Tracy Chou
Tracy is a software engineer and a diversity advocate. She last worked at Pinterest and Quora, and she is a founding member of Project Include.
I got in touch with Tracy after following her push to hold companies accountable for their lack of diversity, and she talked to me about the way that VC-funded companies impact the communities in which they operate in unexpected ways, her opinion that everyone should have a humanities education, and the present and future of govtech and civic tech projects.
JACKIE
What do you think of the idea that technological progress is always good?
TRACY
Have you read Sapiens?
JACKIE
No, I haven't!
TRACY
There’s one part where the author talks about the Agricultural Revolution that’s a fascinating re-frame of the whole thing. Most people think that the Agricultural Revolution was a wonderful advance of human society—we could support much larger populations, specialization of work, not everyone had to be out hunting and gathering all the time. But the other way to look at it is that we were then tethered to the land. We were stuck with the back-breaking work of tilling the soil, planting, harvesting. And we had less diversity in our diets, which came to be built around a few staple crops, and less healthy exercise from just going out and exploring. In a way, we were domesticated by these plants that we thought we domesticated.
So there’s the question—are our lifestyles better because of the Agricultural Revolution? I think people would still argue yes, but it's interesting to have that assumption examined.
JACKIE
It's interesting how ideas take hold and then you can't remember a time when you didn't think this way. I was reading about the Cartesian view of the mind and the body and whatever and just this assumption that everything that makes you you is your mind, basically. So all of these people in tech who are like, "Someday people will be uploaded," and whatever, and just assuming that nothing is lost if the pattern of your mind and whatever is copied onto a server somewhere. It's super interesting.
So—I guess going off what you were saying before, about how technology isn't always a good thing. Do you think there's sort of a framework for determining what it is and what it isn't, maybe? What are the kinds of things you would look at to sort of make that determination?
TRACY
That's an interesting question, I haven’t thought about it framework-level. One way to think about it is, what is it giving, and then what is it taking away? And one interesting pattern of a lot of tech services that have come up in the last few years on the consumer side is that they're all free to use, but...
JACKIE
There's a hidden cost?
TRACY
There’s that saying: “If you're not paying, then you're the product.” You're paying in terms of your attention, and your attention is what's being monetized. You get to use Facebook, Twitter, etc., for free, but they're making money off of you because they're then selling your attention to advertisers.
JACKIE
I guess that's one thing I'm thinking about. It's very hard to tell—for instance, with Google. Obviously, lots of great services, but it's really hard to tell what exactly is being given on my part when I'm using it. Like, to which companies is my data being sold?
TRACY
Yeah. And how much of your data? Because for them to build the best ads modeling and serve you the most relevant ads—they're tracking all of your history. You're giving up a lot of privacy in exchange for better targeted ads.
Another thought on services where the economics are worth examining, because there are hidden, non-financial costs. Uber and a lot of other on-demand services are right now very heavily VC-subsidized, so as a consumer you get to benefit from really cool, convenient services that aren’t actually economically viable or sustainable on their own. But there are negative externalities and impacts on people and communities.
One popular area of on-demand services has been in laundry. Anecdotally, what I’ve heard happened in some places is people who would have used local brick-and-mortar laundromats to do laundry switched to the on-demand services, which were cheaper than they should have been because of investor money subsidies. But those services eventually folded because they weren’t sustainable — only, by that point, they’d already put the local stores out of business. And the people in those neighborhoods could then neither go to their local laundromat nor call the on-demand services anymore. There are costs to engaging with these services.
Let’s go back to the car services. I worry not just about the loss of jobs that will eventually happen—it's bad that all these people are getting recruited to become Uber and Lyft drivers, when those companies are simultaneously working on autonomous vehicles to put those people out of work—but also broader impacts to the community when investment in privately owned and operated car services results in lack of investment in public transit. Sure, it makes sense that Uber and Lyft were founded in a city like San Francisco where the public transit is so bad and cabs are terrible, but it’s concerning that that they then reduce political pressure from the well-off (those who can afford car services) to fix the public transit.
