The Science Colloquy: Professor Hal Abelson, MIT

Interviewed on August 30, 2020

Harold (Hal) Abelson is the Class of 1922 Professor of Electrical Engineering and Computer Science at MIT, a fellow of the IEEE, and is a founding director, both Creative Commons and the Free Software Foundation.

Professor Abelson holds an A.B. degree from Princeton University and a Ph.D. degree in mathematics from MIT. In 1992, Abelson was designated as one of MIT's six inaugural MacVicar Faculty Fellows, in recognition of his significant and sustained contributions to teaching and undergraduate education. Abelson was the recipient in 1992 of the Bose Award (MIT's School of Engineering teaching award). Abelson is also the winner of the 1995 Taylor L. Booth Education Award given by IEEE Computer Society, cited for his continued contributions to the pedagogy and teaching of introductory computer science, winner of the 2012 ACM Special Interest Group on Computer Science Education Award for Outstanding Contribution to Computer Science Education, and winner of the 2011 ACM Karl Karlstrom Outstanding Educator Award.

Interviewed by Satvik Tripathi, Founder & Head, Techvik.

Satvik Tripathi (ST): Sir, You grew up on a chicken farm in New Jersey, about 50 miles from New York, what led you from a chicken farm to computer science?

Hal Abelson (HA): Well, I've always been interested in math, science, and physics and things like that when I was in high school. I got a job, doing some computer work at a local airbase station, where I think they had one of the very first computers, which we programmed by punched paper tape. We did a few things, in those days there wasn't much that you could do. So I had some experience with computers. When I went to college, I majored in mathematics and I got a job in programming again. I was at Princeton University and I was programming one of Princeton's first computers, virtually mostly as a game, as a hobby, and for helping people, and that was an IBM 360. That now seems like an ancient, ancient computer.

ST: And what were you working at back then?

HA: I think we were doing some amount of text formatting, which most of the scientists thought was a complete waste of time. “Why would anyone want to do that?”. thought the Princeton should be used for solving equations and doing astrophysics. But we got a little bit of time to play with the computer for doing these very non-serious things. I mean, why would you waste a computer doing things like text processing, printing, and publishing when the serious work was all going to be mathematics and astrophysics. But then I ended up as one of the few people at Princeton who knew about the new language IBM was promoting.

The new IBM language called PL/1, which was supposedly meant to be the general language that was going to solve all people's problems. Mostly, I worked at the computer center as a consultant, helping faculty think about changing their programs from Fortran to PL/1. what I did as an undergraduate. When I went to graduate school, I went to MIT and I got my Ph.D. in mathematics. Soon after arrived at MIT I got involved in the Logo research project that was run by Professor Seymour Papert, aimed at teaching kids about mathematics and things by having the program in a computer language called Logo -- the first computer language designed for kids.

So I don’t know how it very much relates to the chicken farm.

ST: Which scientist do you look up to, one from the past and one from the present?

HA: Wow. That's hard because there are so many. Well, I mean one of the great Geniuses was Richard Feynman. He was just so incredibly creative, has done so many things, and just has a brilliant track record of original views.

At present, I’m impressed by a lot of my colleagues at MIT. I mean, of course, Papert, who died about 10 years ago. He was somebody I worked with. And again, he had a marvelous way of integrating deep mathematical ideas to working and collaborating with kids, he had a very different outlook on how we can work with kids. They're so many outstanding people at MIT that you want to emulate in various ways. It’s really hard to pick.

ST: You have seen the actual evolution of computers and computer science, What major difference has come in computer science in recent years?

HA: Oh well, the major difference is that computers now have an impact on ordinary people. So I remember when we started working on the Logo, this would have been around 1970 or the end of the 1960s, computers cost a million dollars. And those computers were used by the military and large companies doing data processing and the whole idea of a personal computer wasn't even around. Nobody didn't even envision that until, a little bit into the 70s when Alan Kay and Xerox talked about the Dynabook. But the notion that you'd have a computer, that it would be something that you would use personally just wasn't there much. And when we started the Logo project, our motto said “computers are for kids.” And that seemed, not revolutionary, just crazy. How can anybody think that you take these giant expensive million-dollar machines and make it something that a kid can use? Everyone knew it was for serious data processing and military calculations. So that's an enormous, enormous change between then and now. I mean, when you think of a computer now, you think of a personal computer. And that just did not exist 50 years ago.

ST: Yeah, it's like a crazy thing to think about how giant and less efficient computers were back then, and now compared to that we have Macbooks and iPads, which is like maybe 20 or 40 times more efficient than that.

