What Is The Net Native Model?

My partner Brad is fond of reminding us at USV that taking the offline model for something and porting it to the web is not often the best way to build a business online.

John Markoff's piece in the New York Times on online education got me thinking about that this morning. MOOCs are all the rage in the online ed world these days. And most of the MOOCs I have used remind me a lot of the traditional classroom model of teaching. The question I am noodling is if there is a better way to teach when you have tens of thousands of people wanting to learn something that you can teach them.

John contrasts the MOOC model to our portfolio company DuoLingo in his piece. He says:

there are early indications that the high interactivity and personalized feedback of online education might ultimately offer a learning structure that can’t be matched by the traditional classroom.

Although DuoLingo was built by one of the most popular teachers at Carnegie Mellon, there are no teachers in their learning model. It's all software, content, and users. Now maybe language learning is easier to teach this way than other things. I don't have a fully formed opinion on this. I am just thinking outloud.

But what we have seen over and over again is that taking a model that was optimized for the analog world and porting it to the internet is almost always suboptimal. And that the person or team that finds the optimal model for the internet is almost always the one who ends up with the big win.

And I think that will be true in education as well.

#hacking education

Comments (Archived):

  1. Matt A. Myers

    I’m not sure everything can be / can reach the same level of effectiveness as an online / virtual version as compared to real-life. There are many visual cues that people use when learning. Video can capture some of these, though not them all – real-time video interaction would allow for customization per individual, though then still on video.It really depends what you’re trying to teach someone, and what level of fluency you’re hoping they gain.

  2. takingpitches

    I think a lot of the difference between an application like Duolingo versus MOOCs, where MOOCs are just college courses brought online is in this quote from the researchers:β€œThe main factor for higher effectiveness was the motivation of the participants with people studying for travel gaining the most and people studying for personal interest gaining the least.”College courses are designed for a model where much of the audience feels that it is something they have to do while folks coming onto the net to learn have motivation either from need or personal interest.For example, one of the techniques of college education that is done poorly is the section that accompanies a lecture courses where students are supposed to learn from each other through discussion. 9 of 10 of sections in my education were a throwaway.Contrast that with a peer-to-peer model such as the discussions on AVC where motivated folks come here without the pressure of a penalty of a grade…

    1. Matt A. Myers

      This is a problem with all education, or anything in life really. If people don’t feel driven, aren’t incentivized strongly enough / properly, then they won’t act or at minimum won’t take in information as easily. There will be more friction and thus frustration with the process; Less fun. And everyone’s different of course as to the type of reward they need, when they need it, etc..

  3. Luke Chamberlin

    For those who also had to look it up, MOOCs = massive open online courses.

    1. jason wright

      those who have to look it up may be in a better place to find the model πŸ™‚

  4. Jan Schultink

    What online education needs is a verified testing/grading platform, then you can let the market take care of bringing the best courses to the surface.At the moment, people are starting with the classes themselves

    1. Matt A. Myers

      Define best course, define minimum understanding required for value or success, define success. A testing / grading platform that works for everyone of all capabilities isn’t an even problem to solve.

    2. fredwilson

      i agree with that

    3. William Mougayar

      Proper curation is lacking perhaps?

    4. Luke Chamberlin

      This gets back to what we mean by “education”. Are we talking about a degree as a signifier to society or are we talking about actually learning something?Depending on your objective you have a very different solution.

      1. kidmercury




        1. Jeffrey Hartmann

          This is so powerful a concept. Wouldn’t a statistically verifiable model of the students performance, strengths and weaknesses be much more interesting then a piece of paper and a GPA? Definitely alot more useful for a business if you can look at what skills need additional work and where they can immediately add value to your business.

      3. BillSeitz

        I’d argue that current mainstream education, driven by StandardizedTest-s, fails at both. http://webseitz.fluxent.com

    5. Guest

      layers and surfaces! sounds tasty

    6. JLM

      .Not perfectly focused on your comment but tangentially relevant.The issue of testing is critical to the effectiveness of anything.I learned to fly an airplane later in life. Bucket list kind of thing. Now I have a few thousand hours and would consider myself a pretty good pilot.When you learn to fly, you have to take:1. Classroom training from an instructor who gives homework as you work your way through a 2 1/2 inch book. [First homework assignment I received, I wondered who would actually do it — my assistant or my CFO. Quite a surprise that I had to do it.]2. Supplemental training in which the instructor adds his personal touch to your training. Classroom, whiteboard, hand held props.3. You have to take interim academic tests often given by other instructors.4. You have to learn to fly the airplane physically applying what you have learned in the classroom — aerodynamics, ATC system, flight controls, maneuvers, stalls, engine out, takeoff, landings — on a sometime uncooperative piece of steel and in unfavorable winds.5. You have to take interim practical tests often given by other instructors.6. When you are deemed ready, you have to fly the plane without the instructor in the plane — solo.7. You also have to take a licensing examination which is academic and then a flight test from a licensed designated examiner. These are FAA mandated exams.Much of what is taught is now supplemented by computer learning — both academic and simulator time — and multimedia presentation. In fact, everything is mirrored these days.The delivery methodology is much richer and the best, most complete instructors are also the most computer literate and best communicators.Still, there is that salty old Navy Captain with thousands of carrier landings who when you inevitably decide you simply cannot land a plane in a stiff crosswind who knocks a bit of sense into your head — “God damn it, you are in control of the plane, the plane is not in control of you. The plane is stupid. You are smarter than the plane. Fly the God damn plane.”A learning style which given my background is absolutely the best in the world for me. I loved it.The testing is critical. I don’t care what the delivery methodology is but the test is the test and it does not lie to you. You want to take that test, if you are honest with yourself.If you are seeking mastery — not just trying to get through — you want to be tested..