Imagine you're a wealthy person in San Francisco and always take Uber everywhere because it's the fastest way to get anwyhere—you would never take MUNI, and the joke is that MUNI is for when walking isn’t slow enough—and then bonds come up to support MUNI or public transit development. You might think, "I don't want my money to go toward that because I never use it," and that’s because you have your own rich-person alternatives. So I'm much more a fan of New York City in that regard, where it’s much more democratic with transit options, where in many cases it’s fastest to use the subway to get to where you want to go. No matter your level of wealth, it still makes sense to use public transit in many cases. And so then you're much more invested, political capital-wise, in making sure public transit is good.
There’s a reason that the public sector exists. It creates options at price points that are actually accessible to the general population. Sometimes when I'm using these car service apps, I feel guilty, like I should be supporting public transit instead, giving more ridership to public transit. I don't want to be fueling something that ultimately may be harmful to a wider swath of society.
It's also a little bit funny in a sad way to see a lot of these car service companies rediscover the idea of public transit. It's like, "Oh, we're going to have this service where there are predefined routes and people can get on and get off!" Okay, great! You rediscovered bus routes.
JACKIE
And these ridesharing services are subsidized with VC money.
TRACY
Oh, they totally are! They're losing money on every ride.
JACKIE
Like, Uber rides—they had that period when rides in San Francisco were $2.
TRACY
$2.25 to match the MUNI price, yeah.
JACKIE
It's just ridiculous.
TRACY
I worry about those interims. In the short term, it's easy to take things that are given to you without critically examining what the cost is—and, you know, actually, this pretty interesting question has come up with Uber and people who are saying, "Delete Uber, protest Uber." Do you want to give support to something that's convenient to you in the short term but really terrible in the long term? Hard to say.
JACKIE
Cool. So what do you think can maybe change that? What do you think are the conditions that are making these things happen right now? I feel like with these laundry startups, it's kind of —do you think VCs thought they would be economically viable or something, or...?
TRACY
They must have believed it at some point, right? They believed that it would become a big business, make them money back. I think there's a few things at play here. One is a very capitalist society where we're just very driven by financial returns. And that causes a lot of people to turn a blind eye to ethical questions. And, again, this comes up a lot with Uber—why do we keep funding people who are terrible human beings? Another example is Parker Conrad, who was the CEO of Zenefits, who basically told people at his company to commit fraud and then helped enable that. And he just raised $7 million to do something else. So there’s a willingness of people who have power and money to keep funding people and businesses that are ethically questionable because they think they'll get financial returns.
There are also those people who are not evil actors but not aware of the ethical considerations. Some of it's just a lack of humanities education and what we were talking about earlier—this unquestioning belief in technology being the right thing, the best thing—and never being forced to examine those assumptions or to consider what the consequences of their work might be. Or, in some cases, feeling like it's okay to do a day job that might be questionable and earning an ethics offset some other way. There was an interesting Medium article about this concept—you can't have ethics be your side hustle. Have you seen it?
JACKIE
No, I don't think so.
TRACY
You should look it up! It calls out people who work on things like Uber, or Facebook news feed, in their day jobs and then try to earn “ethics offsets” by volunteering at nonprofits on the weekends. And you can't do that! You need to make the core of what you do something that you can stand behind the values of. But yeah, I think a lot of people just don't really think about the implications of what they're doing. For example, the people who work on Facebook’s news feed were not necessarily thinking in a broader sense about what they were enabling.
I worked on the news feed for Pinterest—we built machine learning models to optimize engagement, where engagement for Pinterest was defined by repins, likes, clicks. So if we were successful, our models got people to repin and like and click more. I imagine Facebook is doing something very similar, with some engagement metrics they're trying to optimize.
My guess is that the people working on the news feed were more concerned about the immediate boundaries of their problems, working with their data, mapping it onto machine learning algorithms, tweaking parameters, trying to achieve their optimization targets, than what impact their work might have on the political environment. In their initial conversations about newsfeed they were certainly not debating the propagation of fake news or things that reaffirm people's beliefs whether or not they are true. And then unfortunately that got very large in the election cycle—people starting to be self-segregated into their own little filter bubbles of news or fake news.
Obviously real issues. But then I think without a stronger grounding in humanities and ethics and thinking about these questions, it's just very easy to pretend they don't exist.
