Can You Learn Faster Without School?

I’m nearly at the halfway point of my challenge to learn MIT’s computer science curriculum in 12 months. I decided to put together a short video explaining what my daily routine looks like, and also share some of the tactics I’ve been using to learn at a faster pace.

Learning MIT at 4x the Pace

After having completed a degree at a regular university, I was surprised to find that learning faster is actually easier if you’re not in school. There are a lot of methods I’ve been able to use to speed up the process which are cumbersome in a normal academic environment.

Are these speed ups legitimate? After all, if an actual MIT student has access to everything I do, how could I possibly learn faster using fewer resources? To explain how that’s possible, let me explain how the challenge works.

From the beginning, my goal was to pick a basis of evaluation that could get close to MIT’s actual curriculum. In the end, I settled for final exams and programming assignments. It’s not perfect, but my goal was to get as close as possible to the evaluation MIT uses, while being flexible in how I actually learn the material.

You may argue that, as evaluation tools, these miss some elements of learning, but I’d say they’re a pretty good approximation. They account for the majority of an MIT student’s GPA and they cover the almost everything taught.

The advantage of a non-traditional learning method isn’t in what it includes, but what it leaves out. With a fairly quick evaluation method, I can be more creative in using assignments, lectures, recitations and problem sets to extract the majority of the value in substantially less time.

Learning Hacks

Despite being the main vehicle for scientific progress, universities are hardly a model of pedagogic excellence. The standard teaching format in most schools has changed little in the last few hundred years, despite new scientific results showing those methods are often inefficient.

Hacking Lectures

A simple example of this is the lecture format. I can extract huge time savings by not watching lectures at their normal speed. By speeding up the lecture to 1.5-2x the pace, I can save nearly 20 hours of work, with just one change. (Note: you can do this by downloading the lecture, then using VLC Player to speed up the playback rate)

Yes, this makes lectures harder to follow, but because I can rewind and rewatch, that isn’t a big problem. If you’re taking a class you’re properly prepared for, it should only be minor sections that cause confusion—why not carefully rewatch those sections and speed up the rest?

Turn Assignments into Deliberate Practice

Assignments and problem sets are another area that can be hacked. Research shows immediate feedback and targeted practice are essential elements in rapid skill building. So why do the assignments in-order and wait weeks to get your results?

I usually do assignments out of order, focusing on questions on the hardest possible topics, or on the core skills I sense I need to master most. I also check the solutions immediately after working on a problem, to improve the feedback mechanism.

Under this view, assignments are all about preparing you and serving as an outlet for deliberate practice, not just busywork that needs to be completed.

Recursive VS Iterative Learning

Most students learn in an iterative fashion—take lesson one, master it, then move onto lesson two. The academic system basically enforces this method of learning since lessons and homework are trickled out in lockstep.

There are two weaknesses with this approach. First, you don’t get to see how early concepts are going to be applied to later ones. Second, it doesn’t allow you to invest your time in the topics you find most challenging, instead you conform to the pace of the group.

I prefer a recursive strategy. First I “learn” the entire course material, which usually doesn’t mean I’ve mastered it, but I understand the basic principles. Then I recursively deepen my understanding on harder topics until I’ve mastered it. This deepening can be done by deliberate practice in problems as mentioned previously or by using intuition-generating methods like the Feynman Technique.

How Can You Apply These Methods?

It might be an interesting experiment to see if there are more efficient ways to learn, but you might wonder how it’s useful. After all, if you’re an actual student, assignments and attendance are often mandatory and rarely can be done in the out-of-order, rapid-feedback ways I’m suggesting.

But imagine spending a month to do what I’m suggesting, and master the concepts of your classes using free resources from Stanford, Harvard or MIT? The rest of the semester would be significantly easier if you’d already learned 80% of the course in the first few weeks.

Not everyone can learn a class in one week, even with the methods I’m describing. But I’m certainly not 4x as smart as an MIT student, so at least some of the advantage has to be structural, not just intellectual.

Changing the Costs of Learning

I know a lot of otherwise intelligent people (typically businesspeople) that scoff at learning from a university. It’s too academic, devoid of real-world implications or theoretical for their liking.

