The Fate of the Incompetent Teacher in the YouTube Era

Up until recently, learners have had little freedom of choice as far as teachers went. If you were stuck with a bad teacher, you pretty much had to suck it up. You simply didn't have access to alternatives.

I remember being taught thermodynamics by a completely incapable teacher. I'm convinced everyone who sat in that class still has mental scars from the experience.

Not even the best students understood what he was saying. Whenever someone asked him a question, he would just go back one page in his notes and repeat himself word-for-word.

We quickly learned not to bother inquiring.

But what could we do? It wasn't like there were a half-dozen other teachers laying around, standing ready to provide a perfectly comprehensible explanation of the law of entropy at the snap of a finger.

Fast-forward to today, and that dream scenario is exactly what we're getting incredibly close to. Look at Salman Khan. He's the real deal. For a few years, the guy has been delivering a steady stream of clear explanations on hundreds of topics relating to mathematics, physics, finance and a few other fields.

His YouTube channel has more than a thousand videos. Each is about 10 minutes in length, which translates to a pretty thick stack of DVDs. Do you need to learn about the ideal gas equation? Khan's got it. The law of cosines? No problem. Moving pulleys? Check. Collateralized debt obligations? Sure. You name the topic, chances are it's already available, or will be soon. For free.

At this very moment, students all over the planet have just discovered Khan's treasure trove, and they are dancing around in their rooms, feeling blessed to have found someone who explains the subject they have to study this year in a way that they can understand.

Think about it. Even assuming, conservatively, that Khan's calculus videos are only slightly above average, roughly half the students taking calculus this semester would save time and pain by watching his lessons instead of paying attention to the mediocre teaching happening in front of them.

And I'm not talking about students who don't have a teacher, or eager minds who are stuck in a class below their ability level. The latent demand for this kind of stuff is huge.

"But these are just videos, not a real flesh-and-blood person you can interact with!" True. But I maintain that a great video compares favorably with a live, but bad, teacher in a classroom setting. You can't interact productively with a bad teacher anyway.

The Snowball Effect

Now, one of the great things about clarity of explanation is that most people tend to recognize it pretty quickly when they see it. Students who stumble upon Khan's videos remember him. They will go back to him; they will recommend him to their friends.

It is important to note that, thanks to YouTube's bandwidth, Khan's teaching scales very well. He has nearly 25,000 subscribers as I write this.

At some point, he will be helping a quarter million people learn.

Expect a similar dynamic to play out in every blackboard-teachable field with a standardized curriculum.

How fast is this going to happen? Well, Khan is already becoming famous. Last year CNN gave him airtime to explain the financial crisis. Why him, and not an economics Ph.D. type, you ask? Because he is understandable, and because some genius at CNN figured out that at least some of their viewers were able and willing to learn a little bit in order to understand what is going on.

So, for a change, instead of viewers being fed stodgy, professional-sounding but indigestible prose from a self-important expert, for several minutes there was a guy on TV with a pink tie and amateurish-looking drawings finally giving a simplified, but clueful explanation of the financial crisis and possible ways to get out of it.

In a fast-changing world, people are beginning to recognize the value of explanation.

Teacher Fame Goes Global

Good teachers have always had some measure of fame at the local level. Let's not kid ourselves: within a school, the students know who is a good teacher and who is no more illuminating than a wet pack of matches.

The net takes that to a whole different level. Eventually everyone will know who the good teachers are, and will be able to tune into them. They will be rock stars.

But what will happen to the bad teachers then?

There's a quote by Warren Buffett that I like to bring up from time to time: "It's only when the tide goes out that you learn who's been swimming naked."

Well, the incompetent teachers have indeed been swimming naked, and in a world where learners are free to tune into many other, competent teachers, it will inevitably show. When you have something to compare to, bad becomes tangibly bad.

Very well then. There can't be so many bad teachers anyway, right? Well... It only takes one extremely talented biology 101 teacher to raise the bar for all biology 101 teachers. In effect, the top 5% of teachers stand to make the other 95% look bad, if they put themselves to it.

