WEBVTT 00:00:03.646 --> 00:00:06.612 So, this is my grandfather, 00:00:06.612 --> 00:00:08.113 Salman Schocken, 00:00:08.113 --> 00:00:12.550 who was born into a poor and uneducated family 00:00:12.550 --> 00:00:15.987 with six children to feed, 00:00:15.987 --> 00:00:19.683 and when he was 14 years old, he was forced to 00:00:19.683 --> 00:00:23.643 drop out of school in order to help put bread on the table. 00:00:23.643 --> 00:00:27.063 He never went back to school. 00:00:27.063 --> 00:00:31.210 Instead, he went on to build a glittering empire 00:00:31.210 --> 00:00:33.318 of department stores. 00:00:33.318 --> 00:00:36.571 Salman was the consummate perfectionist, 00:00:36.571 --> 00:00:39.135 and every one of his stores was a jewel 00:00:39.135 --> 00:00:41.064 of Bauhaus architecture. 00:00:41.064 --> 00:00:44.254 He was also the ultimate self-learner, 00:00:44.254 --> 00:00:47.157 and like everything else, he did it in grand style. 00:00:47.157 --> 00:00:49.845 He surrounded himself with an entourage 00:00:49.845 --> 00:00:53.232 of young, unknown scholars like Martin Buber 00:00:53.232 --> 00:00:56.371 and Shai Agnon and Franz Kafka, 00:00:56.371 --> 00:00:59.296 and he paid each one of them a monthly salary 00:00:59.296 --> 00:01:02.347 so that they could write in peace. NOTE Paragraph 00:01:02.347 --> 00:01:07.284 And yet, in the late '30s, Salman saw what's coming. 00:01:07.284 --> 00:01:10.598 He fled Germany, together with his family, 00:01:10.598 --> 00:01:12.687 leaving everything else behind. 00:01:12.687 --> 00:01:15.897 His department stores confiscated, 00:01:15.897 --> 00:01:18.986 he spent the rest of his life in a relentless pursuit 00:01:18.986 --> 00:01:21.455 of art and culture. 00:01:21.455 --> 00:01:23.394 This high school dropout 00:01:23.394 --> 00:01:25.674 died at the age of 82, 00:01:25.674 --> 00:01:29.803 a formidable intellectual, cofounder and first CEO 00:01:29.803 --> 00:01:31.948 of the Hebrew University of Jerusalem, 00:01:31.948 --> 00:01:33.712 and founder of Schocken Books, 00:01:33.712 --> 00:01:37.083 an acclaimed imprint that was later acquired 00:01:37.083 --> 00:01:38.753 by Random House. 00:01:38.753 --> 00:01:42.729 Such is the power of self-study. NOTE Paragraph 00:01:42.729 --> 00:01:45.245 And these are my parents. 00:01:45.245 --> 00:01:49.292 They too did not enjoy the privilege of college education. 00:01:49.292 --> 00:01:53.263 They were too busy building a family and a country. 00:01:53.263 --> 00:01:57.291 And yet, just like Salman, they were lifelong, 00:01:57.291 --> 00:02:01.049 tenacious self-learners, and our home was stacked 00:02:01.049 --> 00:02:05.212 with thousands of books, records and artwork. 00:02:05.212 --> 00:02:07.608 I remember quite vividly my father telling me 00:02:07.608 --> 00:02:12.440 that when everyone in the neighborhood will have a TV set, 00:02:12.440 --> 00:02:17.058 then we'll buy a normal F.M. radio. (Laughter) NOTE Paragraph 00:02:17.058 --> 00:02:20.087 And that's me, 00:02:20.087 --> 00:02:21.811 I was going to say holding my first abacus, 00:02:21.811 --> 00:02:24.822 but actually holding what my father would consider 00:02:24.822 --> 00:02:27.774 an ample substitute to an iPad. (Laughter) 00:02:27.774 --> 00:02:31.436 So one thing that I took from home is this notion 00:02:31.436 --> 00:02:34.277 that educators don't necessarily have to teach. 