      1. Jan Schultink


  5. Matt A. Myers

    I feel like adding too that our educational structure overall, even worldwide in developed nations, is mostly backwards. It doesn’t match inherently how people learn. If you take this into account when comparing MOOCs and other online learning tools then the absolute/actual advantage difference wouldn’t be as high. They certainly are needed and useful though. What the exact best model is is still unknown / unseen.As part of Neomae I’ll eventually be writing up and publishing my thoughts and creating a framework relating to education to follow, which hopefully I’ll have a chance someday to see tested, and show its success to a great enough degree that its more widespread adoption simply makes sense.

  6. Elia Freedman

    I can concur. I spent two years working on a web-based education product, getting excellent positive feedback from educators, only to realize when I shipped that the entire thing didn’t work. There were computer issues, focus issues, everything I didn’t anticipate. I understood all of it from a one-to-one classroom model but moving to the web meant a lot of things were very different. I learned a lot from the experiment but what a waste! This lesson actually started me down the path of lean startups. Interestingly, while it didn’t have any market impact, the product started me down the path of a great education.

  7. Elia Freedman

    I think I’ve mentioned this here before, but the founder of Palm Computing, Jeff Hawkins, used to preach “don’t shrink, rethink” to us Palm developers. He was right for mobile. He was also right for the web.

  8. Aaron Klein

    The model is suboptimal in the offline world. Somehow we’ve decided that putting a six figure faculty member on the assembly line of rote lecture to 35 students at a time is the best use of resources. It’s driven by union rules and sclerotic thinking and I’m bound and determined to pilot several new models at the community college whose board I sit on.And just to underscore the lack of competition the incumbents will give you: I actually had staffers try to convince me that delivering online education is more expensive than bricks and mortar. “The servers and bandwidth are not cheap.” I was almost speechless at that one.

    1. fredwilson

      they aren’t cheap at scale, but then you have scale πŸ™‚

      1. Aaron Klein

        Yes, but to claim it’s more expensive than building millions of dollars in buildings, parking and infrastructure…I just had to laugh.

    2. Luke Chamberlin

      That’s when you say, “okay, show me the math”.

      1. Aaron Klein

        None has been forthcoming… πŸ˜‰

    3. Matt A. Myers

      Higher education faculty members aren’t only paid for the assembly line of students: There’s research they do which could lead to IP, they publish articles / write books and market the university, etc..I agree time could be better managed.

      1. Aaron Klein

        No research in community colleges.I think faculty play a critical role but not doing repetitive uncreative things.

        1. Matt A. Myers


        2. tsts

          True. But no six-figure salaries either, for the most part.

          1. Aaron Klein

            Untrue. They are commonplace in California.We had one faculty member whose W-2 form had a higher number on it than the college president one year…and the President made $200K+.

    4. Anne Libby

      I take little issue with professor’s salaries. My friends and mentors who are professors offer great value to their students and the community at large.What I mind is the focus on facilities. The growth in facilities comes along with a bloated staff size. Every add to staff is an increase in the institution’s fixed cost. I was taken aback recently to see one of the schools I attended tweeting about their Pilates classes. (Next up, Soul Cycle comes to campus?)

      1. Aaron Klein

        Very true.To be clear, I’m not taking issue with the salaries. I take issue with treating well-paid smart people like assembly line workers.

    5. Dave W Baldwin

      Hang in there Aaron.

      1. Aaron Klein

        I enjoy serving there, despite the challenges. πŸ™‚

    6. Donna Brewington White

      Academia can be insulated and illusional. I say this as someone with a certain fondness for the academy. The academy cannot be left to its own devices. Marriage with business is a very valuable thing and I love the movement toward more business types engaging in the management and leadership of schools.Glad you are there Aaron.

      1. Aaron Klein

        True. This could be a reason why our state’s community colleges have boards of trustees that must come from the community! It’s a privilege to serve there, even if the outside perspective isn’t always beloved. πŸ˜‰

    7. sdefor01

      Yes, and why not save that expensive resource (the physical professor or teacher) for the action that provides the most value; the one-on-one interaction with students at the local level.At home, instead of struggling with homework, would be the time for lectures. Those lectures, to communicate the foundations, can be delivered and archived from the very best profs (including great lessons from those deceased).Questions and followups from the lecture can be answered and rated, aggregated over time, and accessed by the student if there are questions.Since there are many different learning types, a lecturer (for a given topic) can be selected that’s most effective for the specific student.

      1. Aaron Klein

        We are singing from the same hymnbook.

  9. William Mougayar

    I think we are still at the experimentation stages with online education. The issue is that it copied the old “distance learning” model, i.e. removing distance and replacing it with the Internet. The “flipped classroom” model sounds good as a concept, but there’s got to be more to it.

    1. Dave W Baldwin

      You’re right William. I’m struggling with this currently. The knowledge of the Board/Administration reaches out for simple solution. The danger in flipping is this mandates what is on either side of the beam. IMHO, you cannot do all lectures at home, assignment work in classroom and what happens on whatever day needs to be decided by the teacher and that instructor needs the flexibility.