Some of the poor teachers will look so bad that their students will simply laugh and walk out if they can, or tune out if they can't. They will only show up in class to get evaluated.

Of course, this kind of behavior will bring some questions into sharp focus, among them: "What good is it to pay an incompetent teacher to come in and give lessons that nobody actually listens to?"

Education systems typically move very slowly, so I don't expect incompetence to be magically chased all of a sudden because of the sudden availability of zero-cost, high-quality explanation.

Thus, it will be interesting to see exactly in what way the pressure coming from students (and their paying parents) will collide with institutional inertia.

Who knows? It could be that quite a few bottom-of-the-barrel teachers will have to find a new line of work.

I won't miss them.

Digging into Google Autocomplete Data

Following up to "What Google Really Thinks About Social Media", I put together a Yahoo pipe I called "Expose Google autocomplete data", which shows additional data about the Google autocomplete query of your choice.

The numbers below each result correspond to the value of num_queries provided by the Google Suggest API. I'm not sure exactly what they mean, but they are in all likelihood correlated with the frequency of the corresponding searches.

For some reason the Pipes widget doesn't let you pick a string, but the pipe itself does. Just click through to Yahoo Pipes if you'd like to try different queries.



Update: D'Arcy Norman: Google on Google. The snark bites back!

What Google Really Thinks about Social Media

I'm not making this up! See by yourself. (You may have to turn Google Suggest on.)

Further reading: danah boyd: is Facebook for old people?

Stocks, Flows, and Upkeep in Social Media

Last week a few folks including Michael Nielsen and I conducted a little experiment on Drawball. In case you don't know about Drawball, it is a kind of a "collaborative art" site where you are given a certain amout of "ink" which you can use to draw wherever you like on a gigantic, but shared, canvas. Your ink replenishes slowly, so you pretty much have to take breaks. Which is good if you have obsessive tendencies.

I find Drawball interesting because it provides a visualization of community dynamics. You see, some people go there and try to create nice things. Others are bent on destroying any and all nice things. People team up, so they can think big - and draw big. There are (good and evil) mobs. There are turf wars. There is drama.

The action on the ground is perhaps best described as ferocious pixel warfare. If you're wondering what the process looks like from 30,000 feet, watch this (there's a better quality video here):



I believe Drawball One is by now pretty much covered in protected areas that no one can overwrite, as a result of a preservation effort by the site's mastermind; On Drawball Two, however, it's the wild wild west.

Okay. On to our experiment. First we drew and colored in an "impossible triangle" of modest size. This didn't take too long and was not really difficult. Surprisingly, maybe owing to its small size, it stayed nice and clean for a few days.

Then we got started on something more ambitious: a fractal shape called a Sierpinski gasket. I picked this shape because it could grow without bounds from a small seed, and I figured its simplicity would make it relatively easy to repair, even for someone who didn't know exactly what it was. You can see the initial effort to the right.

Very quickly we had a budding replica of the seed, and one of the largest solid black areas on drawball. I think someone working alone would have taken a few days to do this, taking ink refill delays into account.



By the next day we had a pretty nice start. The shape had grown to a size that was noticeable on the whole-ball view (without magnification).

And that's when some vandals noticed us. They actually organized raids against our glorious gasket. Ink limitations mean raids are the only way to get any serious harm done quickly on Drawball. The bottom picture shows what the gasket looked like on day three. Nature abhors a vacuum, and those empty triangles obviously just looked too tempting.


Now, there is no revert button on Drawball like there is on Wikipedia; every vandalized pixel requires a pixel of repair. We didn't really try to fight the vandals. It was easy to redraw outlines, but filling in the shapes again is time- and ink-consuming.

The experiment is pretty much over by now. The only interesting thing that is left to do is to observe and measure decay. The structure is large and regular enough that I don't expect that it will vanish too soon, but it will definitely get dirtier every with every passing day. You can visit the region today and see if you recognize anything.