00:02:34.277 --> 00:02:37.506 Instead, they can provide an environment and resources 00:02:37.506 --> 00:02:42.654 that tease out your natural ability to learn on your own. 00:02:42.654 --> 00:02:46.748 Self-study, self-exploration, self-empowerment: 00:02:46.748 --> 00:02:50.068 these are the virtues of a great education. NOTE Paragraph 00:02:50.068 --> 00:02:54.263 So I'd like to share with you a story about a self-study, 00:02:54.263 --> 00:02:56.817 self-empowering computer science course 00:02:56.817 --> 00:03:01.003 that I built, together with my brilliant colleague Noam Nisan. 00:03:01.003 --> 00:03:03.760 As you can see from the pictures, both Noam and I 00:03:03.760 --> 00:03:07.178 had an early fascination with first principles, 00:03:07.178 --> 00:03:08.700 and over the years, as our knowledge of 00:03:08.700 --> 00:03:12.275 science and technology became more sophisticated, 00:03:12.275 --> 00:03:15.471 this early awe with the basics 00:03:15.471 --> 00:03:17.571 has only intensified. 00:03:17.571 --> 00:03:21.270 So it's not surprising that, about 12 years ago, when 00:03:21.270 --> 00:03:24.752 Noam and I were already computer science professors, 00:03:24.752 --> 00:03:28.504 we were equally frustrated by the same phenomenon. 00:03:28.504 --> 00:03:31.322 As computers became increasingly more complex, 00:03:31.322 --> 00:03:34.526 our students were losing the forest for the trees, 00:03:34.526 --> 00:03:37.613 and indeed, it is impossible to connect 00:03:37.613 --> 00:03:40.255 with the soul of the machine if you interact 00:03:40.255 --> 00:03:43.664 with a black box P.C. or a Mac which is shrouded 00:03:43.664 --> 00:03:47.040 by numerous layers of closed, proprietary software. 00:03:47.040 --> 00:03:50.730 So Noam and I had this insight that if we want our students 00:03:50.730 --> 00:03:52.749 to understand how computers work, 00:03:52.749 --> 00:03:55.837 and understand it in the marrow of their bones, 00:03:55.837 --> 00:03:57.398 then perhaps the best way to go about it 00:03:57.398 --> 00:04:01.992 is to have them build a complete, working, 00:04:01.992 --> 00:04:05.704 general-purpose, useful computer, hardware and software, 00:04:05.704 --> 00:04:09.234 from the ground up, from first principles. NOTE Paragraph 00:04:09.234 --> 00:04:13.442 Now, we had to start somewhere, and so Noam and I 00:04:13.442 --> 00:04:16.932 decided to base our cathedral, so to speak, 00:04:16.932 --> 00:04:20.071 on the simplest possible building block, 00:04:20.071 --> 00:04:22.872 which is something called NAND. 00:04:22.872 --> 00:04:26.810 It is nothing more than a trivial logic gate 00:04:26.810 --> 00:04:29.848 with four input-output states. 00:04:29.848 --> 00:04:32.514 So we now start this journey by telling our students 00:04:32.514 --> 00:04:34.582 that God gave us NAND — (Laughter) — 00:04:34.582 --> 00:04:38.567 and told us to build a computer, and when we asked how, 00:04:38.567 --> 00:04:41.751 God said, "One step at a time." 00:04:41.751 --> 00:04:44.844 And then, following this advice, we start 00:04:44.844 --> 00:04:47.803 with this lowly, humble NAND gate, 00:04:47.803 --> 00:04:49.743 and we walk our students through an elaborate sequence 00:04:49.743 --> 00:04:53.480 of projects in which they gradually build a chip set, 00:04:53.480 --> 00:04:57.592 a hardware platform, an assembler, a virtual machine, 00:04:57.592 --> 00:05:01.169 a basic operating system and a compiler 00:05:01.169 --> 00:05:07.358 for a simple, Java-like language that we call "JACK." 00:05:07.358 --> 00:05:09.906 The students celebrate the end of this tour de force 00:05:09.906 --> 00:05:12.740 by using JACK to write all sorts of cool games 00:05:12.740 --> 00:05:15.732 like Pong, Snake and Tetris. 