You’re widely recognized as the world’s leading campaigner for software freedom, having led the development of the gnu operating system. Today, the gnu /Linux operating system, and free (libre) software more broadly, underpin much of the internet, yet new structures have emerged that can wield a great deal of power over users. We’d like to talk to you about the present computing landscape and the political relevance of free software. But may we start by asking about your formation, as a programmer and as a thinker—how did it all begin?
I grew up in Manhattan, born in 1953. I was a behavioural problem—I couldn’t go to a public school without getting in trouble—and started working with computers at an early age. In 1969, during my last year of high school, an ibm lab let me come and use their computers. In 1970 I had a summer job there. They gave me a project to do, implementing a certain algorithm to see how well it would work. I finished that in a few weeks, so they let me spend the rest of the summer being paid to write whatever I felt like. I went to Harvard to study physics, and carried on programming there. Towards the end of my first year I started visiting computer labs to look at their manuals, to see how the computers differed. When I visited the Artificial Intelligence Lab at mit, they didn’t have much by way of a manual, because they had developed their own time-sharing system. The administrator there decided to hire me more or less straight away. So although I graduated from Harvard in 1974, I had actually been an employee at mit for three years. Harvard’s computer was a lot better to play with than ibm’s, but it didn’t have a lot of memory, whereas mit’s computer at the ai Lab had plenty. Not only that, they let me change the time-sharing system; in fact, that was my job—they hired me to work on that system. I added lots of features to lots of different programs—whatever I thought of, or people suggested to me, that seemed like a good idea, I would implement and then people would use it. And this was absolutely delightful—and gratifying to make things that people used and appreciated—so I kept working there. From that point on, I did programming using the machine at mit.
Were there any people at mit who were influential in how you learned programming?
There was Richard Greenblatt, who later started the Lisp Machine project; to some extent there was Don Eastlake. And Bill Gosper, although he was more of an inspiration in terms of hacking and math than in how to program. There were a lot of smart people there. But also I was inspired by the attitude at the mitai Lab, where the hackers said: ‘We’re not going to let the administrators tell us how to do things; we’re going to work on what they need, but we will decide how; and we won’t let them implement computer security to restrict us with.’ This was a conscious decision of the hackers who had written the time-sharing system, which they’d started a couple of years before I got there. Their attitude was, yes, the administrators could fire us, but we were not going to suck up to them. They weren’t going to stand being treated like ordinary employees. I wouldn’t have had the strength to do this on my own, but as part of a team, I learned it. We were the best, and most of us weren’t getting paid an awful lot—any of us could have got a much better-paying job someplace else if we’d wanted. We were there because we were free to improve the system and do useful things, the way we wanted to, and not be treated like people who had to obey all the time.
And this had the consent of the ai Lab’s directors?
To some extent. The Lab’s leaders, Marvin Minsky and Patrick Winston, were perfectly content with it. Minsky, I was told, didn’t like having doors locked, because he had a tendency to lose his keys. So the doors to the Lab and all the offices inside it were always open. There were no passwords for the time-sharing system. There was no file protection—literally: anybody could sit down at any console and do anything.
To what extent do you think that the collective ethos at the mit ai Lab back then—which is a striking feature of this history, and is typically cited in the origin stories of free and ‘open source’ software—was premised on the fact that you were working with a different kind of technology?