JACKIE
Do you think humanities education, then, is a potential sort of fix?
TRACY
I think so. My experience studying engineering in school is unfortunately that I learned nothing about the humanities, and only much later after I started working did I discover topics like feminism and social justice and privilege and theories of oppression and social change and... I wasn't trying to be ignorant, but my attitude in school was, "I'm trying to get my degree, there is a certain number of classes I need to take, I’m just focused on getting through those." Now I wish I had had a more thorough grounding in humanities during the very formative educational period of college.
I've had really frustrating experiences trying to talk to some tech workers about really basic concepts, things that I feel like they should know as citizens of human society. I don't think they were malicious in any way; they just had never encountered these concepts. With one person that I talked to about social justice and criminal justice, his response was an earnest question: "Aren't they the same thing?" I was like, "Um, no, not exactly, and you should probably know the difference as a citizen, as a human being, kind of part of society." He was an Asian-American man, had grown up in America in a pretty privileged coastal area, and so had just never had to encounter the criminal justice system, really, or think about social justice.
If you're in a position of privilege... the whole definition is that you don't have to worry about it. You have the privilege of ignoring injustice. And so for a lot of these people, it's just really easy to ignore societal, systemic problems or the implications of what they're doing because they've never been confronted with these issues and have no real compunction to. So I think part of producing human beings who will be contributing members of society is making sure that the educations that we're instituting as standard include those parts around humanities and ethics.
JACKIE
What about regulation, in terms of accounting for some of the negative externalities of what companies are doing? It seems like, the better technology gets, the fewer people are required to make decisions that impact billions of people, potentially.
TRACY
The tricky thing about regulation, especially around technology, is that it's so hard to know how to regulate correctly, and the whole point of the technology sector is that it's innovating and creating new things, so it's very hard to know what should be regulated. And it's changing so fast—the pace of politics is just not going to be anywhere near that. So I think tech has to be much more self-regulating, than depending on government as regulatory actor. And there's so much that's unknown about information technology. Before Facebook became what it is, no one knew what it would become, so it was hard to say, "Oh, this is how it should be regulated."
There's new technology with CRISPR where you can edit genes—I don't know what the implications are going to be. I'm sure the government doesn't know, either, and it's really difficult to get lawmakers and people who set policy to be such deep technology experts, who simultaneously have to predict the future and what effects of different policies might be. When things have played out for a little bit longer, there's a role for government to ensure basic standards. But I think it's hard for government to be at the forefront of technology.
JACKIE
So, kind of a pivot, I guess—what's your personal relationship with technology, just in your life on a daily basis?
TRACY
I use it way too much.
JACKIE
Do you feel like that? Do you feel like it's too much and you would like to use it less?
TRACY
Yeah, for sure. It’s addictive. I'm constantly on my phone checking Twitter, even though I don't need to be checking Twitter.
I had a really amusing conversation with someone about Twitter. It was a friend who doesn't use Twitter very much, and someone else was trying to pitch him on why to use it, saying, "You can get everything in real-time!" And he said, "I don't want to get the news in real-time! I don't want to get bombarded with terrible things happening. I just check in once a day to know what terrible things have happened. Why do I need it pushed to me? It's all so negative."
A lot of consumer apps are designed to make you feel this compulsion to check for notifications. You get the little badges, the red dots. You have this many notifications, you want to go check them. You get the quick hit of, "Oh, someone liked my thing, or responded to me.” But those things don’t make you more satisfied in life. And I've actually found myself appreciating the times where I'm away from my phone and don't care to check it.
When I go to workout classes, I'm like, "Yes! My phone will be in my locker, and I’m not going to check it for the next hour because I'm working out." Or I'm with friends and I don't want to check my phone because I'm with people, in person, and I want to be fully engaged. I find myself appreciating those times much more.
But then when I'm not in those situations, I'm still stuck on my phone. It would be so much better if I spent less time on Twitter, less time on Facebook, less time on my phone generally. I’ve actually tried to reduce my time on Facebook, Twitter, Snapchat, by moving those icons away from my home screen top row into a folder, and then I keep moving the folder around—first page, last page, top row, bottom row—so I'll lose the muscle memory. I actively want to spend less time on technology, but it's just so addictive.