So, what have I learned? Well, the first tentative "Drawball law" that becomes apparent very quickly is this:

Drawball Law no. 1. Noticeability brings attention; attention brings participation.

A second law would be:

Drawball Law no. 2.
The effort required for upkeep grows in proportion to the surface area of your turf.

On Drawball strength lies in numbers, not finesse; consequently, if one person is able to maintain a drawing of area X, it will take two people to maintain an area 2X.

Beyond Drawball

How general are these laws? I suspect their applicability goes beyond Drawball. Here's an application of Law no. 2 to wikis and blogs.

To any social media site there is a visitor-editable, "living" part and a non-editable, "still" part. The living portion of a weblog is typically the very edge of the flow - the "now"; the rest of it is still life. For a wiki, the living portion is essentially the whole site (though at any moment some parts are typically more alive than others).

Now, the content of most blogs arguably feels more well-kept than most (open) wikis, even though wikis have a whole community around them and blogs typically have a single maintainer. Why is this?

I believe there are several underlying reasons, but I'm wondering if the chief one might not follow from Drawball Law no. 2. Wiki builders expand their turf with every paragraph added, each page created. If the community doesn't grow with the wiki, its turf becomes unmaintainable. By contrast, as a blog grows, its past content becomes museum material - it quickly fossilizes and requires very little upkeep.

To conclude using the terminology of stocks and flows, I would assert that as a general rule, participative flows (blog-like media) are easier to maintain than participative stocks (wiki-like media).

I leave design-for-community implications as an exercise for the reader. :)

What do you think?

More reading about drawball:

• Drawball: from Chaos to Community - a nice essay positing that the social aspect is the reason artists bothered visiting and revisiting Drawball.
  


• Wikipedia article on Drawball. Alas, Wikipedia doesn't consider Drawball notable.

Citizen Science at SciBarCamp II: Semantic Eco-Blogging!

At the last SciBarCamp, Joel Sachs from the SPIRE team at university of Maryland presented a very inspiring project that opens environmental/bioscience in a way that enables regular people to get involved usefully.

First, a bit of background. Every living species has a certain geographic distribution that is known to evolve over time, especially in recent years due to global warming. This data is extremely useful for environment scientists. As you might expect, though, comprehensively cataloging which species live where is a monumental undertaking, especially if you have to repeat the effort every year!

Now, most people nowadays have digital cameras and internet access. Some gadget freaks actually have both, plus a global positioning system (GPS), all in a neat package, which makes it ultra-easy for them to go out in the woods and snap pictures of plants and animals around them. Special days called BioBlitzes have emerged recently where many people go out all at the same time but in different locations, to do just that. Then they report what they found on their respective blogs.

So where and how does it feed into science? Well, Sach's project provides Spotter, a Firefox extension, and a spreadsheet-to-RDF utility called RDF123, that together enable participants to easily do all of the following:

• Snap a picture of a specimen;
  
• Upload it to Flickr;
• Provide geographic information on where the photo was taken;
• Tag it with descriptive terms, as specifically as they can;
• If needed, request further crowdsourced identification (volunteers can watch RSS streams corresponding to certain tags like "bird" or "butterfly").

Faithful readers of mine (who have good memory!) will recognize this as a neat incarnation of that old structured blogging idea. The result of each contribution is stored in the RDF representation on a central aggregating database managed by SPIRE. This enables semantic queries to be made, such as: "Show all observations of species that are classified as being of concern by the U.S Fish and Wildlife Service" or "What was the northernmost spotting of the Emerald Ash Borer last year?".

The overall result is obviously a global, collaborative database of samples from everywhere that gets richer with each participant's contribution. For participants, obvious motivations include:

• an interest in natural observation (think about birdwatchers);
• the satisfaction of contributing to a worthwhile scientific endeavor;
• the pride of offering something quite unique (info sampled at your particular location).

I could totally picture 6th grade classes going out and having a blast doing this. (If you're an educator reading this, think about doing it!)

So far there are about 1200 observations in the system, but hopefully this number will grow as more people join the effort. I don't know too much about the viral properties of participation, but for starters, the fact that photos are posted into each people's Flickr stream exposes the project to their friends.