    2. Richard

      a 50,000 year old experiment

  10. kidmercury

    I don’t see how analog models are usually suboptimal for the digital world. I think they are usually the same. Do you have any examples to support our elaborate on that viewpoint?MOOCs are the budget option. Premium will involve experts. Just like how small clssses and tutoring is premium in the real world.Just to reiterate my most frequent gripe with online stuff these days, I dont think the udemy/khan academy approach is preferred because there is no focus on niche.they try to bill themselves as a universe learning resource.in the real world, colleges are often associated with a certain expertise; MIT is known for education and NYU is known for film. Companies trying to create an all encompassing learning experience will struggle agsinst google, In my opjnion.

    1. Dave W Baldwin

      Tsk, tsk my friend. You cannot take a batch of tutorials and compare to MIT, NYU and so on. In order to achieve the best output at MIT and NYU, you need a broader group that communicate in a language understood by that broader group.The bigger danger is the larger batch of models governed away from where they take place is truly screwing with what we can accomplish for the HS classes of ’19-’24 and beyond. I’m referring to broad mandates thrown down on the nation’s teachers that are made worse by the truly ignorant middle people on the staff level. As the cognitive ability lessens across the population, then worse madates come down leading to accelerating returns moving the bar lower.

      1. kidmercury

        i graduated from NYU and have spent a lot of time creating and consuming batch online tutorials. the former cost me 80K+ the latter cost me near zero (if we factor in the cost of a computer and high speed, a few thousand dollars i suppose). the latter has proven to be far more effective in education. in terms of fitting in society and convincing people that i’m of sufficient social class that they should hire me, the 80K piece of paper has proven to be a worthwhile investment. but that is more a function of a society that discriminates on social class rather than on education and capability.

        1. Richard

          yep. Education more a verb than a noun.

        2. Dave W Baldwin

          Good point. Sorry you have to put up with lack of intelligence and common sense of others.

    2. Tony Karrer

      “Premium will involve experts.” – I think the bigger issue is that online learning is very scalable so that high quality content does not need to be premium. Right now it does because of limits of scale.

      1. kidmercury

        i meant that interaction with experts will be premium. i.e. content is free or low cost, but personalized responses and tutoring is premium.

      2. Richard

        yes, and best in class professors

    3. Richard

      yep, there is only one way to learn calculus 1, practice.



    5. fredwilson

      yes. at first they took job classifieds online and replicated the offline model. then entrepreneurs created search engines like indeed.com and produced a native model

      1. kidmercury

        i guess i don’t see how that is really any different…..is craigslist a net native model?

  11. Tom Labus

    Some schools districts in Ca use the Khan Academy in the classroom. I don’t know if this is to supplement classes or control costs but leaving education/mentoring without a human element is kinda weird.

    1. Dave W Baldwin

      It depends. In our case, we have Odyssey. As an instructor, you can walk around the room giving positive reinforcement (encouraging logic and confidence) to students. In our Alternative school, you can do this with multiple levels, ranging from PreAlgebra thru Trig and Calculus. From that, you’re able to see pattern happening in the student’s cog ability and help break obstacles that can form in the thought pattern.

    2. Tony Karrer

      I personally really like the Kahn Academy and my kids use it to supplement what happens in the classroom.That said, it is more an example of moving traditional online.

  12. Luke Chamberlin

    Massive open online courses (MOOCs) = passiveNet native model for learning = activeI think that’s the key difference right there.Duolingo is active. Codecademy is active. Wattpad is active (I think Wattpad is an education startup). Heck, I think running a Kickstarter project is probably a better education than most business school classes.I agree completely that massive open online courses are a relic of an old model of education ported to the web. Giant lectures barely work even in the offline world. The net model has to be personalized and engaging. Watching a two-hour video with the occasional chance to comment is neither of those things.To be honest, I am incredibly suspicious of most “education” startups. I think that word means a lot of different things and that many people in the education space are actually at odds with each other. Are we talking about adult education? Are we trying to figure out the best way to integrate technology into primary school classrooms? Some different meanings of education:education = accreditation(your HS diploma, your BA, i.e. you put this on your resume and it’s a symbol of social status as much as a symbol of what you have learned.)education = certification(you “officially” know how to use Microsoft Office or how to repair and HVAC system. you might be able to convince an employer this is work something, but maybe not.)education = general learning(why is it colder in the winter? what is photosynthesis? how do you solve for x in a two-variable equation? when was the signing of the Magna Carta?)education = skill-based learning(how to build a website. how to speak a language. how to repair an electrical circuit.)education = continuing education(you already work in the hospital, but you want to learn how to use the new machines to protect your job for the future.)These are all different goals. How can you hope to solve a probably when you haven’t agreed upon the definition of the problem?Clearly the net is better suited to solve some of these problems than others. I don’t know what you do with accreditation. That probably requires the involvement of governments and large cultural institutions. Some of the other problems require hybrid solutions (you can’t learn carpentry entirely online).So to get back to your question Fred, I think the winners will be focused and really understand the specific education problem they are trying to solve. Either that or they won’t have the best solution but they’ll have the right politician friends to sign it into law. That happens too.