00:05:15.732 --> 00:05:19.347 You can imagine the tremendous joy of playing 00:05:19.347 --> 00:05:22.371 with a Tetris game that you wrote in JACK 00:05:22.371 --> 00:05:25.161 and then compiled into machine language in a compiler 00:05:25.161 --> 00:05:27.569 that you wrote also, and then seeing the result 00:05:27.569 --> 00:05:30.452 running on a machine that you built starting 00:05:30.452 --> 00:05:33.736 with nothing more than a few thousand NAND gates. 00:05:33.736 --> 00:05:36.926 It's a tremendous personal triumph of going 00:05:36.926 --> 00:05:41.651 from first principles all the way to a fantastically complex 00:05:41.651 --> 00:05:44.002 and useful system. NOTE Paragraph 00:05:44.002 --> 00:05:48.399 Noam and I worked five years to facilitate 00:05:48.399 --> 00:05:51.607 this ascent and to create the tools and infrastructure 00:05:51.607 --> 00:05:54.780 that will enable students to build it in one semester. 00:05:54.780 --> 00:05:58.819 And this is the great team that helped us make it happen. 00:05:58.819 --> 00:06:03.066 The trick was to decompose the computer's construction 00:06:03.066 --> 00:06:06.020 into numerous stand-alone modules, 00:06:06.020 --> 00:06:09.920 each of which could be individually specified, 00:06:09.920 --> 00:06:15.155 built and unit-tested in isolation from the rest of the project. 00:06:15.155 --> 00:06:18.144 And from day one, Noam and I decided to put 00:06:18.144 --> 00:06:21.680 all these building blocks freely available in open source 00:06:21.680 --> 00:06:23.160 on the Web. 00:06:23.160 --> 00:06:27.631 So chip specifications, APIs, project descriptions, 00:06:27.631 --> 00:06:31.535 software tools, hardware simulators, CPU emulators, 00:06:31.535 --> 00:06:35.237 stacks of hundreds of slides, lectures -- 00:06:35.237 --> 00:06:37.464 we laid out everything on the Web 00:06:37.464 --> 00:06:40.491 and invited the world to come over, 00:06:40.491 --> 00:06:42.027 take whatever they need, 00:06:42.027 --> 00:06:44.417 and do whatever they want with it. NOTE Paragraph 00:06:44.417 --> 00:06:48.316 And then something fascinating happened. 00:06:48.316 --> 00:06:49.899 The world came. 00:06:49.899 --> 00:06:53.057 And in short order, thousands of people 00:06:53.057 --> 00:06:54.601 were building our machine. 00:06:54.601 --> 00:06:58.525 And NAND2Tetris became one of the first 00:06:58.525 --> 00:07:01.755 massive, open, online courses, 00:07:01.755 --> 00:07:04.499 although seven years ago we had no idea that what 00:07:04.499 --> 00:07:06.660 we were doing is called MOOCs. 00:07:06.660 --> 00:07:10.518 We just observed how self-organized courses 00:07:10.518 --> 00:07:13.600 were kind of spontaneously spawning 00:07:13.600 --> 00:07:14.983 out of our materials. 00:07:14.983 --> 00:07:18.074 For example, Pramode C.E., 00:07:18.074 --> 00:07:19.983 an engineer from Kerala, India, 00:07:19.983 --> 00:07:21.792 has organized groups of self-learners 00:07:21.792 --> 00:07:24.836 who build our computer under his good guidance. 00:07:24.836 --> 00:07:27.858 And Parag Shah, another engineer, from Mumbai, 00:07:27.858 --> 00:07:30.703 has unbundled our projects into smaller, 00:07:30.703 --> 00:07:32.725 more manageable bites that he now serves 00:07:32.725 --> 00:07:36.527 in his pioneering do-it-yourself computer science program. NOTE Paragraph 00:07:36.527 --> 00:07:39.938 The people who are attracted to these courses 00:07:39.938 --> 00:07:42.358 typically have a hacker mentality. 00:07:42.358 --> 00:07:44.463 They want to figure out how things work, 00:07:44.463 --> 00:07:46.223 and they want to do it in groups, 00:07:46.223 --> 00:07:49.034 like this hackers club in Washington, D.C., 00:07:49.034 --> 00:07:52.699 that uses our materials to offer community courses. 00:07:52.699 --> 00:07:55.645 And because these materials are widely available 00:07:55.645 --> 00:07:58.339 and open-source, different people take them 00:07:58.339 --> 00:08:01.282 to very different and unpredictable directions. 