All in all, I think this is a very promising effort that exemplifies how amateurs and pros can complement each other given the right tools. There are obvious parallels to draw between this effort and other citizen science endeavors such as GalaxyZoo, which harnesses people's ability to classify galaxies better than computers, or the OpenStreetMap project, which aims to create free datasets of where streets are everywhere in the world.

SciBarCamp II, Toronto

I'm writing this on the train back from SciBarCamp II in Toronto.

To briefly sum up the experience, I've just spent an intense 24 hours in the company of quite a few admirable thinkers.

As the name suggests to people who know about BarCamps, the event was set up as an unconference. Attendance was on invitation and limited to a hundred participants. The chosen theme was "Open Science".

It really felt like the organizers knew what they were doing. No time was wasted. Pretty much all of the participants (two thirds of whom actually weren't at the first SciBarCamp) slipped seamlessly into the unconference mold.

The participant makeup was quite diverse, with maybe half of the participants being scientists and others being businesspeople, technologists and artists with a strong interest in science. I was struck by the all-around open-mindedness and ability to speak in a jargon-free manner.

---

The event went as follows: At the Friday evening reception, we spent 30 minutes on personal introductions, then had a mix & mingle session where people basically talked to whoever they liked, and filled in session proposal sheets. Around 30 such sheets were filled in. Then, participants were invited to go around the sheets and were able to comment on them and indicate interest by filling in circles below the session descriptions.

The interest indicators enabled the organizers to build Saturday's schedule on the fly. Rooms were allocated to sessions, with the more popular sessions obviously being put into larger rooms. A few proposals were merged together. The final schedule had five one-hour time slots, with 5 parallel sessions underway at any one time.

You can have a look at the schedule here. Two sessions especially stood out for me. The first was Joel Sachs and Jesse Greener's session on citizen science. The second was Michael Nielsen's talk on open mass collaboration in mathematics. I will cover them in upcoming posts.

I especially appreciated the opportunity to meet again with WorldChangers Hassan Masum, Mark Tovey and Karl Schroeder, and I was stoked to shake hands with quantum-information-authority-turned-science-futurist Michael Nielsen for the first time. It was also a great pleasure to meet new people like physicist Rob Spekkens and Kaitlin "I am not a lawyer" Thaney from Science Commons.

Huge kudos to Eva Amsen, Christine Buske, Jen Dodd, Jamie McQuay, Mark Tovey and Sunny Tsang, who pulled this event off, well, flawlessly on a shoestring budget.

I think a Montreal SciBarCamp will obviously need to happen at some point. However, it feels like the requisite network has yet to gel together. (Either that, or I haven't connected to it for some reason.)

Update: Here is Glendon Mellow's report on SciBarCamp.

Quick Guide: Four Easy Ways to Protect your Gmail Account

If you are using Gmail as your main email account, you'll readily admit that you'd be in a tough spot if you were to lose access to it. But these things do happen, be it because you forgot your password (most likely right after changing it!) or because your archenemy (or some random joker) has been able to hijack your account. Hey, if it happened to Kanye, it could happen to anyone :)

If you have registered for various services on the web using your Gmail account, things could get really messy. "I forgot my password" links would no longer be of any use, as they send help to an account you can no longer enter. Moreover, a hijacker could use your Gmail as a gateway to gain access to your accounts on those services.

The long and short of it, really, is that it would be best if you didn't lose your Gmail account.

I did a quick search today to see what precautions I could take to protect against this and found out que you and I can reduce risks in four easy ways. (The links below work for me today; your mileage may vary.)

1. Always use the secure https protocol (look at the bottom of the page) to communicate with Gmail.
   
2. Make sure you have chosen a secret question to get your password back.
3. Make sure you have entered a secondary address in your Gmail settings.
   
4. Pick a password that is not easy to guess.
   

In this guide you can find additional info, including a discussion of what your options are if you find yourself in the unfortunate position of having lost access to your account.