JACKIE
Do you think there's a fix for that? Because it seems like there's kind of this fundamental conflict between what companies are optimizing for and what's actually good for people.
TRACY
Yeah. Especially because the number one thing that companies seem to have figured out how to monetize is attention, all they want is your attention. For the companies that monetize in other ways, like Amazon, which makes money off transactions, they don’t need you glued to their apps. When more companies figure out ways to monetize that are not solely dependent on attention, then it'll be different. This comes up a lot with media, journalism—right now, it's based on attention, and that's why everything is so clickbait-y and reinforcing what you already believe. It's not about quality journalism. But no one has figured out how to monetize quality journalism to the same extent as they've monetized eyeballs. And we operate in a capitalist society, so unless you can tie these other goals to the financial motivators, then it'll be really hard to change behavior.
JACKIE
Do you think you interact differently with different platforms?
TRACY
Yeah, for sure.
JACKIE
It's interesting how—it's hard to tell how conscious it is on the part of Twitter, for instance. Twitter puts me in such a specific sort of mindset, I guess, and mode of interaction with other people on Twitter. What are some of the ways that you experience that? Not necessarily with Twitter, anything. How the design shapes how you behave on it.
TRACY
For me, Twitter and Facebook are really interesting contrasts. Twitter is very real-time; things just keep happening and I feel like I want to keep dipping in to see what's going on. When I tweet, it will get responses, likes, retweets. And things that's happened in the past just kind of flows down.
On Facebook, which has a much more deliberate news feed algorithm and things don’t move as quickly, I'm not as incentivized to post constantly. Only the “best” stuff gets surfaced, and it's actually kind of frustrating. Things I find much more important never get the news feed attention. Photos get bumped up really fast, certain types of contents do really well on Facebook, and others don't. And it's really annoying to me to feel that difference. On Twitter, it feels like everything is much more... equal. Every tweet I send out gets the same sort of prominence, just in the reverse chronological stream.
So I post a lot on Twitter and not that much on Facebook, but on Facebook I post the most important things I want to give attention to. Those things might get a few hundred likes. The things I post on Twitter rarely get more than a hundred. But I have a pretty constant stream of stuff on Twitter, like twenty things a day, while I'll reserve just a couple of things I really care to post to Facebook. Like, oh, there's one article that I will share every couple of months onto Facebook.
JACKIE
How conscious do you think Facebook is about that? Because I would think that they want people to post more, so it's interesting that their design kind of reflects more of the "big life updates once in a while" attitude.
TRACY
I think they have enough content from all the people that are using it that they want the content quality to be higher. Each person can only consume a certain amount of content per day. So even if I were to 10x my volume of updates, that doesn't mean you want to see 10x of my stuff. You’d probably still not want to see more than one post a day from me. If my posting ten things meant the best content from me was better and that improved your experience, then they want me to do that. Otherwise, I don't think they need me to be posting all the time, when all of it is just going into the void.
JACKIE
How do you feel about stories? Or... what is it called, Messenger Day, Facebook Stories?
TRACY
Oh my god. I'm not super into it. I think it's just too many things to check. I do kind of buy their argument that it's like news feed. Like, before Facebook did news feed, it was not a format that people consumed information in, but then once news feed came out, it became standard. I buy their argument that stories is just a new format of consuming content. But to me, it's just too much. I don't need any more places to check.
JACKIE
Yeah. So, kind of shifting gears again, this is super random, but I thought about this—what are your thoughts on, for instance, Peter Thiel's interest in young blood, or basically people in tech who kind of want to use tech as a means toward immortality or something close to immortality?
TRACY
I find it very strange, kind of weird. I personally don't care very much for immortality, so I don't really get why people are so into it.
For me, on a more systemic level, I care about improving quality of life for everyone, instead of quantity. So this whole focus on extension of life is not particularly meaningful to me unless we also can figure out quality of life.
And then there are these questions around the future of work, and how people are going to have incomes to sustain their lifestyles. People will live longer, but there are not necessarily jobs that are good for people at an older age (or even at prime working age). I'm a bit skeptical that we're going to be able to solve these issues.