    1. andyidsinga

      MOOCs don’t *have* to be passive at all ..in fact they should be more like MUDs (if they aren’t already becoming like that). Connect those with F2F meetups solely to be social and practice an activity in the physical realm ..and tada the whole education model has been inverted.Re accreditation – this will become broader than a diploma or degree include demonstrable knowledge and practice in subject based on non-traditional learning.Jack Andraka – who devised a cancer detector or Taylor Wilson – messing with nuclear fusion in his garage (see this http://conferences.ted.com/… ) seemed to have demonstrated some level of advanced non-traditional learning and gained a level of accreditation & celebrity as a result.

      1. Luke Chamberlin

        Active vs passive is a bit of a false dichotomy. It’s probably more like a spectrum.But the massive online class models that I’ve seen tend to be very broadcast-like. The old model of we have information, and we’re going to broadcast it to you. That is a mainly passive model for the learner.Re: accreditation I hope what you are saying is true. This is easier with domain knowledge. You see this in business and tech all the time.But what about the rest of industry, where you see job postings that say “BA required” as more of a social proof than actual education or skill? What replaces that indicator?

        1. Modernist

          The Half-life of Facts: Why Everything We Know (About Education) Has an Expiration Date http://www.amazon.com/dp/15



        1. andyidsinga

          πŸ™‚ telnet FTW baby! (or ANSI BBS …)

        2. falicon

          I used to be Bruceleroy on RealmsMUD in a past online life…good (and unproductive) times…

          1. FAKE GRIMLOCK


    2. Brian Stoner

      More than just active vs passive, I think it’s about building experiences tailored for the domain vs. the lowest common denominator approaches. Duolingo is great for language learning and Codecademy is great for learning code. They each have well crafted experiences specific to their domains. If you set out to build an active learning product that taught code + foreign languages it would never be as good as each of these on their own.A lot of the MOOC’s do have active functionality, but IMO it’s not very innovative. It tends to be lowest common denominator across a broad range of domains. It feels like they’re trying to replicate the whole university before we’ve figured out the best way to teach each individual class.

      1. Luke Chamberlin

        Or… music? πŸ˜‰

        1. Brian Stoner

          perhaps πŸ˜‰

    3. AndrewsProject

      I think that models that encourage teaching as a means of learning will be the most successful online.

    4. fredwilson

      our education investments to date have either been networks/marketplaces (edmodo & skillshare) or focused vertical approaches (duolingo and codecademy).i like the way you think about this stuff

  13. Donna Murdoch

    So true Fred. Think of xMOOCs (Coursera, Udacity, edX, MRUniversity, etc.) as Web 1.0. cMOOCs (Connectivist) are more about user content generation (and ownership) so might not fall in the same category, but they are really time consuming. With xMOOCs, there is nothing new here in terms of pedagogy. They are geared toward adults, and adults learn in certain ways – so I am not sure how deeply learning is happening. These ways can be optimized – situational, experiential, contextual – and the biggest one of all, relevancy. Also, note the MOOC “drop out” rate. We skim MOOCs, as there is no cost, no reason not to move on, no grade – the list goes on. It’s great that we are coming together to learn (instead of game….not that there is anything wrong with great games) and that’s a first step. Now great minds have to go to work. But I do believe they need to work with educators. One thing that does not change is how learners learn. How we facilitate that learning is what is evolving.

    1. Donna Brewington White

      Excellent clarification, Donna. (Very cool name.) There is a difference between sharing information/content and educating and each of us learns differently based on intellectual capacity, cognitive ability, developmental stage, personality, etc. Although there are common stages of learning that a true educator will recognize and have the ability to shape the learning experience accordingly. I am a learner by nature and so almost everything is educational. Which reinforces for me something that I’ve heard so often including AVC comments — that the most important thing that can be taught is how to learn. And then we can go from there.

    2. fredwilson


  14. Rohan

    I think online education will always be a good add on. Like all things online, it will seriously hit all those teachers who get away by being mediocre.But, if I were to dig deeper, I would say – when it comes to educating people, you tend to have (broadly) 2 sets of personalities — Obsessive compulsives: Or folks who actually have the inherent temperament to stick to a course online, diligently complete courseworks..- Attention deficits: Or folks who would rather let their attention wander to “other” things. External accountability helps attention deficits study better. As a result, we tend to learn better in a real classroom.My belief is that the best learning model is the jedi model/apprenticeship model. That model doesn’t scale and doesn’t work for our large population. As a result, our large classroom model makes a lot of sense.If online education has to work, it has to either -1. Target the obsessive compulsives2. Be an excellent add on for “real” classrooms and aid learning (e.g. homework)3. Focus on ways to build accountability and atleast rein in a few of the attention deficits – use models like “Lift” to ensure daily study, for example…(And, here, I am making the assumption that the purpose of education is learning. If it’s to get a brand or find a pathway to a job, then we’re solving a different problem altogether)

    1. Donna Brewington White

      Good points Rohan. Except I’ve seen that online learning can be very effective for those with attention deficet. But that is probably most true when there is a strong incentive or motivating factor. Although to some extent this is applicable to most of us, online or off.

      1. Rohan

        Strong incentive makes sense. ties into my thesis on accountability though.. πŸ™‚

    2. Modernist

      are all students not potential obsessive compulsives?

      1. Rohan

        Would love to understand what makes you think that…

        1. Modernist

          curiosity is addictive. each student needs the right molecular cocktail of gamification, which is a drug. personalized medicine for pedagogical disengagement. zynga’s backers know this about games–addiction must be targeted. it can be applied to education. we need to unlearn pessimism (ie. only a minority of students can be outstanding) as we unbundle self-fulfilling labels. diagnose the studential problems from the student, don’t diagnose the student from the “class”.

