10 ms·

Where robots fail: Why education can’t just be digital

41 points by akharris 13y ago

j_s 13y ago
> This is a guest post by Tutorspree CEO Aaron Harris [...] > a New York-based startup that finds the perfect tutor for every student.
- pyre 13y ago
  The same could be said of advocacy for online-only education by (e.g.) the Kahn Academy.
  - j_s 13y ago
    Interesting that you picked that example! From the article: > Sal Khan isn’t pushing for his videos to be the final say in teaching. > They are meant as a gateway to unlock significantly higher rates of > one-to-one teacher student interaction in the flipped classroom.
    - pyre 13y ago
      It was just an example, Sal Khan seems like a reasonable guy, so I wouldn't have expected him to take a hard-line, but they were the first 'online only'-type education startup I could think of.
droithomme 13y ago
Yet another weekly post saying the same thing about completion numbers, the only argument they have. According to this argument, 1500 students from all nations of the earth successfully completing a CS class at Harvard, provided at absolutely no cost to them, is a complete failure. Meanwhile they don't mention how the schools they represent have programming weeder classes with 900 people enrolled at the start and 100 at the end. This is considered OK since 800 officially dropped the class and the class is supposed to be tough. Every week, the same argument. If 90% completed the class the exact same people would write their PR articles claiming that the classes have failed for not being rigorous enough.
- akharris 13y ago
  Author here - that's not at all what I said. I'm saying these courses set their own standard of measurement, and haven't lived up to it yet. They're iterating by bringing in different support systems to raise completion rates. I hope the MOOC model succeeds, but it's a long way from fulfilling its promise thus far.
  - droithomme 13y ago
    > I'm saying these courses set their own standard of measurement No, you are claiming that and saying that that is their standard. It's a straw man, which is a dishonest form of argument. Your arguments here are without merit. If you want to succeed with your tutoring company you need to establish yourself as a reliable and honest person, not someone trotting out absurd self serving arguments attacking other ventures that you incorrectly and very foolishly see as your enemy. To succeed focus on your product, make it good, and show how it benefits people and is worth paying for. Focusing on these other people instead is a waste of your time and shows that you presently don't understand either what to do to succeed in your own market, or how to run a startup business. You might take what I am saying as criticism, but it is highly targeted and good advice. You should think about it. Ignore it at your own peril.
    - cinquemb 13y ago
      Exactly. People seem to expect that MOOCs/or the like are for the same old feed-people-into-the-system. It's like people can't see the glass houses around them with all the stones flying around, to see that something else will emerge when their glass house shatters and not somebody elses. Long student loan defaults and tuition hikes.
  - barryfandango 13y ago
    I'm currently halfway through "Introduction to Improvisation" on Coursera, offered through Berklee college and taught by the great Gary Burton. The enrollment is around 39,000 and I'm sure the completion will be low. People worldwide and of all skill levels are welcome to ignore the prerequisite guidelines and there's no cost to sign up. I bet twenty thousand didn't make it past week one - either discovering how difficult the class is, or losing interest since they only signed up on a whim anyway. This low barrier to entry probably inflates enrollment considerably compared to a traditional school. There is a human element you did not address in your article. The discussion forum for the class is very active and full of enthusiastic students helping each other, asking questions, sharing resources, posting recordings of their music work, providing suggestions to the Coursera staff about curriculum improvements (many of which have been implemented already.) So it's not all robots.
tbatterii 13y ago
misleading headline. there is no mention of robots teaching anyone anything or even how robots fail. How's tutorspree any different than sylven? What was the point of this article again?
purplelobster 13y ago