I suspect a lot of their distaste comes from the normally high cost in time and money it takes to learn. If you could learn twice as fast, with complete flexibility and zero cost, would that change your perspective of whether learning calculus, finance or genetics is worthwhile?

When mainframes still took up a warehouse, many intelligent people couldn’t imagine why anyone would want a personal computer. Now we have them in our phones. If education were to become free and less time consuming, what would stop you from learning?

I don’t know if free education will change the world. But it has at least changed my life. I’ve already started downloading the resources for graduate classes on quantum computing and machine learning I want to take once my challenge is over, and I’m looking forward to classes in physics, brain science and biology. Why not you too?

  • Steven

    This really isn’t four years of study in one. Although you previously mentioned that it ignores all labs and homework, etc, you also don’t take into account all the other non-engineering classes that every university including MIT requires. It’s more like 2-3 years of classes in 1

  • Vinay Bhat

    It’s incredible seeing you halfway through your challenge, and thanks for the link to Feynman Technique as well, looks extremely interesting.

    I have a question, Scott. Everyone around tells me that to master or “deepen” the understanding of anything, it is very important to use it or apply it on a day-to-day basis.

    So with this CS knowledge, what are your plans for applying the knowledge? A new start-up?

    PS: Btw, Scott, you resemble Stuart Broad a lot!

  • Nelson

    Great post. I absolutely agree. Another huge benefit of being self-taught is that typically, you enjoy what you’re learning more — usually because you choose to teach yourself, and you wouldn’t choose something that you didn’t enjoy learning. While you may not enjoy every class maximally, you chose the CompSci curriculum because you like and want to learn about CompSci. Choosing a Philosophy curriculum wouldn’t have made much sense (unless you analyzed and ruminated beforehand about why you’d want to do so).

    Also, I found one mistake: “Second, it doesn’t allow you to invest your time in the topics you find most challenging, instead you conform to the pace of the group.” This isn’t a sentence. 😉

  • Michael Sieler

    Really interesting information! Thanks for sharing the Feynman Technique!

  • Peter

    Hi Scott,
    Quick question. How much sleep do you get every night?

    I’ve seen in previous articles from over a year back you mentioned you get 8 hours, but do still maintain that now?

    I’m especially interested since you’re studying advanced concepts and using creative problem solving for algorithm problems. As someone who does some programming myself, I’ve found that my ability to solve new problems is noticeably decreased with reduced sleep. Knowing your experience with this would be very helpful.

    Thanks a lot for being so inspiring, take care,

  • Sally

    Hello Scott;

    My question is quite similar to Peter’s. With such a workload, do you still do the things you used to do–read 60 books a year, go dancing lessons, workout regularly etc…

  • Jared Romey


    Just thought you might be interested in Khan Academy. I’ve run across the name several times in books I read (most recently Abundance). I pulled this from their website: “Khan Academy is on a mission to provide a free world-class education to anyone anywhere.”

    Thanks for the post, it’s extremely well written and includes great learning information.


  • Nathan Shepperd

    Scott, I think you’ve confused iteration and recursion as concepts. I’d say a traditional syllabus of learning is sequential. What you’ve actually described is iterative – get the basic principles and then repeat areas of detail with a wider view of the whole course. I suppose recursion might apply, as the concept of “depth” is in there, so this may fit your meaning well enough.

  • IC

    I like the recursive learning trick. Perhaps if universities allowed students to take exams to pass a course before the end of term it would be beneficial. I recently wrote a blog post about some changes that might make the education system more in tune with societal needs, and your idea might provide students with a method for accelerating their university turn around time.
    The problem, as I see it, is that our society is a meritocracy, i.e. it doesn’t care what you know, only what you can prove you know through certification or degrees. That doesn’t mean learning outside of institutions is a waste of time, but it will be less helpful in providing individuals with the credentials to obtain jobs.

  • Amelia

    Hi Scott,

    I’ve been following your blog loosely for about two years now and This MIT project really interested me for a number of reasons, but I want to pose some questions for you concerning how you’re presenting this project to yourself and your readers as well as your evaluation methods.

    Summary of Points:

    1. A focus in any major takes only 2 years, even MIT requires basic math and humanities courses from all students in the first 2 years. So aren’t really only going 2x as fast?

    2. Are you leaving out the engineering aspect of the MIT degree? All computer science majors are attached with this, no? This further decreases your rate to < 2x if it is the case.