00:08:01.282 --> 00:08:04.195 For example, Yu Fangmin, from Guangzhou, 00:08:04.195 --> 00:08:07.041 has used FPGA technology 00:08:07.041 --> 00:08:10.766 to build our computer and show others how to do the same 00:08:10.766 --> 00:08:14.129 using a video clip, and Ben Craddock developed 00:08:14.129 --> 00:08:17.730 a very nice computer game that unfolds 00:08:17.730 --> 00:08:22.699 inside our CPU architecture, which is quite a complex 00:08:22.699 --> 00:08:25.184 3D maze that Ben developed 00:08:25.184 --> 00:08:29.092 using the Minecraft 3D simulator engine. 00:08:29.092 --> 00:08:32.110 The Minecraft community went bananas over this project, 00:08:32.110 --> 00:08:35.494 and Ben became an instant media celebrity. NOTE Paragraph 00:08:35.494 --> 00:08:38.396 And indeed, for quite a few people, 00:08:38.396 --> 00:08:42.473 taking this NAND2Tetris pilgrimage, if you will, 00:08:42.473 --> 00:08:45.185 has turned into a life-changing experience. 00:08:45.185 --> 00:08:48.374 For example, take Dan Rounds, who is a music 00:08:48.374 --> 00:08:51.143 and math major from East Lansing, Michigan. 00:08:51.143 --> 00:08:54.587 A few weeks ago, Dan posted a victorious post 00:08:54.587 --> 00:08:57.207 on our website, and I'd like to read it to you. 00:08:57.207 --> 00:09:00.665 So here's what Dan said. NOTE Paragraph 00:09:00.665 --> 00:09:03.477 "I did the coursework because understanding computers 00:09:03.477 --> 00:09:06.823 is important to me, just like literacy and numeracy, 00:09:06.823 --> 00:09:09.840 and I made it through. I never worked harder on anything, 00:09:09.840 --> 00:09:12.549 never been challenged to this degree. 00:09:12.549 --> 00:09:14.936 But given what I now feel capable of doing, 00:09:14.936 --> 00:09:17.094 I would certainly do it again. 00:09:17.094 --> 00:09:19.344 To anyone considering NAND2Tetris, 00:09:19.344 --> 00:09:23.090 it's a tough journey, but you'll be profoundly changed." NOTE Paragraph 00:09:23.090 --> 00:09:27.990 So Dan demonstrates the many self-learners 00:09:27.990 --> 00:09:32.636 who take this course off the Web, on their own traction, 00:09:32.636 --> 00:09:37.225 on their own initiative, and it's quite amazing because 00:09:37.225 --> 00:09:41.812 these people cannot care less about 00:09:41.812 --> 00:09:43.054 grades. 00:09:43.054 --> 00:09:46.959 They are doing it because of one motivation only. 00:09:46.959 --> 00:09:50.445 They have a tremendous passion to learn. NOTE Paragraph 00:09:50.445 --> 00:09:52.369 And with that in mind, 00:09:52.369 --> 00:09:56.973 I'd like to say a few words about traditional college grading. 00:09:56.973 --> 00:09:59.333 I'm sick of it. 00:09:59.333 --> 00:10:00.766 We are obsessed with grades 00:10:00.766 --> 00:10:02.632 because we are obsessed with data, 00:10:02.632 --> 00:10:07.219 and yet grading takes away all the fun from failing, 00:10:07.219 --> 00:10:09.695 and a huge part of education 00:10:09.695 --> 00:10:11.286 is about failing. 00:10:11.286 --> 00:10:13.761 Courage, according to Churchill, 00:10:13.761 --> 00:10:16.707 is the ability to go from one defeat to another 00:10:16.707 --> 00:10:19.321 without losing enthusiasm. (Laughter) 00:10:19.321 --> 00:10:23.151 And [Joyce] said that mistakes 00:10:23.151 --> 00:10:24.620 are the portals of discovery. 00:10:24.620 --> 00:10:27.966 And yet we don't tolerate mistakes, 00:10:27.966 --> 00:10:29.275 and we worship grades. 