Costas Spanos: Thank you, Camille, and thank you all for being here. It is actually shaping out to be a wonderful even and I would like to thank the organizer for all the effort that went into that.
The first event that I have the honor to introduce is a Fireside chat between author Ellen Ullman and Berkeley engineering dean and WITI@UC co-founder Tsu-Jae King Liu. Our guest Ellen Ullman is a computer programmer, an
essayist on technology and culture, and an author of four books, two non-fiction, and two novels on the human side of technology.
Her most recent book, Life in Code: A Personal History of Technology, in 2007 was named by the San Francisco Chronicle among the best books of the year.
Life in Code bookends her earlier work, in 1997, where that was named Close to the Machine: Technophilia and Its Discontents, recounting life as a woman technologist amongst and almost exclusively male workforce at the start of the global digital revolution. Twenty years later, Ullman reflects on digital technology’s loss of innocence and reckons with all that has changed and so much that hasn’t.
She has earned a BA in English at Cornell University in the early 1970s, and began working professionally as a programmer in 1978. Now, my colleague, professor of ECS, and dean of engineering Tsu-Jae King Liu was named, in her recent post, this in July of past year. This is the first woman to hold that post in Berkeley. She’s among only 60 women deans of 368 engineering schools in the U.S. She co-founded WITI@UC, and she’s a very close friend of CITRIS. She has been faculty director of the Berkeley Microfabrication Laboratory for six years overseeing it’s transition into the Marvell Nanofabrication Laboratory at CITRIS. She’s a former chair of the Department of Electrical Engineering and Computer Sciences, and she has been known for innovations in semiconductor devices and technology, and member of the National Academy of Engineering, fellow of the National Academy of Inventors.
She was recently inducted Silicon Valley Engineering Hall of Fame, so she has very well recognized for her many accomplishments. Tsu-Jae has earned her BS, MS, and PhD degrees in electrical engineering at Stanford University in ’84, ’86, and ’94 respectively.
Now, it turns out that Tsu-Jae and Ellen hold something in common: Cornell University. Tsu-Jae was born in Ithaca, New York, where her parents, who have immigrated from Taiwan, were both graduate students there. And, of course, Ellen has graduated from there as well. So, thank you for being here. It’s gonna be an exciting conversation.
Tsu-Jae King Liu: Thank you very much for that warm welcome, Costas, and thank you all who are
here in person and online. And welcome, Ellen, to UC Berkeley. Delighted to have you here.
Ellen Ullman: Great experience for me.
Tsu-Jae King Liu: So, you know, I learned just this morning that you actually majored in English.
Ellen Ullman: Yes.
Tsu-Jae King Liu: And I thought it might be interesting for you to share with us how you, with your background, how you came to be interested in computer programming, and to make that actually your early career as a software engineer.
Ellen Ullman: Well, I took a path that I think, sadly, is not available anymore, because the first generation of programmers I worked with were- came from- a bunch of crazy people. And, smart, crazy people. Former Sufi dancers, all but dissertation in Greek, anthropologist from Harvard. These are people who just flock to it and learn to love it. I went to college at Ithaca. I went to Cornell. Beautiful school. Loved having the education I did. I valued it. From there, I got involved in video. There was a group called The Ithaca Video Project and I learned to use the Portapaks and the cameras and editing decks. We did local community things.
We were on a forerunner for cable, community antenna. We did pieces for the New York State Assembly and various political activity and artistic attempt. And then I learned I loved crawling around on the floor with cables, and operating machines, and I sort of graduated from there, left Ithaca. It’s very important to get out of the college town, especially in a small town like that. People who linger in Ithaca long after they graduate are a strange group.
Tsu-Jae King Liu: Okay. Good to know.
Ellen Ullman: It’s like they haven’t found their place yet. So, I left. I went to San Francisco. And one day I was walking down Market Street, and in the window of a Radio Shack- I miss Radio Shack, you could by all kinds of cables and connectors. You could build a heath kit.