HA: Oh more than that, you know. You know that your mobile phone has more computing power than that used by all of NASA to do the support computing for the first moon landing in 1969. Your mobile phone has more computing power than all of NASA did on that day.

ST: Okay, so what are the areas that you're finding most compelling right now?

HA: Well, what's happening right now is that everyone is fixated on machine learning. There's a real issue about how deep that is, but it certainly has an enormous, enormous impact. So, I mean that's certainly what everyone is experiencing and thinking about. hat's more interesting? There's the work that's going into the computer-brain interface Everyone's waiting for Elon Musk to demonstrate a neural link. Right? He said about having it working in a pig. So we'll see what this pig can do with a direct brain interlink to a computer and then see what happens after that.

There’s enormous potential in merging psychology and neuroscience with the work that's going on in machine learning and artificial intelligence. I don't know how we could even predict what's going to happen. But it's going to be striking and revolutionary.

ST: Indeed, it’s always very intriguing to see what Elon Musk is doing and the neural link is something people have been hyped up about for several years now.

ST: So then, what advice would you give to students who are aspiring to go into computer science or are pursuing CS Major?

HA: The main thing is to be broad-minded. We see a lot of students who get a lot of experience in high school and they focus and maybe they program in Java or JavaScript and they do a lot of things, but they don't do much of anything else. So it turns out, if you want to do something good in computing, you have to apply it and know about something else. One easy thing people can do is solve some puzzles and get some experience in advanced mathematics. There are so many applications of computing, more than any past decade, and you can try various things by actually getting some experience applying it. So, I would suggest that if you want to prepare yourself in computer science, you have to think about computer science and ‘what is it for?’

ST: This is a very interesting point of view, to understand the importance of ‘and’ in computer science. Maybe that’s the reason for computer science having such interdisciplinary applications.

HA: Yes, absolutely right.

ST: I recently started with your course, 6.001: Structure and Interpretation of Computer.

HA: That’s amazing, thank you!

ST: And in the very first lecture, you said, “Computer science is not science, maybe engineering or arts, and has a lot in common with Magic” Can you please explain the idea behind it?

HA: Computer science is very much an experimental science. You kind of has to see what works. A lot of people use it without really understanding how powerful it is. You've written some very nice stuff about that too. It's when you have these systems that do something complicated, even the experts, using machine learning, say it has to be experimental and you have to wait and see what the algorithm does. So that's not very much of a science, and even you know, perfectly well, even the people who are designing these things tend to say it's an art. I mean you'll get to know some of the world's top designers and you will ask them “how did you do it”? “Well, they answer, you get experience and it's an art” all of these things I think what Jerry Sussman and I said in those lectures in the 1980s are even more true now.

ST: How do you think taking classes like psychology, linguistic, or cognitive science would help in the understanding of Machine learning or Artificial Intelligence?

HA: Oh, they're very very strong links in that. Some of the original Artificial intelligence programs and research were directly inspired by ideas from psychology and linguistics and all of those things. That's been true historically and it's still true. Even at MIT right now, there are very strong links between people conducting research in computer science and neuroscience. As I just said, I think one of the really exciting upcoming things is to be able to make brain-computer interfaces, and so the question is can you use the ideas of computer science to understand how the brain might be organized. And at the same time, can you use what's being understood about neuroscience to think about the new kinds of computing structures and architectures? It is also very important to have a strong sense of liberal arts subjects. So there's an enormous enormous impact there.

ST: What do you think has better outcomes? Taking liberal arts classes as well as computer science, or getting out there and having hands-on experiences with research or software development?

HA: Well, who are we talking about, and what age people?

ST: Mostly Graduates and Undergraduate students pursuing computer science or a related subject.

HA: It’s important right now to get some experience by doing things. But of course, as I said, you want to get a broad background. The key is computing ... and what? And, of course, liberal arts in that sense is really important. What is worth making? And again, you have an enormous impact from having a broad education, thinking about liberal arts, thinking about the impact of what you're designing because. Is there an impact on the lives of people? And what should you think deeply about if you're going to be designing something? What effect will that have on people That t has tremendous implications? So taking classes in computer science versus taking classes in liberal art are both important and it’s not an either-or.

ST: I also started reading your book Blown to Bits: Your Life, Liberty, and Happiness After the Digital Explosion. It is a very interesting read.

ST: If you could add a chapter to the book now in 2020, what would it be about?

HA: Well, we're coming out with a new edition. We've just sent it to the publisher. I suspect it's going to be a couple of months before it's out, but the big difference is expanding the way people think about privacy. That has changed since that book came out 12 years ago. There's just so much more impact, so much more on data collection that people read about that in the papers every day. People still don't comprehend the scope of it and what it could be. So that's a big change.