    3. fredwilson

      i agree Rohan. your comment pretty much sums up what i have observed in the past three years in my/our deep dive in this area

  15. gregorylent

    self-discovery is missing in classroom lectures … online can have that, multi-media opens the door.

  16. Anne Libby

    (And on an unrelated note, a particularly odd set of articles is popping up on avc.com as “recommended for you” — “‘Pregnant’ Teen’s Boyfriend Allegedly Sat on Her Until She Suffocated.” Uh, ok?)

    1. Anne Libby

      (Is this what reading “crafting” blogs gets me?)

    2. fredwilson

      yes, disqus is working on their targeting algorithms and they use avc as a test bed. so we are going to see the good, the bad, and the ugly

    3. Donna Brewington White

      And I’ve got five on cruises and one on the guy who funds Lance Armstrong.So does the algorithm show that I need a vacation and someone to fund it?

  17. pointsnfigures

    we have to make it our mission to revolutionize education by any means necessary. we know that government, and entrenched organizations will try to stop us-but the future depends on us re-thinking the way we educate our population. MOOCs are a start-but as thinking evolves I bet entrepreneurs are going to figure out some incredible ways. Because of the structural constructs in the way, it’s going to be more expensive to accomplish a revolution. But, we can do it if we try. I am pretty sure we can validate demand, now it’s time to revolutionize supply.

  18. howardlindzon

    We are seeing this everyday on stocktwits…nothing works better than reading and digesting lessons about the market, but having a 24/7/365 interactive journal and feedback about ‘selling’ rules and support are the real key to investing and trading success.

    1. William Mougayar

      yup. bite sized, daily learnings.

      1. Rohan

        Ah. You mean like those on that Alearningaday.com blog?? πŸ˜‰

        1. William Mougayar

          yes, exactly!!

    2. Modernist

      i also encourage the students on my social learning platform to keep a journal of their bets-Seth Davis, Queens Casino Education Company, LLC

    3. fredwilson

      and nothing like taking some losses teaches you investing

  19. howardlindzon

    also – sorry to promote again here – but Colingo has started to figure this out for english learning by using big google hangout sessions with full interactivity. Its cool.

  20. William Mougayar

    Before applying the Learning model to the Internet, first we need to unlearn how learning was done before.Unlearning is the biggest obstacle to learning.

    1. EmilSt


    2. Modernist

      “systemic problem, integrated solution”. we need to act in parallel

      1. William Mougayar

        yes. unbundling, unlearning, disagregating…then re-assembling in a very different ways with elements not possible in the old way.

  21. falicon

    ‘Moon shots’ are what will continue to drive education…the trick is in finding the big problems/challenges that the younger generation actually buys into and cares about.Once someone is truly hooked on a problem the required education towards a solution tends to work itself out.

    1. Dave W Baldwin

      You are right! You could do some interesting slides on the Dawn sattelite leaving Vesta, attaching the interest in mining raw material and its next step forward intercepting Ceres which apparently contains an incredible amount of water.Don’t make the presentation too long, just something to be real world. You can then have students during/after give answers regarding area, volume, %volume related to element of interest, distance and so on.For that matter, it offers a transition to writing based on developing a plot that moves into that realm.

    2. BillSeitz

      I disagree, this is just another attempt to “engage” students in a passive/reactive/push model. Seriously unstructured and individualized project/product framings let students create their own curriculum base from which they want to *pull* learning. http://webseitz.fluxent.com

      1. falicon

        I was talking about unstructured framing…with a focus on big and interesting problems…without the motivation or the drive to question and to learn, it’s nearly impossible for anyone to teach…

  22. tsts

    This is an interesting discussion. But I think there is unlikely to be one model. Rather, there will be a large set of tools or templates that can cover many learning scenarios, and then there are lectures and maybe face to face for unstructured content and for motivation.Let me try to explain what I mean by unstructured, and relate it to duolingo (which I personally use) and coursera. Duolingo works very well because much of what you have to do for language learning can be boiled down to a small set of templates (structured tasks or exercises), like vocabulary learning, translation, listening. While there are other important things, this set of templates gets you most of the way. This is a main reason why duolingo works.Now compare that say to your typical senior undergrad or early grad database course. There are some templates that you can identify and implement and drill in, say writing SQL queries, relational algebra queries, normalization, and a few other tasks that usually feature prominently in homework assignments and tests. But that does not get you to 50% of what you should learn in such a course, and there is a lot of other content that is unstructured.Now think of a course in American HIstory, or many other subjects. It is all unstructured. Not many repetitive (even if intellectually challenging – like writing really complicated SQL queries or solving differential equations) tasks you can train. In this case, your default template is the lecture (in person or online) or the reading assignment. So lectures are for unstructured learning, what (early) search engines were for unstructured web content — you do not try to formally model the structure because the structure is too complicated. (That is what unstructured actually means in many cases: no simple structure – lots of structure, too much to model in detail). But eventually, you find that identifying and modeling structure (say, tables and entities in the case of search engines) improves your quality. But it is hard and progress is slow.So my point is that there is no one model. Online learning will be important, but it will be a long slog trying to figure it all out. It involves inventing and implementing new templates (models, mechanisms), as duolingo has done for languages, and then identifying these in the topics that have to be taught (basically, decomposing the topic into structured and unstructured parts), and then covering the structured parts with templates. It will take a decade or two to make much headway.

