesome and unique seems to ignore the thousands of equally awesome teachers that are out there teaching like that every day.
I would say he hits the sweet spot between 'inanimate book' and 'personal teacher' pretty much dead on.
Computers and internet connections reach places on the planet where very few physics books go, and embody potentially every book ever written + all the free content available out there.
I had the same impression. Nothing can educate like a good teacher who coaches you directly. Looks like many americans (or britons) are not getting this individual attention in their schools and colleges.
Some professors are able to teach you complex things in such simple terms that it damn well blows you away. I remember being tought van der walls equation's proof using some very basic concepts and the taylor series. I can't find that simple a proof on wikipedia now, and unfortunately I have forgotten much of physics. SICP is also a good example of a good teacher making complex problems simple.
You can check out my lectures here http://inperc.com/wiki/index.php?title=Category:Courses (math only). These are written lectures but they start at about the same place as Khan's (in a real classroom though). Hundreds of illustrations but I haven't figured out the "interactive" part yet.
Khan Academy is a good resource if you haven't a good teacher and you have a lot of time for learning.
With a good teacher you can learn more with less time. If you are in a small group you can ask the teacher and you will learn that he can give you not only the answer you are looking for but a new perspective about that subject.
So Khan Academy is a good resource in certain occasions, but never as a good teacher should be.
Why I love ? Because almost after a decade from high school class, I realize why it is that way. His lessons are short, engaging and teach you the basics. I could only wish I had someone with quarter of Khan's talent to teach me math and science.
I'm taking Math and Physics at uni in a pre-engineer year. Khan Academy is making me learn a lot! My physics teacher is not that good, Sal however is an amazing communicator. If you only visit the site shortly you will only see the youtube videos but looking closer you will find the real gem: The application guiding you through math and physics. If you can answer all the questions correctly you know have learned everything you need to know to pass a exam! :-)
I would highly suggest watching the Walter Lewin videos from mit. They are what got me through a mediocre at best, terrible at worst teacher in physics. I used him heavily for electricity and magnetism and some for the classical mechanics stuff. Excellent teacher!
Cool thanks! I have a test coming up this Wednesday on the first chapters: Motion in a straight line and Force and Motion. I understand the basics/theory behind it but i need to solve more problems. :-)
My current physics teacher's philosophy, as I understand it, is that he teaches something once, if no one says they need more instruction he moves on, and he does not work out problems to a final answer in class. His explanation is that he wants us to learn to learn on our own and ask questions, but of course he won't answer questions if they relate to material we should already know.
I haven't been able to figure out if this is a good thing or bad thing, rather it may come down to how students learn and what is the actual job/purpose of the instructor. If you learn the material before his lecture then there is no problem but then there is little reason to come to class other than it being a requirement and to enjoy the half of class he isn't instructing, discussing random topics and shooting the shit.
That kind of makes me sad, physics can be so engaging with a good teacher even though physics turned out to be my hardest non engineering course because of the teacher's assigned workload and tests it was by far my favorite class to attend.
I struggle with proving big oh, thus I wonder if (Kahn) or anyone else can help me prove that:
t(n) <= cg(n) for all n >= 0
and here is the example I do not grok, where I'm gonna prove this assertion: 100n + 5 ∈ O(n^2).
This proof goes like this:
100n + 5 <= 100n + n = 101n <= 101n^2
I struggle basically with wrapping my brain around the proof there. What is going on?