Tsu-Jae King Liu: I mean, there were those days.
Ellen Ullman: I saw TRS-80 in the window. Affectionately known as a Trash-80. And, no reason, I just bought it because, well, it was like, “Is this video? Can you make art with this? Can you do things that are socially interesting? Or just challenging? Y’know, fun.” So, I bought the thing. And I went, “Oh, okay. This involves something called basic programming.” And so, I learned basic programming, and when I got my first program working. It worked. It was so exhilarating that I felt I had to keep doing this. It’s very hard work, but it was really good, hard work. It involved creativity, high tolerance for failure, which is important in programming, a drive, and the idea that you love and hate this machine. It’s a love-hate relationship.
I had to make money. I had to do something and I saw ads for computer programmers everywhere. It was the late 70s, early 80s. Business computing was exploding. And the need for programmers went far beyond what came out of electrical engineering schools. So, I answered a couple of ads, and anyone who kind of knew what a compiler was got a job.
Tsu-Jae King Liu: Wonderful.
Ellen Ullman: It went on from there. I went from programming and went to side based, dearly departed, eaten up by Oracle. Worked in. Got a look at the core of Unix. Went on from there. Went deeper and deeper from the high level stuff to the networking between client to server, which is what we were doing at that time, and then down into the machine level. I never wrote device drivers. Oh my god, people who did that went insane. Because what they say, it manufactures kind of a black box interaction. They say it works like this, and it really doesn’t and you have to improve and tweak. I guess that’s the end of the story. I went on to do it. I mean, something that started on kind of a whim and I went, you know, “Can I have fun with this?” It turned into a job that I thought was just gonna be a job. And then I was hooked, and I went on from there.
Now, as I say, I think, sadly, that path is not open anymore.
Tsu-Jae King Liu: Why do you say that?
Ellen Ullma: Well, especially for women-
Tsu-Jae King Liu: I know Trash-80s no longer exist, of course.
Ellen Ullman: They still do. They’re in museums somewhere.
Tsu-Jae King Liu: Okay.
Ellen Ullman: I also had a K Pro, that was my next machine.
Tsu-Jae King Liu: I see.
Ellen Ullman: That was called a portable. Ha.
Tsu-Jae King Liu: And, you think that maybe today it’s more difficult for people to teach themselves how to be software developers, programmers?
Ellen Ullman: I think people can teach themselves, but, at some level, to succeed you really need the degree. You need to go through the engineering. If you want to work at a place where you’re actually working on the algorithmic level where you’re tweaking servers, then I think more and more people need degrees.
I don’t know. I think, especially for women, it gives you confidence. I’m not sure. I mean, I hope that we’re getting more English majors, we’re getting more philosophy majors, history majors especially. I find how few people know the history of science, the history of computing, politics and so forth. It is my hope that more are coming.
But I see a lot of engineering students, not to put down engineering students, but the blend of the technical with the sense of if you work in the humanities, you get a sort of questioning wider view of what you’re doing. You look at it, and you go, “Well,” let’s say I read this, “What’s wrong with this? What’s right with this? Oh, this is wonderful.” So, I think it gives a person a different way to look at technology.
Tsu-Jae King Liu: Certainly your training in the social science or humanities helped you transition to become a writer.
Ellen Ullman: Right.
Tsu-Jae King Liu: Acclaimed author.
Ellen Ullman: Right.
Tsu-Jae King Liu: Which is wonderful. But, I was just curious if you- what challenges did you encounter since there were not that many women who followed your path and also since you didn’t have a college degree in computing or computer science. How did you overcome those challenges that you must have faced?