    1. fredwilson

      i agree with you

  23. EmilSt

    I think the platform itself should offer range of models that fit different users. Something like pandora for education where you can subscribe to channels on different subjects, on which you can get different algorithmicly selected content like textbooks, articles, studies, videos,tweets … All relevant and up to date. You could also chose the increments in daily, weekly, monthly delivery.Then you can have curated channels from anyone who wants to curate. Great teachers will have thousands of subscribers. They can do it for free or for fee. However, there might be different levels of interaction with the teachers/curators with a scale of fees. Every curator can set his/her fees and the platform gets percentage. Also curators could invite contributors/assistants to their channels. Eventually it can have some very high level of interaction with examination and awarding points/credits. Like this I can live in India and study Philosophy with A Harvard teacher, History with Columbia teacher and tech startups with Fred. All on different interactivity level as I find appropriate. I can have my own personal virtual university for low cost and maybe I can even get credits and some edu-pandora diploma if I ever need to prove to someone what I have studied (I believe more in proving what someone can do). With the social layer I can see what my friends and peers are studying and also what the great achivers are and where studying

    1. Donna Brewington White

      You are on to something here.

    2. Jack Sarvary

      Love the idea of subscribing to a professor.

      1. Modernist

        How about a competitive game you play with professors?

        1. EmilSt

          Or with peers…

          1. Modernist

            false dichotomy, in my humble anti-credentialist opinion!

          2. EmilSt

            I agree. We are all peers teachers/students.

    3. fredwilson

      that’s not far from what Udacity is trying to do

      1. EmilSt

        I feel they are still tied to “class” model. Fixed knowledge in fixed time. I think the learning curve could be more individualized. Flexible knowledge in flexible time.

  24. AVCoholic

    I think the issue here is that we’re taking an offline model that many people seem to agree is very flawed and ported it to the web. It’s not the most optimal solution in the offline world so it’s not a surprise at all that it isn’t the best way in the online world. Just my humble opinion though so feel free to disagree.

  25. Tony Karrer

    Great post Fred. Thanks for shining light on this.It seems like a lot of the comments miss what you are really saying here. Online learning has an opportunity to be radically different than traditional education which relies on sage on the stage and testing. Instead, online you can make learning adaptive. This is a hard problem to solve, especially generally, but its been something that’s been shown to work extremely well.The problem we have is that most of what people are looking at right now resemble traditional learning with MOOC and Khan Academy. I think both are great, but you are exactly right that this misses the real opportunity here.

    1. Jeffrey Hartmann

      Tony,I hate to respond with a ‘me too here’, but I couldn’t resist. Tailoring education is where education software can really shine. A human teacher can not keep every response of every homework question in mind when they make a decision on what to teach, but this sort of statistical approach is child’s play for the computer. Traditional education beyond one on one tutoring also doesn’t have the ability to really focus on what each individual needs at the time. Neuroscience teaches us that the most valuable time in the student/teacher relationship is the one on one time a teacher spends with a student. Imagine if we can amplify their one on one time by providing the teacher information on what the student needs help on. The one on one time can then be spent helping the student rather then trying to figure out their problems. Imagine if we can emulate this one on one time as well, and provide the student constant feedback on their progress as they practice? The possibilities here are endless to really make a meaningful impact.

  26. Modernist

    i think edmodo gets overlooked because it resembles facebook superficially, but will prove a tremendously influential platform. if you feel similarly, please consider sending a tip to MIT Review, because as my favorite story source it bugs me they have not covered edmodo yethttp://www.technologyreview…

    1. fredwilson

      i agree with you. but you already knew that

      1. Modernist

        we love you, but you already knew that. “family” and “community” are synonyms.global family is a scarily powerful idea, but we (the world) could learn from chicagoans. familiarity is the union explanation for: 1) chicago known for nepotism, *and* 2) statistically asymmetric (nonrandom) emergence of the great social forces (Darrow, King, Obama, Pincus, Edmodo etc) from the windy cityplease everyone, let’s spam MIT Review

  27. panterosa,

    My team is watching the Udacity lectures for an incubator at Columbia and Steve Blank came to talk last week. How to successfully engage kids/students is of particular interest to me. I am building digital learning via gaming and looking to engage kids in groups of their friends or classmates beyond the game learning moment.As a team watching Udacity, we find the benefit is our discussion after, as well as interaction with faculty on our conclusions. I went to a school which used round tables, Harkness tables, which work well for participation, and I’d like to make the round table possible for my product. Possibly EdModo is a tool to leverage in this venture, not sure yet, but for sure sure some kind of hangout is a great add to mulch the learning and encourage discussion.

    1. Timothy Meade

      While a lot of people talk about a “post-PC” world, it seems obvious that the first time a young child of the near future might see a full-sized keyboard is in school.

      1. panterosa,

        Please expand on this.