Personally, I see it as a great opportunity. Today you need a great teacher that can sit down and figure out why a student has problems understanding a specific concept etc, and explain in many different ways. I don't see how computers can't do that. Online education will actually level the playing field and make the best material and pedagogy available to everyone. Today, if you get a lousy mathematics teacher, you're set back years potentially. I know, I had one for 3 years in high-school. Compared to my peers, it really set me back. Had to work very hard to catch up, but fortunately I was motivated enough to do it. While Khan Academy might be good today, what I think would help even further is to find the absolute best teachers out there and "capture" the way they teach. Capture all the questions students have and incorporate them into the videos. For example, if there's a video talking about the number PI, the student should be able to pause in the middle of the video and see commonly asked questions and answers about this portion of the video. This also has the great benefit of students being able to go through the material at their own pace. Today teachers have to either leave half the class behind, or spend too much time on the weaker students.
- zwieback 13y ago
  I think this is heading in the right direction but teaching methods of the best teachers have been studied for years and I don't think the algorithm is simple. I have no doubt that computers can be better than the worst teachers but I think we're far a way from beating the best ones. I'm surprised at the negative comments here, I'm guessing there are a lot of education software entrepreneurs who somehow interpreted the article as a swipe at their industry. I thought it was a good counterpoint to the hype around online courses although it was also clearly a plug for another product. My kids are already growing up with a mix of online and warm-blood teachers. They talk a lot about how they like or dislike this or that teacher and they try very hard to succeed in the classes of the teachers they respect. They use Khan Academy and stuff like that but it's just this thing on the web they can draw on, they have zero emotional connection to it.
  - purplelobster 13y ago
    Let me clarify that I think computers would help the most with explaining concepts in an engaging, pedagogical and interactive way. Having a personal tutor that helps you with your homework and example problems is much further into the future. But my biggest problem in school was not that I didn't get help to solve problems and do homework, it's that the teachers often didn't explain high-level concepts in a way that made it click for me. It's the one-to-many teaching that is especially bad today if you get a teacher that either can't explain well, or just doesn't understand the subject deeply enough to explain it well. Edit: As an example, listen to Feynman talk about anything and everything, and compare that with your average or below average college lecturer. The difference in your understanding is immense. A good teacher helps you build a mental model of the world and gets you excited about it. A bad teacher just regurgitates what the course material says with no real engagement. I wish I could have had Feynman-level lectures on everything from the start, it would have saved me years.
adaml_623 13y ago
"While MOOC proponents argue that these completion rates are irrelevant because of varied interest levels of participating students at the outset, they’ve set the stage for their own measurement by using enrollment factors as their topline metric in interview after interview." What's the term to describe this logical fallacy whereby you dismiss a valid proposition by quoting something that your opponent has said about something different. Completion rates and Enrollment numbers are connected but both must be weighed separately when judging online education. (Also where are the robots :-( )
- infectoid 13y ago
  It sounds close to a Strawman argument but not quite... https://yourlogicalfallacyis.com/strawman https://yourlogicalfallacyis.com/strawman
BasilAwad 13y ago
Summary: MOOCs have (1) really really low completion rates and (2) a user base that is already highly literate and educated. So MOOCs don't target the majority of students who actually need a person to direct their learning. (Author then mentions his tutoring directory company called Tutorspree). Then the author cuts MOOCs some slack and says they are hiring TAs and encouraging meetups.
btian 13y ago
The article is a massive fail. I don't see where the "robot" is. Is Sal Khan not a teacher? Andrew Ng is a "robot"?
- rimantas 13y ago
  How do you define a teacher? If it is anyone who teaches, then yes. If it is someone who understands most of the issues related to teaching, then no, by a long shot.
  - tbatterii 13y ago
    > If it is someone who understands most of the issues related to teaching, then no, by a long shot. I would love some more details on this.
    - saraid216 13y ago
      As I understand it, pre-college teachers generally have Actual Training for dealing with kids, like child psych.
      - tbatterii 13y ago
        I don't know if that's what rimantas meant though.
alan-crowe 13y ago
The article made me think about home schooling. Some home schooling is done for religious reasons, but other home school is motivated by academic concerns. There is not so much of the second kind because it raises an awkward question: how is the parent to do better than school at teaching a subject about which they themselves are ignorant? Enter the MOOC. The content is available, at a level and pace to suit a child. The stay-at-home parent can learn along, providing the child with both guidance and flesh-and-blood encouragement on learning and learning to learn. I anticipate that MOOC will change the balance in home schooling with a big expansion of the academically motivated aspects.
