Yesterday, I took my last final exam and finished 5 years of study at MIT. I learned a lot of amazing things here, and I gained a lot of skills – some more obvious than others. When I walked out of the gym after my final, all I felt was exhausted (hell, I had been computing matrices by hand for 2.5 hours…), but a strange thing happened when I woke up this morning. I felt a sensation that I wasn’t expecting – lightness, almost as if a great weight had been lifted off my shoulders and I was free to really begin my education on my own terms. But why did I feel this way when MIT was over, and not every day over the last 5 years?
I think the answer to this question lies not in a scrutiny of MIT, but in an exploration of what we look for in an education and what we actually receive. I had a great conversation with two of my thesis advisors, M.S. Vijay Kumar and Brandon Muramatsu, regarding the positive and negative aspects of my experience at MIT and what a future version of the school could be. M.S. Vijay, or Vijay Uncle as I refer to him, gave me one last homework assignment: come up with 5 features of your ideal version of MIT, the version of the institute you’d like to have attended.
Here’s what I came up with:
1. Grades Are Flexible
In the MIT I’d like to have attended, my grade in a class is not a mark that is given to me on a single day at a single point in time, but rather a constant reminder of all the work I have put into mastering a subject. My grade starts as a zero, an F, what now indicates to students at the end of a semester that they are a failure. In the new MIT, this just means there’s more that lies ahead on my journey of learning. As I complete assignments, my course grade increases. If I complete a lab report and it receives a grade of B or C, that grade is given with qualitative criticisms, and by improving that lab report, I can elevate my grade to an A. After a certain point, my grade is high enough that I have ‘passed’ the class, but when I reach this point, it’s at a C minus. If I choose to move on to the next subject, it’s my choice, but that choice carries the consequence of having a C minus on my transcript – but in the MIT I’d like to have attended, a C minus means I stopped trying to master the course, not that I was “bad” at that course.
2. Grades Open Doors
In the MIT I’d like to have attended, navigating the curriculum is like playing a game of Super Mario Bros. Salman Khan uses this strategy with the Khan Academy, where students complete practice problems grouped by concept, and after finishing 10 problems correctly in a row, a student is declared to have “mastered” that concept, and new, more advanced concepts are only then made available. Colleges now only do this at a low-resolution level, where courses have prerequisite courses. But what if concepts had prerequisite concepts? This might just give students the reality check they need to not rush through a course and get frustrated with the many holes in their knowledge of the material.
3. Teachers have a Pulse on the Classroom
In the MIT I’d like to have attended, lectures are high-bandwidth interactions between students and teachers or between students, and digital media is used to make a single teacher more aware of the state of learning of all his or her students and also to empower that teacher to react to this information.
In today’s MIT, much of the time that we spend in class is in the form of a lecture, that, until smaller graduate-level classes, consists in large part of a single speaker simply lecturing to an audience. Is there something wrong with this picture? I say there is. I look to my left, and a student is passed out. I look to my right, and I see a kid jotting notes furiously, but if I asked him what was going on, he’d just stare at me with a blank look on his face. His students are dropping like flies, but the professor keeps on talking. But this is not his fault – it’s the lack of technology that could help facilitate the rather difficult task of communication between a single teacher and many students in a single room. What if each student had a “I’m lost” button that they could press at any time? Now, the same professor could stop when he sees that 80% of his class is lost. Sure, he isn’t sticking to his timeline, but with 100 students, by not finishing the last 20 minutes of lecture and going back to address the misconceptions, he is saving 20 * 80 = 1,600 minutes of class time. Better feedback leads to better teaching.
Similarly, what if students could ask questions not just by raising a hand but by subtly posting in a real-time forum, where questions could be up-voted or down-voted just like on StackOverflow.com, and what if after a certain number of up-votes (from students saying “hey, I’d like to know the answer to that, too!”), the teacher is notified in real-time that many students share a single question? Would this disrupt the flow of the lecture? Hell yes. Would it lead to lectures that are somewhat more akin to a real conversation? I believe it would. People like the folks at Piazzza are already working on solving this problem as I write this.
4. Educational Content is Served Deli Style
In the MIT I’d like to have attended, students no longer use traditional textbooks. Instead, they are given a choice of a variety of educational media that is heavily tagged with metadata such as the specific concept being covered, the way in which the concept is explained (analogy, real-world example, application of theory, etc.), the type of media (document, video, interactive game, simulation, etc.), and a wealth of data regarding how students across both an individual class and all learners in the institute engage with each given resource.
Students are not assigned to read a chapter of a book each week; rather, they are told to expose themselves to and try to learn a new concept each day, and each student has a choice of which resources to use when learning this concept, along with information about which resources other students are using more or less. Students can even contribute resources to the course library that they find while browsing the web. I had the fortune of building a primitive version of such a system for my Masters thesis at MIT, and I think that harnessing the vast amount of educational content that exists in corners of the internet and surrounding that content with relevant and useful metadata is key to improving the user interface of education overall.
5. The Golden Rule: Low Student to Teacher Ratio
In the MIT I’d like to have attended, there is one rule that reigns over all others in the design of the institute and of individual classes: a low student to teacher ratio.
I had the fortune of attending a small, private high school in San Jose, CA where the benefits of a low student to teacher ratio were evident. If I ever got a B or a C on a test, my teacher reached out to me and asked me to come in after class and talk about it. This resulted in high-bandwidth, fruitful one-on-one conversations that resulted in amazing repairs to my flawed learning as evidenced by the test I scored poorly on.
Now compare this to MIT, where I sit in a lecture hall of 150 students, then attend similar lecture-formatted “recitations” with smaller groups of 30 students, turn in problem sets each week and only get feedback on those problem sets a whole week later in the form of just a number grade, which is rendered by “graders” whose job it is to read my work, compare it to a template, and assign a number. Suddenly being a student feels less like being a person and more like being a can of processed food moving along an assembly line.
Learning revolves around failing, receiving rich feedback on why we failed, and through the guidance of a teacher correcting the misconceptions that caused that failure. Until we create the ultimate AI-based tutor that solves this fundamental problem, this problem needs to be addressed by maintaining a small number of teachers to students. Why was my feedback in MIT classes so poor? Because it scaled – it scaled damn well, but at what cost? Just because an education is covering a technical subject like engineering, many people simply assume that the rich interactions of one-on-one communication between teacher and student matter less, and it is easier in these disciplines to discover and correct misconceptions. I respectfully disagree, and offer myself as data point #1.
Now people will of course say that this is an impossible problem to solve, that it requires just too many resources and too much money. But there are many ways this can be tackled. For example, a teaching requirement can be placed on students so that before they graduate, they *must* TA at least one or even five classes, not only solidifying the material of their fundamental courses but also at the same time providing an unending army of dedicated teachers who are large enough in numbers that they are capable of investing themselves in the success of their students to the extent that all teachers really strive to but are unable to because of excessively large classes.
Ok, now that I’ve gotten this list started, here is where I need your help. So far we have 5, but I want 100. Then I want Susan Hockfield, president of MIT to read this list. I have only one rule: don’t restrict yourself with what seems practical, doable in 5 years, or even in 20. Use your imagination, and dream of the best possible learning environment you can.
I invite all of you to post a comment, starting as follows: “In the <University Name> I’d like to have attended, …”
This year, MIT turns 150 years old. Maybe by the time MIT turns 200, the MIT that I’d liked to have attended will be the MIT that everyone attends.