It's kind of interesting to look at the population breakdowns in different countries. Some of them have bulges in population at older ages, and those are less sustainable because as those people move towards retirement, there aren’t enough workers to sustain the growth of the economy. And so there are certain governments that are really concerned about this and they're trying to get their citizens to have more kids.
JACKIE
Yeah, yeah, like in Japan.
TRACY
In Singapore, they were actually giving out pamphlets—maybe a joke, I don't know how serious they were, but—pamphlets on how to go on dates.
JACKIE
Oh, wow.
TRACY
They said things like, "Oh, you should brush your teeth before you go on a date! You know, if they seem interested, you can hold their hand!" It was kind of hilarious, trying to encourage people to reproduce, basically.
But yeah, a population bulge with older people but not enough younger people to support that, that’s problematic. China has issues around this because of the one-child policy. Even though they’ve stopped the one-child policy now, there are ramifications still. In a lot of cases, you’ll see a man who has to support his wife, who may not work, and his parents and her parents. Obviously there’s a bunch of gendered stuff going on, too, but all in all, it just doesn’t work very well.
JACKIE
What are some less sexy problems that you think are important and that need more attention? Because there's definitely radical life extension and there's Elon Musk trying to go Mars, but there are a lot of much more mundane problems that kind of aren't getting as much attention.
TRACY
Yeah. I mean, the sector that I'm working in now is govtech—just trying to improve basic digital services for government institutions that are intended to serve the people but are not very good at it because they have really broken technology.
I worked in the US Digital Service a couple years ago, in the federal government.
Just look at government's inability to provide basic services it intends to provide, like with Healthcare.gov, which is what inspired the creation of the US Digital Service. They had this big healthcare reform, they wanted to roll it out, and then it was almost completely derailed by the technological failures.
The process for applying for immigration—still super paper-based. Federal government spends something like $300 million a year on postage internally for shipping immigration paperwork around. That doesn't seem like it should be necessary, why are we spending so much on shipping paper among processing centers? And it's inefficient.
There's just so much that could be better about how we serve people. Taxes. The whole process of filing taxes is such a disaster in the US. In countries like the UK and Sweden, sometimes you don't even have to file your taxes—they already know how much you're getting paid and taxes were automatically deducted. I think in Sweden, they can even text you how much you get in returns or not.
There are so many ways that government could be better serving the people with better technology. There are a lot of reasons why government tech is hard—not in the technical aspects, but in the politics and the bureaucracy you have to navigate through. I think there's a big space of opportunity there.
A few sectors have been getting more attention, and while I agree the tech there could be better, I don’t think the technology is the primary obstacle to better services; for example, healthcare and education. A decent number of startups and companies in those spaces, but I think fundamentally with healthcare it’s not the technology that’s the problem, it’s the really complicated employer-based insurance system, no single payer; and with education, lots of regulation and bad education policy that is more problematic than the software.
What else? There are definitely underserved communities, the ones that can't afford to pay as much. Oftentimes that overlaps with the ones disproportionately dependent on governmental services.
If you’re an entrepreneur, it's more appealing if you can build a product for rich people that will pay you a lot of money for it, right?
JACKIE
Are you optimistic about the government adopting new technologies and getting more efficient?
TRACY
There are good people working on it now. Efforts like US Digital Service, Code for America, and then generally the socialization of this idea that tech people should do tours of duty in government.
Before USDS, and CFA, good tech people would typically not want to go anywhere near government. Stay away, don't touch with a ten-foot pole. Very different from other sectors, say law or finance. Clerking for the Supreme Court is very prestigious—and finance, working in the Treasury, like Sheryl Sandberg did—it’s prestigious to be affiliated with those institutions. But for tech people, no way were they going anywhere near the government.
USDS has done a good job of changing that perception on the federal level. Code for America has been really good, too, more on the municipal level. People are getting into the idea of serving the country. I think some of it is also accelerated by the changing political backdrop and people wanting to get more involved with the functioning of the country and doing good and giving back, so I think there's reason for optimism about more good people wanting to get involved.
I do think it will be tricky, frustrating, annoying, to make change because so much of government is very risk-averse, bureaucratic, things get ossified. It's hard to make change, but at least there’s more attention on it now.