mattangriffel 13y ago
This article and others like it are arguing against a strawman. No one ever proposed removing teachers from the equation. The idea is to leverage technology to allow teachers to reach more students and/or teach more effectively. Sure, no online educational platform can rival a good teacher in a classroom setting yet, but that's because a good teacher is able to pick up on and respond to a lot more data through a physical interaction (a student's posture, tone, and other social cues) than a computer can online. That being said, in theory an online education platform should be able to track all that and more. They potentially have access to an incredible amount of data including the exact amount of time a user takes to engage with a lesson, how fast he or she is typing / is able to complete tasks, even the varying levels of difficulty at which different concepts are being learning, etc. This is all stuff that none of the current online education platforms have even tried to do anything with yet. But to say that an online education platform will never be as good as a teacher in a classroom is like going back to 60s and saying that computer image recognition sucks so we might as well stop trying.
ivankirigin 13y ago
Founder of company that specializes in adding humans to online educations says that you need humans in online education. Self promotion doesn't need to attack other good causes. There is a bit in the article about the failing brick & mortar institutions, but that should really be the entire focus. That is the only real story here. Online efforts are scrambling as offline continues to be a clusterfuck.
aaron695 13y ago
Point out current problems with X. Somehow proves X is never possible. A real hacker would think, ok robots don't work yet, how can we make them work.
jholman 13y ago
I think this article is all over the map, and I hope akharris is a better entrepreneur than writer. Though he makes one or two good points. The title is total bullshit. "Linkbait" is giving it too much credit. "Robots" is a complete head fake, and the article literally says nothing at all about whether or not education can be totally digitally mediated. The completion-rate issue. First of all, I think we can all agree that completion rate is at least somewhat interesting, and the link to Katy Jordan's piece is useful if you haven't seen it yet. And while I agree it's not a HUGE deal, for the reasons many others have gone over, akharris makes a good point, as follows: The MOOCs opened themselves up to this criticism, because over and over again they post their vanity metrics of enrollments, and compare them to classroom enrollment, and that's crazy. It's like comparing Facebook signups to private plane purchases. If they had been bragging about completion numbers (not ratios, obviously) all along, they'd be fine. Anyway, my point here is that the focus on the completion rates isn't the worst part of this article, it's pretty near the best part. The next chunk talks about K-12. Obviously most of the MOOCs were never aimed at K-12, so it's an evasion. If you look at the forum conversations or the StackExchange sites or whatever, there are obviously lots of students who are adults from eastern Europe, or Asia (and lots who appear to be middle-class Americans, of course). That's access to education, baby. The rhetorical deceit here actually enrages me. I might make a perfectly-parallel criticism of Tutorspree: "[Harris's] company states that its mission is connect students with great private tutors. So who is getting those tutors? Certainly not Tamil-speaking children in Sri Lanka." What a stupid argument I just made. Just like akharris's argument, that because 7-year-olds aren't benefiting from Sebastian Thrun's robotics courses, he's therefore not living up to Udacity's lofty mission statement. And of course, Khan Academy IS improving access to education. Not in some ideal future, but in the present and in the past. Without one-to-one tutoring, which by the way is obviously beyond the budget of, you know, most families (average price for tutors in my Zip: $88/h). So where's your access now? Admittedly Khan requires an internet connection, but $50/month for internet that the whole family can share, plus a $250 computer, sounds better than $88/h to me. Though it is nice that akharris isn't duplicitous enough to grossly misrepresent Sal Khan's vision for an ideal classroom. Nice work there. The weird part is that the thesis that he's trying to weave, about the value of human contact and personalized teaching, is kind of a no-brainer. And although I think he overstates the awesomeness of his business, I'm willing to assume it's more part-of-the-solution that part-of-the-problem. So why did he have to stoop to all these dodgy half-truths and misdirections? Maybe he needs to hire a composition tutor.