00:10:29.275 --> 00:10:33.063 So we collect your B pluses and your A minuses 00:10:33.063 --> 00:10:35.900 and we aggregate them into a number like 3.4, 00:10:35.900 --> 00:10:37.811 which is stamped on your forehead 00:10:37.811 --> 00:10:40.254 and sums up who you are. 00:10:40.254 --> 00:10:43.537 Well, in my opinion, we went too far with this nonsense, 00:10:43.537 --> 00:10:46.218 and grading became degrading. NOTE Paragraph 00:10:46.218 --> 00:10:51.119 So with that, I'd like to say a few words about upgrading, 00:10:51.119 --> 00:10:55.544 and share with you a glimpse from my current project, 00:10:55.544 --> 00:10:57.574 which is different from the previous one, 00:10:57.574 --> 00:10:59.588 but it shares exactly the same characteristics 00:10:59.588 --> 00:11:03.415 of self-learning, learning by doing, 00:11:03.415 --> 00:11:06.190 self-exploration and community-building, 00:11:06.190 --> 00:11:12.050 and this project deals with K-12 math education, 00:11:12.050 --> 00:11:13.954 beginning with early age math, 00:11:13.954 --> 00:11:18.447 and we do it on tablets because we believe that 00:11:18.447 --> 00:11:22.852 math, like anything else, should be taught hands on. NOTE Paragraph 00:11:22.852 --> 00:11:26.007 So here's what we do. Basically, we developed 00:11:26.007 --> 00:11:29.765 numerous mobile apps, every one of them explaining 00:11:29.765 --> 00:11:31.262 a particular concept in math. 00:11:31.262 --> 00:11:34.703 So for example, let's take area. 00:11:34.703 --> 00:11:37.175 When you deal with a concept like area -- 00:11:37.175 --> 00:11:42.419 well, we also provide a set of tools that the child 00:11:42.419 --> 00:11:45.050 is invited to experiment with in order to learn. 00:11:45.050 --> 00:11:49.003 So if area is what interests us, then one thing 00:11:49.003 --> 00:11:53.141 which is natural to do is to tile the area 00:11:53.141 --> 00:11:56.791 of this particular shape and simply count 00:11:56.791 --> 00:12:00.654 how many tiles it takes to cover it completely. 00:12:00.654 --> 00:12:03.136 And this little exercise here gives you a first 00:12:03.136 --> 00:12:06.565 good insight of the notion of area. NOTE Paragraph 00:12:06.565 --> 00:12:09.355 Moving along, what about the area of this figure? 00:12:09.355 --> 00:12:14.012 Well, if you try to tile it, it doesn't work too well, does it. 00:12:14.012 --> 00:12:16.316 So instead, you can experiment 00:12:16.316 --> 00:12:18.996 with these different tools here by some process 00:12:18.996 --> 00:12:21.035 of guided trial and error, 00:12:21.035 --> 00:12:23.873 and at some point you will discover that one thing 00:12:23.873 --> 00:12:26.961 that you can do among several legitimate transformations 00:12:26.961 --> 00:12:29.792 is the following one. You can cut the figure, 00:12:29.792 --> 00:12:33.408 you can rearrange the parts, you can glue them 00:12:33.408 --> 00:12:36.525 and then proceed to tile just like we did before. 00:12:36.525 --> 00:12:41.800 (Applause) 00:12:41.800 --> 00:12:44.523 Now this particular transformation 00:12:44.523 --> 00:12:48.507 did not change the area of the original figure, 00:12:48.507 --> 00:12:50.694 so a six-year-old who plays with this 00:12:50.694 --> 00:12:53.578 has just discovered a clever algorithm 00:12:53.578 --> 00:12:57.618 to compute the area of any given parallelogram. NOTE Paragraph 00:12:57.618 --> 00:12:59.474 We don't replace teachers, by the way. 00:12:59.474 --> 00:13:02.603 We believe that teachers should be empowered, not replaced. NOTE Paragraph 00:13:02.603 --> 00:13:05.898 Moving along, what about the area of a triangle? 