    3. If you actually do a legitimate 4 years of relevant computer science work in 1, that&#039s equivalent to a B.S. and M.S. – maybe you&#039re a super genius, so is this really an applicable lesson for your average reader to take home? I say this, because I don&#039t think a reader who is attempting something at 4x the pace should feel that they are not being maximally productive if they&#039re not reaching that goal.

    4. If you&#039ve gotten through the first six months, you&#039ve proven that you can do it at this rate. Why push to continue so quickly? There are many valuable experiences in life that you may be missing out on, I am curious why this is more important? Is this actually healthy overall? That puts this in the category of “personally inspired, but overall irrelevant to others experiment, save for inspirational” As a scientist I appreciate this sort of thing but “why?” remains. You can&#039t prove to other people that &#039if you set your mind to something you can do it,&#039 only they can prove that to themselves. All seriousness aside, I wonder if you are super excited about life all the time or on the verge of a nervous breakdown every 72 hours!

    6. You are evaluating yourself based on final exams and projects, but being a successful MIT student also means going through the grind of the method that MIT uses. For any institution, normally exams are more general than the detailed work of problem sets – usually the students come out of the course and school with more knowledge than what is required to pass the exam, more intricate examples, etc, which you probably won&#039t be coming out of this with. Which leads to my last point:

    7. [Here is where I will take a positive turn] What you are in essence proving, isn&#039t that you are doing the work 4x as fast (really 2x), because you&#039re actually not doing the work as an MIT student does it or as it&#039s assigned and subjecting yourself to the idiosyncrasies of a professor or department, or the insecurities of a student deciding on a major and the amount of wiggle room a department creates for that. What you are doing is using a much better method to learn information: a recursive approach to their entire curriculum as well as each class, with, iterative steps in between and you&#039re doing it at a high level, hence using MIT (plus they have their online courses available); I think it&#039s implied but I&#039m stating it clearly for other readers.

    I have studied using a variety of methods, from the poor cramming where I recall nothing 2 weeks later, to the well paced iterative method, to finally, and the best in my opinion, a recursive method. I started this approach after reading How To Read A Book by Mortimer Adler. I think one of the great strengths anyone can possess (and you do) for being efficient is to be able to look at the whole of what you are trying to accomplish and then create a plan and follow that. It&#039s akin to a recursive method in some ways and certainly leads to the experience you&#039ve described: a more intense beginning, and efficiency speeding up as you go.

    Anyway, that&#039s my input. I have a lot of thoughts about the educational system and learning methods that I&#039ll probably post on my blog later on, but I definitely appreciate this experiment, and I certainly enjoy watching you do it, though, I must say, I don&#039t think the project on the whole it&#039s something people should pursue unless it is necessary.

    For the record, I haven&#039t taken computer science or engineering. I studied astrophysics, and so some of the things I have said may be better applied to a physics or mathematics major, but in my experience, people tend to divide: science kids:liberal arts kids:: left brain:right brain . So I went ahead and shared my opinions.

  • Amelia

    As I said, I follow it loosely, so I went and saw your curriculum list, a good list. But I’ll add more to what I said earlier. I also think it’s great that you’re doing it in a year, because a lot of information from year one can drift away by year four, sometimes. It makes recursive learning much more efficient in your own mind and what you might walk out with that others in university might not, is a great overall look at the entire subject of computer science and create a strong web of concepts. Overall, big props for learning this, and enjoy South by Southwest [very jealous]!

  • Aurooba Ahmed

    Even if it’s possible to learn faster without college, not everyone could do it. Many people need the extra motivation and accountability that the social environment of a learning institution can provide. However, I admire your effort and agree that not being in school can help learning progress much faster, in the right circumstances. 🙂

  • Andrew

    Hey Scott, I wanted to chime in and say I do much the same thing. Over the winter break I listened to every calculus video on khan, read The C Programming Language by K&R and listened to an economics lecture series from the teaching company.

    As a result, I have a first run knowledge of how the topic was going to go and homework serves, like you said, as dilberate practice. The only problem I ever have is that the quiz format for my C course is fill in the blank for the way the textbook phrases things. This is annoying but it doesn’t prevent me from rocking the course out.