41 points by akharris 13y ago

j_s 13y ago

> This is a guest post by Tutorspree CEO Aaron Harris [...] > a New York-based startup that finds the perfect tutor for every student.

pyre 13y ago
The same could be said of advocacy for online-only education by (e.g.) the Kahn Academy.
- j_s 13y ago
  Interesting that you picked that example! From the article: > Sal Khan isn’t pushing for his videos to be the final say in teaching. > They are meant as a gateway to unlock significantly higher rates of > one-to-one teacher student interaction in the flipped classroom.
  - pyre 13y ago
    It was just an example, Sal Khan seems like a reasonable guy, so I wouldn't have expected him to take a hard-line, but they were the first 'online only'-type education startup I could think of.

droithomme 13y ago

Yet another weekly post saying the same thing about completion numbers, the only argument they have. According to this argument, 1500 students from all nations of the earth successfully completing a CS class at Harvard, provided at absolutely no cost to them, is a complete failure. Meanwhile they don't mention how the schools they represent have programming weeder classes with 900 people enrolled at the start and 100 at the end. This is considered OK since 800 officially dropped the class and the class is supposed to be tough. Every week, the same argument. If 90% completed the class the exact same people would write their PR articles claiming that the classes have failed for not being rigorous enough.

akharris 13y ago
Author here - that's not at all what I said. I'm saying these courses set their own standard of measurement, and haven't lived up to it yet. They're iterating by bringing in different support systems to raise completion rates. I hope the MOOC model succeeds, but it's a long way from fulfilling its promise thus far.
- droithomme 13y ago
  > I'm saying these courses set their own standard of measurement No, you are claiming that and saying that that is their standard. It's a straw man, which is a dishonest form of argument. Your arguments here are without merit. If you want to succeed with your tutoring company you need to establish yourself as a reliable and honest person, not someone trotting out absurd self serving arguments attacking other ventures that you incorrectly and very foolishly see as your enemy. To succeed focus on your product, make it good, and show how it benefits people and is worth paying for. Focusing on these other people instead is a waste of your time and shows that you presently don't understand either what to do to succeed in your own market, or how to run a startup business. You might take what I am saying as criticism, but it is highly targeted and good advice. You should think about it. Ignore it at your own peril.
  - cinquemb 13y ago
    Exactly. People seem to expect that MOOCs/or the like are for the same old feed-people-into-the-system. It's like people can't see the glass houses around them with all the stones flying around, to see that something else will emerge when their glass house shatters and not somebody elses. Long student loan defaults and tuition hikes.
- barryfandango 13y ago
  I'm currently halfway through "Introduction to Improvisation" on Coursera, offered through Berklee college and taught by the great Gary Burton. The enrollment is around 39,000 and I'm sure the completion will be low. People worldwide and of all skill levels are welcome to ignore the prerequisite guidelines and there's no cost to sign up. I bet twenty thousand didn't make it past week one - either discovering how difficult the class is, or losing interest since they only signed up on a whim anyway. This low barrier to entry probably inflates enrollment considerably compared to a traditional school. There is a human element you did not address in your article. The discussion forum for the class is very active and full of enthusiastic students helping each other, asking questions, sharing resources, posting recordings of their music work, providing suggestions to the Coursera staff about curriculum improvements (many of which have been implemented already.) So it's not all robots.

tbatterii 13y ago

misleading headline. there is no mention of robots teaching anyone anything or even how robots fail. How's tutorspree any different than sylven? What was the point of this article again?