00:13:05.898 --> 00:13:08.693 So after some guided trial and error, 00:13:08.693 --> 00:13:12.628 the child will discover, with or without help, 00:13:12.628 --> 00:13:16.486 that he or she can duplicate the original figure 00:13:16.486 --> 00:13:19.894 and then take the result, transpose it, 00:13:19.894 --> 00:13:23.642 glue it to the original and then proceed [with] what we did before: 00:13:23.642 --> 00:13:31.370 cut, rearrange, paste — oops— paste and glue, 00:13:31.370 --> 00:13:32.966 and tile. 00:13:32.966 --> 00:13:36.793 Now this transformation has doubled the area 00:13:36.793 --> 00:13:40.643 of the original figure, and therefore we have just learned 00:13:40.643 --> 00:13:44.409 that the area of the triangle equals the area of this rectangle 00:13:44.409 --> 00:13:46.863 divided by two. 00:13:46.863 --> 00:13:50.311 But we discovered it by self-exploration. NOTE Paragraph 00:13:50.311 --> 00:13:56.391 So, in addition to learning some useful geometry, 00:13:56.391 --> 00:14:00.778 the child has been exposed to some pretty sophisticated 00:14:00.778 --> 00:14:03.805 science strategies, like reduction, 00:14:03.805 --> 00:14:06.719 which is the art of 00:14:06.719 --> 00:14:10.305 transforming a complex problem into a simple one, 00:14:10.305 --> 00:14:12.674 or generalization, which is at the heart 00:14:12.674 --> 00:14:15.517 of any scientific discipline, 00:14:15.517 --> 00:14:18.107 or the fact that some properties are invariant 00:14:18.107 --> 00:14:21.131 under some transformations. 00:14:21.131 --> 00:14:24.416 And all this is something that a very young child 00:14:24.416 --> 00:14:28.725 can pick up using such mobile apps. 00:14:28.725 --> 00:14:31.819 So presently, we are doing the following: 00:14:31.819 --> 00:14:36.128 First of all, we are decomposing the K-12 math curriculum 00:14:36.128 --> 00:14:39.019 into numerous such apps. 00:14:39.019 --> 00:14:41.596 And because we cannot do it on our own, 00:14:41.596 --> 00:14:44.690 we've developed a very fancy authoring tool 00:14:44.690 --> 00:14:47.965 that any author, any parent or actually anyone 00:14:47.965 --> 00:14:50.318 who has an interest in math education, 00:14:50.318 --> 00:14:53.590 can use this authoring tool to develop similar apps 00:14:53.590 --> 00:14:56.995 on tablets without programming. 00:14:56.995 --> 00:15:00.215 And finally, we are putting together an adaptive ecosystem 00:15:00.215 --> 00:15:03.043 that will match different learners 00:15:03.043 --> 00:15:08.420 with different apps according to their evolving learning style. NOTE Paragraph 00:15:08.420 --> 00:15:10.941 The driving force behind this project 00:15:10.941 --> 00:15:13.888 is my colleague Shmulik London, 00:15:13.888 --> 00:15:16.532 and, you see, just like 00:15:16.532 --> 00:15:20.605 Salman did about 90 years ago, 00:15:20.605 --> 00:15:24.576 the trick is to surround yourself with brilliant people, 00:15:24.576 --> 00:15:27.332 because at the end, 00:15:27.332 --> 00:15:29.235 it's all about people. 00:15:29.235 --> 00:15:32.654 And a few years ago, I was walking in Tel Aviv 00:15:32.654 --> 00:15:35.484 and I saw this graffiti on a wall, 00:15:35.484 --> 00:15:37.275 and I found it so compelling 00:15:37.275 --> 00:15:39.418 that by now I preach it to my students, 00:15:39.418 --> 00:15:41.684 and I'd like to try to preach it to you. 00:15:41.684 --> 00:15:43.424 Now, I don't know how many people here are familiar 00:15:43.424 --> 00:15:45.252 with the term "mensch." 00:15:45.252 --> 00:15:47.860 It basically means to be human 00:15:47.860 --> 00:15:49.795 and to do the right thing. 00:15:49.795 --> 00:15:52.224 And with that, what this graffiti says is, 00:15:52.224 --> 00:15:54.310 "High-tech schmigh-tech. 00:15:54.310 --> 00:15:56.873 The most important thing is to be a mensch." (Laughter) 00:15:56.873 --> 00:16:00.268 Thank you. (Applause) 00:16:00.268 --> 00:16:05.051 (Applause)