  • Amelia

    Hi Scott,
    I really believe in deliberate practice when it comes to learning, either in an academic or vocational setting. I see a lot of students reading, re-reading, and copying notes from textbooks and their own class/lecture material… because it’s easy. But when it comes to doing practice questions or demonstrating their ability (which is how they will be assessed most of the time too), students shy away from it. Why? Like you said in a previous post – because there is a possibility that they might get it *wrong*. My philosophy is that you learn a lot more when you get something wrong than when you get something right (usually because you feel silly and think – “I’ll never make that mistake again!”).
    I am curious to know more about your method of choosing the “right” or “hardest possible” questions/assignments to stretch yourself when you are so new to the subject matter. As a tutor, I can do that for my students because I know the subject material inside and out. But when you are new to a topic, it can sometimes be hard to separate the basic concepts from the more complex concepts. Would love to know your thoughts!
    Thanks for the update on your progress too, I’m really enjoying reading and hearing about the fun you’re having 😉
    Cheers from Australia!

  • Prince Sam

    Yea yea!

    Scott is @ it again just as usual . BLAST THE SYSTEM AND GO SCOTT FREE (without pressure ) to your achievements & objectives.

    Quite informative, and enlightening very explicit. Thanks for sharing these realities.

  • Scott Young


    If you check my course list you’ll see I have 8 classes to approximate the HASS requirements and the 6 GIRs. I wasn’t able to replicate all elements of the curriculum, but I’m taking the same number of credit hours as listed in MIT’s description of the 6-1 degree:


    You can’t master everything. It’s about triage–deciding what is maximally beneficial for your time.


    I get 8 hours of sleep each night. Sleeping too little just isn’t worth it, so I’d rather kill an hour or two in the morning if it means I get enough sleep.


    Well I’m not doing everything, but that’s because I’m never doing everything. I talk about a lot of things I do on this blog, but I’m not usually doing everything at the same time (people tend to forget I’ve written articles over the span of 6 years). My reading has gone down quite a bit, I’ve only read about 3 books outside of my curriculum since I started, but that’s mostly because after reading all day I like to go outside and spend time with people.


    True–my analogy wasn’t really clear in the article. I’m borrowing the dichotomy from computer science, where iteration is like a while/for loop and recursion is functional. Recursive methods, like the learning strategy I describe, are usually based on dividing a problem into smaller subproblems in order to solve them.


    1. As I addressed to Steven, my MIT curriculum is the same size (in credit hours) as the one an undergraduate student would take. I did have to substitute courses when I couldn’t do a lab, I picked another CS class. So overall the volume of the curriculum is comparable.

    2. Depends on what you mean by engineering. If you mean engineering projects, then aside from programming assignments, I’m indeed leaving those out. If a whole course was lab work, then I substituted it, so volume is preserved even if there is some drift in the actual curriculum.

    3. As I state in the article, not everyone can learn that fast, that’s true. My point is only to (a) show that it’s possible and (b) suggest that there may be some structural advantages to my approach, which I outline. In the end, my entire experiment is only a sample size of 1, so I’m not trying to prove generalized theories of learning.

    4. If you watch the video, you’ll see that I’m not on the verge of a nervous breakdown. I still watch television, exercise, go on dates, party and I’m flying out to Austin for a vacation. I’m working more than full-time, but I’m hardly exhausted.

    As to the “why”–because it’s a fun challenge! I want to see if I can do it! 🙂

    6. Certainly problem sets explore more difficult problems than are encountered solely on exams. But problem sets are often done as group work and are therefore (in most cases) primarily a teaching tool, not an evaluation tool (which is why they count for so little of the grade in a typical class). My point isn’t to replicate every learning tool used at MIT, but to see if I can be creative to get efficiency gains.

    7. Exactly. I’ve tried to be clear from the beginning that my goal was to *learn* MIT’s curriculum, not prove I can do all the work in one year.

    Thanks for the comment!


    True. But many of the things I describe on this blog aren’t things most people will do on their own. I inevitably speak to a minority.


    True–but it really depends. For example, I already have a university degree, so the marginal benefit of obtaining another undergrad degree is significantly reduced. I’m not saying it’s right for everyone, just some people.

  • Tynan Hitchens

    Well, i think you are working hard sure. But since starting a new job ive learned probably as much as i did in my degree in 6 months. So i learn faster than you, earned 20k and contributed to a major project. Who win?