purplelobster 13y ago

Personally, I see it as a great opportunity. Today you need a great teacher that can sit down and figure out why a student has problems understanding a specific concept etc, and explain in many different ways. I don't see how computers can't do that. Online education will actually level the playing field and make the best material and pedagogy available to everyone. Today, if you get a lousy mathematics teacher, you're set back years potentially. I know, I had one for 3 years in high-school. Compared to my peers, it really set me back. Had to work very hard to catch up, but fortunately I was motivated enough to do it. While Khan Academy might be good today, what I think would help even further is to find the absolute best teachers out there and "capture" the way they teach. Capture all the questions students have and incorporate them into the videos. For example, if there's a video talking about the number PI, the student should be able to pause in the middle of the video and see commonly asked questions and answers about this portion of the video. This also has the great benefit of students being able to go through the material at their own pace. Today teachers have to either leave half the class behind, or spend too much time on the weaker students.

zwieback 13y ago
I think this is heading in the right direction but teaching methods of the best teachers have been studied for years and I don't think the algorithm is simple. I have no doubt that computers can be better than the worst teachers but I think we're far a way from beating the best ones. I'm surprised at the negative comments here, I'm guessing there are a lot of education software entrepreneurs who somehow interpreted the article as a swipe at their industry. I thought it was a good counterpoint to the hype around online courses although it was also clearly a plug for another product. My kids are already growing up with a mix of online and warm-blood teachers. They talk a lot about how they like or dislike this or that teacher and they try very hard to succeed in the classes of the teachers they respect. They use Khan Academy and stuff like that but it's just this thing on the web they can draw on, they have zero emotional connection to it.
- purplelobster 13y ago
  Let me clarify that I think computers would help the most with explaining concepts in an engaging, pedagogical and interactive way. Having a personal tutor that helps you with your homework and example problems is much further into the future. But my biggest problem in school was not that I didn't get help to solve problems and do homework, it's that the teachers often didn't explain high-level concepts in a way that made it click for me. It's the one-to-many teaching that is especially bad today if you get a teacher that either can't explain well, or just doesn't understand the subject deeply enough to explain it well. Edit: As an example, listen to Feynman talk about anything and everything, and compare that with your average or below average college lecturer. The difference in your understanding is immense. A good teacher helps you build a mental model of the world and gets you excited about it. A bad teacher just regurgitates what the course material says with no real engagement. I wish I could have had Feynman-level lectures on everything from the start, it would have saved me years.

adaml_623 13y ago

"While MOOC proponents argue that these completion rates are irrelevant because of varied interest levels of participating students at the outset, they’ve set the stage for their own measurement by using enrollment factors as their topline metric in interview after interview." What's the term to describe this logical fallacy whereby you dismiss a valid proposition by quoting something that your opponent has said about something different. Completion rates and Enrollment numbers are connected but both must be weighed separately when judging online education. (Also where are the robots :-( )

infectoid 13y ago
It sounds close to a Strawman argument but not quite... https://yourlogicalfallacyis.com/strawman https://yourlogicalfallacyis.com/strawman

BasilAwad 13y ago

Summary: MOOCs have (1) really really low completion rates and (2) a user base that is already highly literate and educated. So MOOCs don't target the majority of students who actually need a person to direct their learning. (Author then mentions his tutoring directory company called Tutorspree). Then the author cuts MOOCs some slack and says they are hiring TAs and encouraging meetups.

btian 13y ago

The article is a massive fail. I don't see where the "robot" is. Is Sal Khan not a teacher? Andrew Ng is a "robot"?

rimantas 13y ago
How do you define a teacher? If it is anyone who teaches, then yes. If it is someone who understands most of the issues related to teaching, then no, by a long shot.
- tbatterii 13y ago
  > If it is someone who understands most of the issues related to teaching, then no, by a long shot. I would love some more details on this.
  - saraid216 13y ago
    As I understand it, pre-college teachers generally have Actual Training for dealing with kids, like child psych.
    - tbatterii 13y ago
      I don't know if that's what rimantas meant though.

alan-crowe 13y ago

The article made me think about home schooling. Some home schooling is done for religious reasons, but other home school is motivated by academic concerns. There is not so much of the second kind because it raises an awkward question: how is the parent to do better than school at teaching a subject about which they themselves are ignorant? Enter the MOOC. The content is available, at a level and pace to suit a child. The stay-at-home parent can learn along, providing the child with both guidance and flesh-and-blood encouragement on learning and learning to learn. I anticipate that MOOC will change the balance in home schooling with a big expansion of the academically motivated aspects.