Tsu-Jae King Liu: Well, I don’t know. Sometimes I’m surprised at what I did. I went to this small company and, “Well, you don’t have a degree.” And offered me this very, very small salary. And I said, “Okay. I’ll take that for three months, and if you want to keep me, this is how much you have to pay me.” Now, I don’t know where I got that from, but I think this is what everyone needs to do. Y’know, “I’ll start with that. Six months from now, you’re gonna pay me so much.” And it’s sort of like you have a lease. It says this step up every year. I don’t know, and he hired me. And I stayed.
Tsu-Jae King Liu: And you were good at it too, I’m sure.
Ellen Ullman: Yeah, it turned out I was. And also I go into situations where I don’t know a damn thing, if I may say. And I looked at this system, and suddenly I am taking it apart and people call in when it’s broken. I always talk in tangents, so I’m gonna go off on one, and hopefully you’ll guide me back.
Tsu-Jae King Liu: Alright, I’ll try.
Ellen Ullman: I think there are fields that are so devalued that it’s very important that they be valued. For instance, customer service. In technical situation. We were dealing with technical users on the other side, specifically. If you see where a system fails, you learned a great deal about how a system should be created. You really see how it’s not working.
And testers, the whole, they’re lower than developers. They are the people, again, who see where it doesn’t work, who go back to the programmer and say, “You have a bug.” “Nah, that’s the way it’s supposed to work. There’s no bug.” And there’s great resistance. I know it’s true. You get these bug reports, you go like, “No. I can’t reproduce this bug. It’s nonexistent.” So, the testers are the ones who come back and show you it does exist.
So, I just wanna be a advocate for the levels at which people work. From the customer experience, as it’s called now, to testing, early developing, going deeper into algorithmic tweaking and devising, multiprocessing system. You can go down, and down, and down as closer to the machine as you find yourself interested.
Tsu-Jae King Liu: Interesting. And I know from your book — this is the Life in Code that came out last year — that you think that people who are closer to the machine tend to think of themselves as better or somehow worth more than people who are farther away.
Ellen Ullman: Yes.
Tsu-Jae King Liu: Maybe interfacing with the customers and actually, who might know how to design this system better. Where do you think that bias stems from?
Ellen Ullman: First of all, if you are working on something that normal human beings interact with, that’s not considered a very important job, strangely. And so, the more you work only with other programmers and then machine and machine and machine. That is more highly valued, and that, in general, human interaction is not valued in the development of systems. [crosstalk 00:14:38]
Tsu-Jae King Liu: Why is that?
Ellen Ullman: I think it’s the engineering bias, excuse me.
Tsu-Jae King Liu: Really? Engineering bias. So, what can we do to counteract this bias? Because, clearly you don’t think that it’s fair.
Ellen Ullman: I just think, y’know, bring more people into it. I mean, what happens is the customer support and the front end development and website is becoming a pink ghetto.
Tsu-Jae King Liu: Pink ghetto?
Ellen Ullman: Pink ghetto, you haven’t heard that phrase? Like, typing. Pink ghetto. Secretary. Pink ghetto. In other words, it’s not like- you wear blue, you wear pink. And I fear that that’s happening, and that women are getting pushed to that end.
At Google, for instance, women who work on the website development, on the interacts with people. Even if a woman has a PhD, that is not considered a technical track in the company. If you’re not on the server side, you’re not on a technical track. So, these are great dangers. Can you bring me back? That was a very long tangent.
Tsu-Jae King Liu: Oh, sure. Okay, but I’ll leave it to members of the audience, if you’re interested to continue along that tangent, to ask a question. So, maybe tell us about your transition to becoming an author.
Ellen Ullman: Nancy Peters at City Lights.
Tsu-Jae King Liu: Okay.
Ellen Ullman: City Lights. Nancy doesn’t get the glory that she should at City Lights. She received a proposal for a book called Resisting the Virtual Life. This came out in 1995, to give you an idea of how advanced the thinking was that she was able to latch onto. And she got this proposal, and she said — I know Nancy socially — said, “Oh, yeah. This is all very well, these professors and so on, but you gotta get that Ellen Ullman to give you something. She’s a programmer.”