ST: Yes definitely that’s one of the top concerns right now, it's still like people believe that for something like privacy violation or data collection is basically someone hacking into your computer, but in real life, it’s something far more complex and dangerous.

HA: Yeah, that's exactly right. I mean you put it very well.

ST: Will the new book have the same title as the previous one?

HA: It’s the same title. But of course, some of the stories in ‘Blown to bits’ some of the stories are very much out of date. So there are new stories for today. I mean the stuff that was striking when we made the first edition of that book is not as interesting or not as striking now. The basic chapters are still very much the same. But they're different examples and emphasis on people understanding the implications of data collection and being able to filter data.

ST: You were one of the co-founders of MIT OCW, a resource that I really admire and have taken various courses on it as well.

HA: Yeah, yeah, that was wonderful. Just a wonderful thing that we started that in, gosh it’s almost 20 years ago, 2000 or 2001.

ST: So, how do you think resources like OpenCourseWare are different from learning platforms such as edX or Coursera?

HA: Well the open courses first of all are much broader. OCW is specifically designed so that anyone can get it -- not only get it but also adapt and reuse it. Modify and redistribute it. that's a little harder to do with things like edX and Coursera. All those are great things too and have been very popular lately. But the question when we put out OpenCourseWare(OCW) -- the image of MIT -- is not that MIT becomes the place that defines knowledge and courses that everyone is going to us Rather it's that MIT makes raw materials that other people can build on. hat's the essential difference there.

ST: Will it be beneficial to add course completion certificates for OCW?

HA: We can imagine doing something like that with OCW but MIT has been more doing that with respect to edX. Coursera has been doing that because typically when you want to do course completion, somebody has to demonstrate something that says they've actually completed the course. So I think that might be a good thing. I'll talk with the people who are running OCW right now if they've even thought about it, but so far at MIT, at least they've been leaving that for edX. It’s but completion certificate is a good idea. Thank you.

ST: It’s my pleasure, sir. You also led the MIT App Inventor project, and it had a huge success. More than 34 million apps have been built so far and around 8.2 million registered users.

ST: What was the motivation or idea behind the MIT App Inventor project? Do you see it as an educational piece?

HA: To answer your questions in reverse order: The application of App Inventors for education -- right now, we are majorly focusing on Middle School, and even Primary School, and is about the question of how do you get kids involved in computing, getting them to know that they can build things. And again, in terms of computer science, related to what I said before, the big idea that you want to get across to kids at maybe 11 years old and 12 years old is that they really can make something useful. So it's not so much of the basic technical knowledge of computing, but the fact that you can actually make something. Relative to what's happened in the last five years, the structure of the Internet and services are becoming powerful enough that even young children can do significant development. Even without deep knowledge of coding, you can build apps that can contribute to societal change. That's really changed. One of my favorite examples is... do you know where Dharavi is outside of Mumbai?

ST: Yes sir

HA: Okay, so there’s a group of girls in Dharavi. They are making things using App Inventor to help their community. So one of the things that they built is an app that schedules your time at the community water distribution plan. And people around Dharavi, who have mobile phones, can access that and schedule something. That wouldn't have been easy to make 10 years ago. Of course, that infrastructure didn't exist. And so, one of the really exciting things we're dealing with in App Inventor is something we call ‘computational action.’ Computational action means I can make things that actually have an impact on my life and my family and my community and maybe even the world. And what's really exciting right now is what we were talking about before, about how individual computers are getting powerful, but it's not just the individual computers. It's the infrastructure that allows you to do something. You or anybody can make something that says I have a map and I'm going to put things on the map and anybody in the world can see stuff on the map. No one could do that 10 years ago.

The thing that's happening is that we are talking about Al and machine learning. There are now these services on the web and large databases so that even children can think of writing about significant things. They can look at large scientific databases and large social databases. And they also can get access to the machine learning infrastructure that allows you to process things. That's a change. I mean that's not a thing that kids could do even two years ago. And so when I look-ahead about what's coming out of that App Inventor, it's about increasingly making this power of computing and information available to kids, you know, ten-year-old or nine-year-old kids -- that's a significant change. So one of the things that we are focusing on the App Inventor. It really is about getting kids access to this incredible power and understanding that their wishes and ideas can make a real change now.

ST: I totally agree on that, I still remember it was a great feeling for me to print “hello world” using C++ and that emotion of joy and self-empowerment had somewhat led me to dive more into coding and AI.

HA: Yes, definitely. By the way, I was really impressed with your article on Federated learning. How deeply you have explained it is very very impressive and remarkable for your age.