  • Scott Young


    Haha, well you may have a point. Part of my challenge was showing a more direct comparison in education using an established benchmark.


  • jk

    Cool post. I’ve tried to teach myself many things – sometimes successfully, sometimes not – but the biggest lesson I’ve learned is that the quality of books (and materials) you pick makes a world of difference.

    I’ve browsed through and purchased all sorts of books which seemed to be exactly what I was looking for, only to find a few chapters in that the writing and instruction were unintelligible and convoluted. Occasionally I’ll come across a gem – like Khan Academy, mentioned above, or Morris Kline, great math author worth checking out – which changes everything.

    The difficulty is that I’ve found it takes a lot of digging in order to find a really great resource. But once I’ve found it, it’ll be completely worthwhile, as it’ll make learning more efficient, fun, and meaningful.

    Hope rest of the challenge goes well.

  • Doogie

    What is the feynmann technique? I m confused 🙁

  • Peter

    Thanks a lot for taking the time out of your busy schedule to reply Scott!

    It’s very valuable for me to know that you sleep for 8 hours every night. It reassures me about my findings, after hearing about so many people who say they get 1-5 hrs of sleep a night, and who get a large amount of work done. I’ve found that its possible to get a lot of work done that way, but one loses the extra 10-15% cognitive capacity that consistently getting a full amount of sleep gives.

    That extra capacity is necessary to perform in work that uses one’s full capacity- such as highly creative and deeply thoughtful work like programming(solving fresh problems not ‘grunt work’), as well as learning and practicing these concepts like Scott is doing. That extra 10-15% cognition, allows the solutions with x2 quality and effectiveness to be found.

    That is vs doing work like modelling on a 3d model that has already been fully planned already and is so minimal creativity is needed. I’ve found that I am perfectly able to do this work even for 3 days straight with no sleep.(I don’t regularly do this though) Physical work suffers with lack of sleep also.

    Thanks a lot for the information. Its Very valuable to know, as well as reassuring.

  • Vinay Bhat

    I mean to say, Scott, how are you planning to use this CS knowledge? Surely this challenge was not to just flex some neurons, right?

  • rahul

    I am a great follower of books. They hold so much info, don’t you try books?

  • Shakeel

    Scott, I found out about your MIT Challenge and your blog yesterday and am quite impressed by your studying techniques. I need to go through some more of your posts and videos to get a better understanding of your learning process. I would love to replicate some of your study techniques to learn faster.


  • Maritza van den Heuvel

    Hi Scott

    I fundamentally believe that the way people learn needs to change. In fact, education from primary school onwards needs to change radically if what we call education is going to continue to serve the needs of an increasingly knowledge-based society.

    That’s where I think you brush over Aurooba’s comment about not everybody being able to learn like this too easily. Most people I know are products of a top-down education system and career paths that do not encourage this level of self-learning. In fact, it is often actively discouraged when children at school are taught to learn only what they need to know to pass a test.

    For free education, or at least more open access to learning resources like the MIT library or the Khan library, to have any real chance of changing the world, we need to start teaching more people to learn independently and critically. Otherwise, access to these rich learning resources will become yet another divide.

    It’s a fundamental catch-22, really. Those who lack the skills to learn like this, usually because they were not raised or taught in environments that built these skills, are in fact the ones who can most benefit from the access to these learning resources. Putting the resources down is just half the battle here. The other half – the half that is more likely to ensure a long-term benefit to a broader section of society than just the intellectually savvy and skilled – is still sorely lacking.

    *And yes, I know your experiment is in context of the MIT learning program, therefore geared towards a specific audience. But I think people too easily tout this program as an example of how free online resources can revolutionize education, without thinking of the greater context.

  • Armen Shirvanian

    Hi Scott.

    I continue to be a big supporter of your material. I wanted to comment here in relation to some of the comments you’ve been getting. They’re almost like attacks on you for taking on this effort. When you’re getting attacked regularly, you know you’re doing something right. I think some people are challenged by your challenge because it makes them feel like you’re doing more than they are with your time, or showing your ability to grasp knowledge makes them feel less able.

    The material you are generating from this challenge will be valuable to many for years to come.