    2. fredwilson

      yup. udacity is great for delivering the material you need to learn. but nothing like a face to face discussion to get it to enter the brain and take residence

  28. LaVonne Reimer

    I was the founder and CEO of one of the first wave of e-learning companies. I launched Cenquest in 98, focusing on grad ed. Hummer Winblad, Sevin Rosen and others backed it. Henry Kaufman of Dr. No fame was on my advisory board and our academic partners included UT Austin for tech commercialization, Babson, and NYU/Stern in its international Exec MBA with LSE and others. We disdained the glorified correspondence courses and the talking head video. When Berners-Lee’s Weaving the Web came out, the very first page had words describing our philosophy of education beautifully “the decentralized, organic growth of ideas, . . .”The practical challenges we faced were professors deciding on a syllabus at the last possible minute. Also each of them believing they had the most unique marketing or strategic alliances curriculum out there. Of course they shamelessly cribbed each others content. So we found a way to re-use course “objects” that flattered professors and stayed within the bounds of copyright law. We built the course dev system to deliver on the vision while managing for the practical problems.Our best moments were hearing from working adult students that they had been able to instantly apply new learning to solve a stubborn work problem–the collapse of the ivy-covered brick wall campus and merger of learning with solving real-world problems.That first wave collapsed. How and why it did is not unlike many other similar tales. I’m sure my journey of processing through all that and now the rise of the next wave is pretty common too. For my part I am moving on to what is in some ways the next rev of Cenquest. On the surface we have a trust platform for commercial credit and the utility aspect is being able to easily share segments of financial information with diverse audiences. But the early user feedback that brings me total joy is how educational the experience is, from how to prove credit and trust to understanding the numbers and using that understanding to build a more sustainable company.I apologize for the length of this comment. I am a believer in the transformation of education through technology, and networks. I am hopeful for this new wave.

    1. Donna Brewington White

      If anyone has the chops to write a longer comment on this it’s someone like you. The type of journey you are experiencing is very informative. Some words you used leapt out at me when you said what a total joy it is to hear that the experience is so educational. I am playing on words to an extent but online education has the potential to be highly experiential even if that experience is simulated. Many of us learn better with an experiential element or the ability to readily apply what we are learning.

      1. LaVonne Reimer

        I always learn from your comments! Really appreciate the nice note. The area I am exploring is how to create immersive experiences through web applications. In a sense we “trick” users into learning by how we guide them through using our applications. An open question is how we give users “credit” for that.

    2. JLM

      .Not only is it not too long, you have us on half rations. More, please.The voice of reality and experience drowns out the chorus every time.Well played..

      1. LaVonne Reimer

        Thanks! For a long time I swore I would never talk about e-learning again. That was my valley-of-the-shadow period I guess. I am still puzzling over the meaning of this wave so not sure yet that I have any real wisdom to impart.

        1. Donna Brewington White

          You got a well played from JLM. That’s precious currency. And thanks for your kind words below.

          1. LaVonne Reimer

            Yes, I am a JLM fan too.

        2. fredwilson

          so we did something good. we got you talking about something you know a lot about. failing = learning

  29. Sebastian Wain

    I think it is easier to argue about online education for computer science or math than on educating doctors. Learning to develop software can be gamified with platforms such as topcoder but starting with more basic exercises. The theorical aspect of math and CS needs a different approach, automated proof tools came to the rescue. But for me the most interesting part is like an OkCupid but for education: building good distributed teams that works well and with a clear purpose.

    1. Richard

      Secret is 2.5/4 years of medical school “is online”. This is why 4-6 additional years are as a resident/fellow.

  30. JLM

    .At the end of the day, most everything is out there somewhere, it is the delivery mechanism that makes it palatable or, worse, prevents us from getting it into our heads. That gives it the flavor that keeps learners coming back and teachers teaching.I suspect the first thing to do is to unlearn most of what is known about learning and teaching.The UNC MBA on line program uses Adobe Connect to deliver what is a great mix of face to face (but distance learning), lectures, “pre-noted” lectures, PPT lectures, discussion, Q & A, office hours, “canned” (previous lectures video taped) information, podcasts and group activities.Adobe Connect makes all this possible in one piece of software and there is a professor to do the teaching.If you cannot learn a subject with those varied delivery channels then you have a learning issue.I routinely study podcasts from MIT, Stanford and other places but those are my two favorites just as much for the bibliographies as anything else.But I am looking for “mastery” rather than educational achievement. I will know when I know the subject and continue to be curious about, as an example WWII, though I actually know the outcome.I realized I knew next to nothing about the Federal Reserve and got a couple of podcasts with bibliographies and am now through my 10th book — of course, the books don’t all agree requiring me to think my way around the bases.Egads, I actually have to think critically and with a bit of skepticism. I actually might be getting an education along the way.Also, I do not have to pay tuition.JLM.

    1. Jeffrey Hartmann

      I think you hit on a key point here. Our current system doesn’t seem to encourage or reward students who want to master a subject. For example, my son reads well above his level, but he still has to read simple Dick and Jane like paragraphs for some state sanctioned testing regime on a weekly basis. His teacher knows it is well beyond him, but ‘them’s the rules’. Wouldn’t it be so much more powerful if we recognized his mastery and pushed him to really excel? I think this is where the solution to many of our education problems is hiding.

    2. fredwilson

      lifelong learner!!!