So, the editor called me up, Jim Brooke, one night and said, “We’d like you to give us something.” I said, “Well, what?” And he said, “Well, about being a programmer.”
So, I wrote something. And that programming life outside of time. Y’know, living outside of time. And it was published in the book and Harper’s Magazine picked it up as a reading section. Nancy came back to me, she said, “Why don’t you write us a book?” I said, “Well, what?” She said, “Well, that essay, but longer.” And, there was something wonderful in the beginner’s mind. And, they can only have a beginner’s mind once. And that’s how Close to the Machine was written. I absolutely just went at it. I wrote it at various times. I got stuck, I went to Barcelona. Brought a computer into a café and people stared at me. They couldn’t believe what that was. 1996-7?
And I worked sort of everywhere. Hotel rooms, airports, cafés. And I just didn’t know what I was doing, but did it. And got stuck at various places. It had an ending, Nancy said, “No, can’t stop there. It has to end with you, not that guy.” Really, she said, “He leaves the thing, you can’t end with him leaving. It has to be about you leaving.”
That’s how I came to do it. And it took off in a really weird way. It got a lot of notice, and City Lights had to keep reprinting it. And a lot of people know it.
Tsu-Jae King Liu: That’s great.
Ellen Ullman: I was talking to Fred Turner at a City Lights gathering. He teaches at Stanford. And he teaches Close to the Machine. He says, “Somehow it hasn’t aged the way other books age.” And I’m just like, “I did that?”
Tsu-Jae King Liu: So, the culture within this field of computing or computer science has not changed?
Ellen Ullman: It’s certainly different. I think that you here are the ones who have to tell me. I wrote something. I’ve been writing things for 22 years, yikes. And I think it’s time for the next generation to begin telling me, and you, and the world what it’s like now. To find out ways to articulate it in a fashion that non-technical people will actually read, other technical people will read and recognize themselves. I mean, one of the things that happened to me that I wrote, there was a piece in here called The Dumbing Down of Programming, and I got hundreds of emails from programmers who said, “Thank you for describing my life. I couldn’t have talked about it.”
So, there needs to be way that the next generation- yoo-hoo, hello – begin your work to articulate it. Y’know, take an English course.
Tsu-Jae King Liu: I think English course are required here.
Ellen Ullman: They are! Yay!
Tsu-Jae King Liu: Yes, I think so.
Ellen Ullman: Go English!
Tsu-Jae King Liu: And our chancellor’s a professor of English, so it’s very important here at Berkeley as part of our education. So, I found it interesting that your book is a collection of essays that you’ve written starting in 1994 through 2017, so I think that’s a interesting piece of advice for the members of our audience. You can write your thoughts and collect them over time and, who knows, someday they can be compiled into a timeline. Capture the future of AI. I guess that’s the topic of the symposium today. So, maybe my final question to you would be: I know you have some concerns about the future of AI. There’s a lot of good, of course, that can come out of AI, but also, there could be some dangers. And maybe if you can share your concerns with our audience before we open it up to questions, that’d be wonderful.
Ellen Ullman: My main concern is machine learning. The original programmers write something, engineers write something, design something, and then code writes code, writes code, writes code, writes code. I mean, it turns over the decision making to the system. However that has been configured, who knows. And data sources, experiences that are surely biased, wrong. I mean, we have data scientists here.
And so, what decisions are being made in nanoseconds? I’ve asked people who do machine learning, and I say, “Can you tell me the intermediate decision making that’s going on?” And, “Well, not at this time.” Nobody can do this because it’s all going by so fast. And so, are those good decisions? Are those interesting decisions? Are they biased decisions? Are they working on a piece of data that is so limited that you really can’t use it as representative?