    Also, I recently wrote an article about your effort, and so I thought I’d link here for those who may be interested:


  • Ashvin

    I hold this value about medical school as well. The first 2 years of basic sciences can be learned with 6 months of dedication, and some cost. (significantly less than med school tuition)

    1. Kaplan USMLE 1 Video + Lecture Notes.
    2. USMLE World Q Bank
    3. To add lab dissections, look into Acland’s Dissections.
    4. U. Michigan & WebPath have plenty of additional data (CT Scans & Microscopy)
    5. Plenty of YouTube and web resources are freely available.

    What you’re missing out on is face-to-face interaction with a professor and other students, and med school parties.

  • anonymous

    When you count 3 months of summer, 1 month of IAP, and about a collective month and a half of various breaks, one-day holidays, and finals weeks, an MIT student’s academic year is only around 7.5 months, not a year. Add to that the obligations of extracurriculars (depending on involvement, can be anywhere from less than 3 to more than 12 hours a week), research for many students (at minimum 10 hours a week if you’re serious about it), and the fact that many to most students take several classes beyond the requirements out of interest or obligation, and you aren’t really learning at 4x the pace.

    I’m not saying that this would invalidate your attempt to learn the core CS curriculum at MIT, but I think you’re misleading your readers a bit in the conclusion you’re aiming to draw. You might be replicating the MIT curriculum, but– for reasons not limited to those above– I don’t think you’re replicating an MIT education.

  • obed

    if you want to have learning references people go to i’m sure Scott will agree this is a good website to learn from and best of all its freeee!!!!!!!!

  • Erica

    Dear Scott H Young,

    My friend and I were watching your video and wondering, “If this guy is working nine and a half hour days at least five days a week, when does he have time to eat?!” More generally, what is your break schedule like? How do you mentally/physically/spiritually/emotionally recharge to maintain a healthy lifestyle (e.g. exercise, relax, meditate, spend time with family/friends, service to your community)?

    Thanks in advance for the reply!

    With humility,

  • Nahyan

    That’s brilliant.
    I especially liked the point on recursive learning strategy.

  • Fast Listener

    Hi Scott,

    Thank you very much for sharing your effort and becoming a good role model for the humanity. We have many vulnerabilities such as learning for the sake of titles, spending thousands of dollars for education that could be done in half or even quarter period of time, wasting our skills and capacity and so on.

    I am very excited to share my happiness with you: I am also a fast listener and I completed tens of hours of auditory courses in almost half of the regular time. I am pushing the limits and even I founded a website introducing the concept of fast listening. I am willing to create content about foundations and how-to guides about listening faster.

    My website is:

    I wish limitless success to you. Thanks again for being such a good example for humanity.


  • Kyle Panda

    I sense the stirrings of a long overdue education revolution here. To convince people, you need to show them what’s possible.

    What’s cool about Scott’s approach is that its driven more by Intrinsic motivation. [Which the past decades’ research in behavioral economics and psychology has shown to be more effective when it comes to non-mechanical tasks, like learning.]

    Extrinsic motivation like grades are in their proper place as feedback/evaluation for deliberate practice (they’re less emphasized as carrots and sticks for a far-away extrinsic salary increase).

  • Kyle Panda

    ***To convince people, you need to show them what’s possible – in daily life. Not just “educate/inform” them.

  • Michelle@Growx10

    Great learning “hacks!” It is definitely time to revise the entire schooling system!

  • gcbenison

    In computer programming, I think it is important to draw a distinction between “being able to write a working program” and “understanding the core concepts.” The former can be achieved much more quickly and is easily confused with the latter. So I’d caution anyone contemplating an accelerated education to make sure that not only can they “write a program to implement algorithm X”, but that they can choose algorithm X appropriately when confronted with a real-world problem. It’s that kind of judgement that can be developed from years of seemingly tedious study.

  • Nicolas D.

    Also, don’t forget what I call “passive learning”. That is : setup your environment so you are surrounded by what you are learning every day.

    This mean : work for free for someone great in this field, add all the blogs about this subject as you can, read for the community (easier in computing), find people that use this knowledge at work and are passionate of it and keep following them.

    Active learning is great for a knowledge peak, but without passive learning it will become fade.

  • Http://Esut.Gistvine.Com

    Yes! Finally someone writes about engineer.