  31. Dai Ellis

    Agree with others here that MOOCs (at least in their current v1 incarnation) don’t represent the ideal net-native model. I’ve been working in the no-excuses charter school world and one of the reasons for the slow uptake of blended learning in K12 is that the state of the software is still pretty bleak. I’m now involved in developing a new model for higher ed (http://wp.me/p2bfBq-2c and bit.ly/T9GTB1); we want to curate the best online content/tech out there rather than creating our own but there’s not nearly as much ready-for-primetime content as we need.MOOCs as they exist now are basically like online textbooks with built-in problem sets — just another form of asynchronous online content. Keys to net-native will be interactivity (whether through synchronicity or innovative asynchronous approaches); active learning; adaptivity; and ‘flow’. John Danner is one of the smartest K12 observers because his school network is on the leading edge of blended learning, and his recent post on active learning is a good read: bit.ly/YhoM03. Likewise University Ventures’ super-sharp year-in-review letter (bit.ly/YhoTsy) has a great section on ‘flow’ starting at the bottom of p. 5.Knewton, Dreambox and a few others are on the cutting edge but no-one has cracked it across a wide range of levels and subjects.

  32. Charles Butler

    I think in this case, the lack of success (if you can consider a 20% completion rate unsuccessful considering the situation) is due to neglecting some of the most important parts of the offline model, namely investment and accountability.If you shell out lots of money for tuition and stop showing up you fail. Also, the other people in your life are going to know it, since most people attending (offline) college courses aren’t doing so in secret. Maybe MOOCs aren’t secret, but I also imagine that no one you know cares too much about whether you’re taking one (or dropping out of one).Adding the same level of investment and accountability to the online courses would dramatically increase the completion rate, but of course, it would also dramatically reduce the number of participants. I’m going to speculate that the MOOC model results in far more total education, and that sounds like success to me.

    1. fredwilson

      i would like to see how we can remove or at least drastically reduce the financial cost. so the use of investment as a driver is something i would like to lose.but you are right that these things drive students to perform.maybe we can start with accountability



    1. Richard

      No matter the shoe, education is a mountain, with no summit, where you trip and fall often.

  34. joeagliozzo

    I like DuoLingo a lot. I think there IS a very big difference between something like language instruction – which is mostly a lot of repitition – and say, programming in python, or stats, or machine learning – which requires a lot of explanation and guidance. The explanatory video format is why I had more luck with udacity programming courses than codecademy. I felt like codecademy was just a lot of repitition without explanation of what you were doing, so then when you tried to do something on your own, you really didn’t know where to start.

    1. fredwilson

      yes. i agree. different things require different approaches.

  35. alexosterwalder

    when the film camera was invented people filmed the theater… that’s where we are in online education…

    1. fredwilson


  36. Sam Birmingham

    Two key words: Attention and MotivationWhen the main cost is my attention (ie. time) it’s easy to register, passively wade in, then drift away whenever — think the thousand-and-a-one newsletters we variously subscribe for and lose interest in; traditional broadcast media’s inability to transition to digital; first-gen MOOCs…When my motivations are personal and specific (eg. DuoLingo and its higher effectiveness amongst those who are learning a language for travel rather than personal interest) then I’m inherently more likely to stay engaged.

    1. fredwilson

      gerat point

  37. jason wright

    universities are walled gardens.

  38. Jeffrey Hartmann

    As someone with an education startup I can not but whole-heartily agree that the MOOC model is something that is not optimal for learning. We have a vision of something that is quite different in the end, software that drastically improves a teachers ability to teach not something that just magnifies their reach. Right now there is so much that a teacher does that doesn’t have a meaningful impact. They also have an obscene amount of busy work. If everything has built in measurement and is adaptive to each student without the teacher having to be intimately involved, if we can provide targeted content to address a student’s weakness, if we can provide a system that gives the student a place to practice deliberately their skills we have something that is much more powerful then simply letting a lecturer teach to thousands of students instead of at most hundreds.There is so much opportunity right now in the education market, generally the software that is available really tries to push a very outdated model. I think focusing on making tools that attack failures of the current system like:1. How can we get students to collaborate but not copy off one another?2. How can we make the measuring and grading of students all encompassing, friction-less and effortless on the part of the teacher?3. How can we use available data from this measuring system to adapt our approach to teach?4. How can we engage and empower parents and peers to know how to best help the student that is struggling?5. How best do we identify promising innate skill and properly foster it?In my opinion attacking these problems are what will really move the needle towards a net native model for education. Our startup is attacking fundamental math education first, keeping in mind that solving these problems will give us something very valuable for teachers, students and parents. I’m very excited to see how our approach works over time.

  39. Emily Merkle

    The role of the teacher in the future of education needs to change. A shift from Imparter of Information to Experience Guide. Teacher as Evaluator Of Progress, as Diagnostician of issues, as Child Advocate.

  40. Dasher

    For me going to school (I have a Ph D so I went to schools for a long period of time) was an experience I would not trade for anything. There were many teachers that inspired me, mentored me, cajoled me, punished me, pushed me, entertained me, and helped me be the person I am today – next to my parents and friends they were the most influential people that shaped me. Online educational models can provide content but can not replace what I got from these teachers. So I am not convinced that they can be a complete solution to providing ‘education’ my kids.

  41. invisibleman29

    that’s funny. I’ve taken 10 or so courses on Coursera ‘just for grins’, and being attracted to its ‘focus in ten minute slices’ flexibility — I came to the exact opposite conclusion. Learning was most efficient when you completed a lecture beginning to end, and took the quiz right away (i.e. very classroomy)

  42. Advancedwriters.com

    That is a very interesting thing. I like the thing that it’s good in business as well as in education too. A simple thing yet helpful to those students for their studies.

  43. fredwilson

    stack is certainly a great resource. i don’t think of it as an education oriented investment but maybe it is