Also, I think you can’t see the fortunate mistakes. Mistakes are very important learning tools. Post-It notes, failure. Guy made a glue that wouldn’t stick very well. Just see the steps it’s making, and so we don’t know. I mean, what is the input? Are women’s experience in here? Is it mixed with men and women, people of other color, nationalities? There was a doctor who found AI really, really helpful to diagnose medicine in a certain kind of cancer, then she came out and said, “Well, I don’t know if that’s going to work on another kind of cancer.” So, that data set cannot look outside of itself. So, it’s naturally limited. These are my concerns.
Tsu-Jae King Liu: Yeah. So, actually, not only here at Berkeley, but at other universities, I know this emerging area of explainable AI is gaining, I guess, traction and momentum, because people are concerned that if humans cannot explain how an AI system arrived at the optimal solution or the conclusion then there might be some dangers inherent in that kind of system.
Ellen Ullman: I’m extremely happy to hear this.
Tsu-Jae King Liu: Yeah. So, I thought maybe it would be nice to hear what questions the audience here might have for Ellen. It’s quite a unique experience that she’s had in her career. There’s a question in the back. Please wait for the microphone. Please introduce yourself and, if you’re a student, maybe tell us if you’re a English major or an engineering major. I don’t know if this abides here.
Audience 1 — Rain Hoover: Hi, my name is Rain Hoover, and I got my undergrad in history, you’ll be happy to know, and then found my way into computer science through a similar kind of, “This is cool” kind of path. Now, I’m working at a company called Primer. Amy Heineke will be talking later today, that’s how I’m here. My question to you Ellen is, I think the phrase you used earlier was, “I didn’t know what I was doing, but I did it.” And I’m curious what tools you use to keep yourself going when you don’t know what you’re doing to convince yourself that even if you don’t know, you can still do it.
Ellen U.: Scotch.
Rain Hoover: I’ll get some of that and put it in my desk. Thank you.
Tsu-Jae King Liu: That was an easy, quick answer. Maybe a more difficult question for Ellen. Who has the microphone? There’s another question here.
Audience 2 — Kate Omon: I guess it’s just more an observation. My name is Kate Omon. I graduated with an ECS degree back in 1989, when I guess there were more women, and I’m distressed that the numbers are down. And Ellen Ullman’s observation of the bias in engineering? I saw it even when I was a co-op job at Intel. Even the layout engineers who would actually implement the design, they were seen as a lower level, and the attitude towards test engineers and such. I saw that too, and it just kind of baffled me, the snobbery. It’s like, everyone’s job is worth doing. And it seems like the more hands off your job is, that it’s even better paid. So, and that goes down to people are just actually, physically making things in a low tech level. And doing non-tech jobs, there’s just not the respect there should be, and that distresses me. So, anyways, I’m motivated to go back into AI for the diversity and for all the problems that we observe. My goal is reentry, so.
Tsu-Jae King Liu: That’s wonderful.
Kate Omon: And, yeah. So, I guess I’ll ask smaller questions of you later, unless you had a question.
Tsu-Jae King Liu: Maybe I can speak to this, just to clarify the data. At least at UC Berkeley, I know the percentages of women who majored in electrical engineering and computer sciences, or computer science have been going down. And, only in the recent decade, have they turned around because we are doing things here to turn that around. But the pure numbers, actually, have grown. Because the number of students, as you might know, who are majoring in EECS or CS has skyrocketed. And even though the percentages might not be as high — still around 20% — not as high as in the 80s when it was more like 30%. The pure number of women graduating with degrees in computer sciences or EECS has actually grown. At Intel, being a leader in the semiconductor industry, the previous CEO did make a commitment to address the diversity issue, to close the gender pay gap. They did that last year. The atmosphere within that company specifically has improved dramatically. It’s because the CEO made it a priority. He allocated three hundred million dollars for programs initiatives to really hold managers accountable, to have programs that benefited men and women, and so on. So, it is actually making progress. Companies like Google also are making efforts. I think we’re seeing that the industry does value women and realizes that it must do things to counteract bias, and so I think the picture’s not so bleak. But still, we can do more to empower men to be allies to help women to be more resilient. Like Ellen. I don’t know if we need alcohol, necessarily.