JACKIE
So there's USDS and other sort of organizations working kind of within the government, but then there's also sort of this kind of newer wave of tech efforts that are civic-oriented, but totally unrelated to the government. For instance, YC launched their repository of projects, apps, trying to organize immigration data and things like that. What do you think of those efforts?
TRACY
I think there's reason to be encouraged by more attention on these efforts, though right now the civic tech space is still trying to figure itself out. It's generally hard for these efforts to get funding, since they’re not going to be big moneymakers. Relying on volunteer crews is also hard.
A lot of people want to start their own efforts but don't necessarily know the whole space, what's out there. There's not always been a great connect between the people who are the builders on the tech side and the people who know what the problems really are; I think there needs to be more collaboration between the different actors.
But I do think the civic tech space is interesting. I'm involved with the Arena Summit; it's something that a group of us got together to do after the election in November, convening people around the idea of running for office. We've expanded the vision to also be supporting civic entrepreneurship and community-building, so we actually do see a lot of these civic organizations as part of the Arena community: organizations like Flippable, Swing Left, Sister District, Run For Something. A couple of the YC folks, who worked on that repository you mention, were at our last summit recently, too.
JACKIE
So what is maybe one thing that you're concerned about (maybe it's an area or company) and then one thing that's exciting to you?
TRACY
I guess a couple areas of concern that slightly overlap. One's very obvious given all the work I do, around diversity in tech. A lack of diversity has implications for ethics and morality, but also just sheer lack of perspective. And I think this will play out in really interesting ways with some of the new technologies that are being developed, like AR, VR, AI/machine learning.
So one interesting analysis of current VR tech says that the way they try to simulate the virtual world is primarily geared toward men's perception biology, which is different than women’s, so many women find VR tech now to be dizzying and nausea-inducing.
JACKIE
That's so interesting.
TRACY
On the AI front, people have a tendency to believe that algorithms are neutral and code is neutral, when, really, algorithms and machine learning models are just codifying the biases of the people who write them or the data that they're fed, and without enough perspective on what that means—and some of this just goes back to our earlier conversation on humanities education and thinking about the implications of what you're doing—without the contemplating all the biases that go in, it's very easy to use algorithms, like machine learning, AI, in a way that can actually be really harmful.
There's a really interesting book called Weapons of Math Destruction that just came out, where the author talks at length about how data can be used in these very bad ways. We build opaque models that actually just perpetuate bias. For example, there's this tool that prosecutors or police were using to predict recidivism, and it essentially reiterates all the racial and socioeconomic inequality that already exists in the system, though people treat the output of this recidivism predictor as fact.
It creates a really bad cycle, and people aren’t necessarily thinking about how to create a negative feedback loop. That’s a specific term from systems engineering. The opposite is a positive feedback loop, like the screechy sound you get when a mic picks up a sound and blows it up, which reinforces whatever signal it had originally instead of adjusting based on real world and correcting for bias. We want negative feedback loops that dampen the original bias and bring the system towards an equilibrium.
JACKIE
And then maybe a thing that you're excited about?
TRACY
Right, what am I excited about? I am generally still very excited about software and making the world more efficient.
One thing that's kind of cool about more people understanding what software is, is that they also can start to see the ways that software can be beneficial.
A lot of people ask if coding is new literacy. I don't think everybody has to code professionally or even write code most of the time. I do think it's important for people to get how code works and why it's interesting and relevant and have the mental model for what code is. It’s instructions for computers. And computers are very good at executing those instructions very quickly in the same way every time.
As more people start to understand what code is and what software is and what new technological systems mean in their lives, they'll know how to interact with them better or even design them. You see this with more people being non-technical founders at tech companies, where they understand enough of what software is to have an idea for how software can solve problems. They don't need to be the ones coding, but if they understand what software is capable of doing, they can recruit people to help implement it. As more people become software-literate, they'll see more opportunities for where software can materially improve a situation.
I think this is most meaningful for the communities that are underrepresented in tech right now. A lot of what people in tech are doing is just looking at what problems they have in their own lives and solving them. But for the populations not well-represented in tech, that means there aren’t as many people thinking about how to solve their problems. I’m optimistic that this will change.