mattangriffel 13y ago

This article and others like it are arguing against a strawman. No one ever proposed removing teachers from the equation. The idea is to leverage technology to allow teachers to reach more students and/or teach more effectively. Sure, no online educational platform can rival a good teacher in a classroom setting yet, but that's because a good teacher is able to pick up on and respond to a lot more data through a physical interaction (a student's posture, tone, and other social cues) than a computer can online. That being said, in theory an online education platform should be able to track all that and more. They potentially have access to an incredible amount of data including the exact amount of time a user takes to engage with a lesson, how fast he or she is typing / is able to complete tasks, even the varying levels of difficulty at which different concepts are being learning, etc. This is all stuff that none of the current online education platforms have even tried to do anything with yet. But to say that an online education platform will never be as good as a teacher in a classroom is like going back to 60s and saying that computer image recognition sucks so we might as well stop trying.

ivankirigin 13y ago

Founder of company that specializes in adding humans to online educations says that you need humans in online education. Self promotion doesn't need to attack other good causes. There is a bit in the article about the failing brick & mortar institutions, but that should really be the entire focus. That is the only real story here. Online efforts are scrambling as offline continues to be a clusterfuck.

aaron695 13y ago

Point out current problems with X. Somehow proves X is never possible. A real hacker would think, ok robots don't work yet, how can we make them work.

jholman 13y ago

I think this article is all over the map, and I hope akharris is a better entrepreneur than writer. Though he makes one or two good points. The title is total bullshit. "Linkbait" is giving it too much credit. "Robots" is a complete head fake, and the article literally says nothing at all about whether or not education can be totally digitally mediated. The completion-rate issue. First of all, I think we can all agree that completion rate is at least somewhat interesting, and the link to Katy Jordan's piece is useful if you haven't seen it yet. And while I agree it's not a HUGE deal, for the reasons many others have gone over, akharris makes a good point, as follows: The MOOCs opened themselves up to this criticism, because over and over again they post their vanity metrics of enrollments, and compare them to classroom enrollment, and that's crazy. It's like comparing Facebook signups to private plane purchases. If they had been bragging about completion numbers (not ratios, obviously) all along, they'd be fine. Anyway, my point here is that the focus on the completion rates isn't the worst part of this article, it's pretty near the best part. The next chunk talks about K-12. Obviously most of the MOOCs were never aimed at K-12, so it's an evasion. If you look at the forum conversations or the StackExchange sites or whatever, there are obviously lots of students who are adults from eastern Europe, or Asia (and lots who appear to be middle-class Americans, of course). That's access to education, baby. The rhetorical deceit here actually enrages me. I might make a perfectly-parallel criticism of Tutorspree: "[Harris's] company states that its mission is connect students with great private tutors. So who is getting those tutors? Certainly not Tamil-speaking children in Sri Lanka." What a stupid argument I just made. Just like akharris's argument, that because 7-year-olds aren't benefiting from Sebastian Thrun's robotics courses, he's therefore not living up to Udacity's lofty mission statement. And of course, Khan Academy IS improving access to education. Not in some ideal future, but in the present and in the past. Without one-to-one tutoring, which by the way is obviously beyond the budget of, you know, most families (average price for tutors in my Zip: $88/h). So where's your access now? Admittedly Khan requires an internet connection, but $50/month for internet that the whole family can share, plus a $250 computer, sounds better than $88/h to me. Though it is nice that akharris isn't duplicitous enough to grossly misrepresent Sal Khan's vision for an ideal classroom. Nice work there. The weird part is that the thesis that he's trying to weave, about the value of human contact and personalized teaching, is kind of a no-brainer. And although I think he overstates the awesomeness of his business, I'm willing to assume it's more part-of-the-solution that part-of-the-problem. So why did he have to stoop to all these dodgy half-truths and misdirections? Maybe he needs to hire a composition tutor.