Ellen Ullman: No, no, no, no. I mean, that was a glib answer, really. I apologize for that.
Tsu-Jae King Liu: National chemicalization. Do you have any other questions?
Kate Omon: The bias I observed wasn’t so much sexism as just whatever role you had. Different jobs had different prestige.
Tsu-Jae King Liu: Right, that’s true.
Ellen Ullman: May I add something?
Tsu-Jae King Liu: Please.
Ellen Ullman: What really heartens me now is the activism that is coming from the engineering and programming world. Organizing within the companies. Google’s strike. Saying, “I don’t want to work on deadly weapons.” Questioning what’s going on and actually organizing. So, it’s not one person- someone like me and a friend going like, “Oh my God, this is terrible.” That, they’re organizing to ask questions about what sort of work should they be doing. What direction should the companies take. And that it’s coming from inside the companies and organizing in a political and social direction is, I think, the most heartening thing I’ve heard in a very long time.
Tsu-Jae King Liu: Any last question?
Audience 3 — Terry Meade: Yeah, I do. Hi, I’m Terry Meade. I’m an Angel Investor startup advisor, and I’ve spent 20 years working in the life sciences space on IT related stuff, and I’m specifically focused on women’s health right now. So, one of the things that, in terms of the bias question, and it’s interesting coming from your response from an academic perspective, since I focus so much on early stage startups, and see the bias associated with the 92% male VCs funding the products and services of the future.
And, Ellen, your comment about the next generation — Gen X is kind of the next generation, so I see — I’m really heartened to see the number of women my age in this room, because I think we have a lot of ability to influence change rather than skipping us like the CBNBC infographic skipped us recently. So, what are you advocating, or what are you recommending that, really, at the consumer level or maybe at the investor level, to really push for innovation, to have more diversity in terms of the development in terms of these AI and ML related solutions and innovations, so that we can get more dollars behind the things that have greater inclusion. So, what are you seeing around that? Or what would you recommend for this particular multi-generational group in this particular room to push for that?
Ellen Ullman: Underlying all of this is that, whenever there’s money involved, women are not trusted with large amounts of money.
Terry Meade: Yeah. Yeah.
Ellen Ullman: In the financial industry as well. I mean, there is funding for women, but it seems that the funding goes into things that are — like for clothing and cosmetics and bras and not into harder contributions.
What do I recommend… Now, that is a hard nut to crack. To attract a group of women and men who have a enough money to take up this cause. So, it’s the money. Venture capital is to make money. Lots of money. So, that is a very tough nut to crack. To bring in women who have lots more money and can join that group. That’s a problem. Because women don’t earn as much as men.
So, the whole picture … I think you can write about it. I think you can talk about it. I think you organize with other investors. I think that’s the way to do it. To just make awareness about that, especially writing about it. Your own experiences. The experiences that you’ve seen on the ground. Particular experiences that you can offer to the world in a way that they can understand internally, emotionally, socially, politically what goes on in that world.
Tsu-Jae King Liu: Yeah. I would tend to agree. Maybe we can increase awareness of the bias? And, also, actually, awareness of the benefits of being more inclusive and highlighting the successes of women in the industry, like what they have contributed. So we just make them more visible. So, increasing awareness in different aspects, I think, can help make the difference. And also to help women prepare so that they have credentials to help support their confidence, and to network, and to connect them with people who can be supportive. A lot of success is based on who you know, not necessarily what you know, so we, really, also need to advocate for women in powerful networks as well. So, there are a lot of things we can do, I think, it has to be a multifaceted solution, in my mind.
Alright, I think, because of the limited amount of time, I’d like to close here. I’m sure Ellen would be happy to chat with people individually, but, for now, please join me in thanking her.