ST: Thank you so much, sir, means a lot to me.

ST: What would you like App Inventor to become?

HA: Well, our App Inventor curriculum hasn't really talked about Al a lot. But as I just said, both the phones and infrastructure are getting powerful enough to do that. The kinds of applications people are talking about now, even the very important ones, is are about what you think of as large computing... That’s now within the reach of young kids. So one of the things we're pushing in App Inventor is making that stuff accessible. Even the newer phones right now can do serious machine learning computations. So the question is how do you make that power available to kids and what are the great projects that they can do? So that's one of the things we're pushing. And of course, the other thing we're doing is integrating the apps and physical devices through Bluetooth low energy with the Internet of Things. So it's not just things on the phone. You actually can control devices and all kinds of integration with the same kinds of infrastructure that Amazon, Google Home, and all these places are doing. The other thing that I’m very interested in is the integration of speech and conversational AI. Because again, if you look at a 10-year-old student today, what's their experience of computers going to be like when they are 13 years old? Three years from now and the answer is that a tremendous amount of it is going to be through vision and speech. So when you think about what kind of programming would you want this 10-year-old student to become familiar with? How do you do speech understanding and speech generation? Most of the computer science courses for 10-year-olds don't do that yet. But you know yourself, that's going to change enormously over the next couple of years. So we're moving App Inventor in that direction, too.

ST: Yes, I remember that same thing. When three years ago. I was 13 years old. I was deeply fascinated by Iron Man and my dream was to build Jarvis, and it is still one of the key reasons for my app developments with google assistant. It was something I could talk to in a human manner.

HA: That is so great! Absolutely. So you figure that's going to be everyone's experience, not just yours. So whether it's Google Assistant or Alexa or something else, that's what you ask people how they think about computers three years from now, it's going to be something you talk too.

ST: I believe that kids are very very good at it because most of the time they're reading comics or watching animations, which have these amazing demonstrations of Highly Advanced Computers that are not quite possible today but would be in the near future. So, these characters help them to build a perception of what computers would look like in 5 or 10 years from now.

HA: Absolutely, and it was something which we didn’t have in our childhood.

ST: That brings us to the last question of this interview

ST: What are your views on our Initiative, Techvik, and how do you think educating masses and providing youth with a platform could bring a change?

HA: Well, I think what you're doing is really nice. I mean, I read some of the articles on your website, your website is beautiful. I love the fact that you're opening it to other people to write articles and share experiences. So certainly, my advice is to keep that up. Think more about how the various people you work with can start interacting with each other across the world and get involved with other groups or startups, like yours, that are doing similar things, yeah that’s my advice, it’s a very wonderful thing that you have done and best of luck.

For the second part of your question, people have been talking about theories of education for years and one of the things to realize is every time there's been new technology, people have said this is going to revolutionize education. Thomas Edison strongly believed that a major impact of motion pictures would be in education be education. having the ability to have famous people lecturing and making films of them was a thing that was going to totally revolutionize education. Then people have said that about the telephone, and every other technology. And certainly about computers. And, we're in the middle of saying there's this enormous impact today. But it's all kind of the same story. So I think that could lead us to be a little bit more skeptical about how hard it is to make an educational change. I mean, we can try, and there are great opportunities right now. I think about the college level things like OpenCourseware that you're talking about. They are very very significant. The other thing that's possible is using the fact that you and I can talk across the world. And we can have people all around the world that are discussing with each other and teaching others. But again, education is very complicated, it's embedded in the culture about how people and children think about each other. I'm optimistic, but at the same time, there’s a long history of overpromising about technology changing education.

ST: Sir, any final words for our reader?

HA: I think the main thing for you is to inspire your readers to get involved with what they can to actually do and to work with each other. Encourage people to work with younger people, see yourself as part of the education, but also set up nice examples that inspire other people too. I think that again the big issue with computing right now is that you can do things. Right, not just learn about them. You can actually make things and do things. So the more you can find ways to do that. You know, whether it's contests or hackathons or just organizing groups to do stuff. I think that'd be the thing that might just help you make a major impact.

ST: Thank you so much, sir, for spending your time and accepting our invitation for this talk.

HA: It was a pleasure talking to you and helping your initiative. The main reason I replied to your mail, I normally don’t, was because I looked at your website and I was very impressed with the quality of things you are doing. So keep on with that, it was an amazing experience for me to get to know you.

ST: Thank you so much once again, sir, and I will end this interview with a picture of ours.

HA: Sure, I would love that.

The Screenshot

Thank You, Professor Hal Abelson.