Pragmatic capacity
On burnout and metadata with Karen Su
Juggling a day job while trying to effect change at that job, as well as organizing for other causes at the same time can lead to feeling burnt out. We talked to Karen Su about long-tail content creators, burnout, and using art as therapy.
Xiaowei Wang
Could you introduce yourself?
Karen Su
I'm Karen. I work as a product manager at Spotify on a data team that transforms messy metadata into useful information for other Spotify teams. I'm based in Brooklyn, New York.
Xiaowei Wang
What brought you to the Logic School space?
Karen Su
I’ve been interested for a long time in the role technology plays in social change. I actually joined my current team, “Content Knowledge,” because I thought one of the things that technology is uniquely good at is processing a ton of information and making it useful. I joined Logic School because I wanted to find a group of people who were also interested in applying technology for social change or contributing to social change despite technology.
Xiaowei Wang
I like the two dichotomies that you set up. Do you feel like you landed on one side of that after Logic School?
Karen Su
Well, I can briefly describe my Logic School experience, which I think is best summarized with the words “burnout” and “care.” When Logic School started, I was on the move with my family, hopping between Airbnbs and trying to spend quality time together. But I was still working and feeling the effects of a year-long pandemic. I thought I’d have a lot of energy and at least 10 hours for Logic School every week. In short, I overestimated my capacity.
In the past, I might have gone down a shame spiral, beating myself up for not doing as much as I’d intended to. My biggest lesson from Logic School was the importance of care — that the way to do more is not to be harsher on yourself or to set even higher expectations, but to be realistic about where you are, and to apply care to yourself and to others exactly as you are today. To me, this idea of unconditional care is the common theme across all our readings, speakers, and general interactions. My final project was very therapeutic. I did a bunch of abstract color paintings to capture the feeling of returning to the light from a really dark place. I want to remember how connection helped me get out, so I can extend care to myself and others in the future, if we find ourselves in a similar place.
Back to your question: I set up a false dichotomy. I think it can be helpful for thinking about things but it’s definitely not black or white. With the right intentions and resources, technology can amplify social change. I think of tech as a powerful tool, and I think of values and ethics as guides in how we use that tool. Through Logic School, through learning about the work of so many activists, it was clear to me that tech can be guided by both business interests and values of care. One of my goals was to use art to capture some of these values, to create a reminder and guide for how I want to live and what kind of art and tech I want to create. I think art can influence technology, and technology can amplify art.
Xiaowei Wang
Can you go deeper on your intention with your project?
Karen Su
The primary intention was to remember and convey care. I wanted to capture this moment of feeling overwhelmed and working through it. Nowadays it feels like we have the world’s resources at our fingertips so, in theory, we could learn and do anything. In practice, my execution never lives up to my vision. In the past, I thought this gap was caused by a lack of intention, discipline, and ability — it was not very healthy, and it didn’t help me do more.
As a concrete example, I remember a somewhat traumatic learning experience where students had endless resources, open-ended prompts, and high expectations. Some friends thrived. Others, like myself, got into our own heads. I remember talking to a friend before a mid-semester presentation of our final project. We had been working mostly independently and this review was the first time people would see our progress. As we talked, this usually sunny, carefree guy started to hyperventilate. I tried to help but was not in a much better state myself. In retrospect, I think we both turned inward in shame. But it could have been a powerful moment of connection.
In contrast, Logic School really prioritized care and connection. For me, this has made all the difference. In the first week of Logic School, we formalized a communal agreement on how to treat each other and ourselves. I saw people practicing respect, openness, and generosity. I heard people speaking openly about their capacity (or lack thereof). I felt that everyone was accepted and welcomed exactly as they are. So for the project, I really wanted to capture and remember this care. I also think this is a great guiding theme for any project, especially in tech.
Xiaowei Wang
How has care continued for you, in your workplace or more broadly?
Karen Su
I’ve been more focused on relationships with my immediate surroundings — my neighbors, teammates, friends, and family. It’s helpful to have embodied interactions after a year of mostly digital experiences. One guiding principle I’ve learned is to apply care to myself the same way I would to a friend. Then I start from that unit of one to extend little by little. When it feels like I’ve overcommitted, I try to pull back to refocus on the core. I want to focus on the smaller spheres within my control and immediate responsibility instead of using abstract lofty goals as an excuse to neglect the very real people in my life.
In the workplace, I’m practicing care in two ways. First, caring for my colleagues as people. Trying to understand their needs and motivations, and helping them self-actualize. Second, thinking about the people who may be impacted by our work, especially if they’re not a primary user we serve. I think a foundation of mutual trust and respect, plus a culture where care is baked into every product decision, can be more effective than any reactive top-down policy.
Xiaowei Wang
I remember, when we first started talking, we talked a little bit about your work at Spotify and the role of platforms in moderation. It was interesting because there are some folks in the Logic School community who are trust and safety engineers working in that space. How do you feel now about big corporate platforms and the ways they moderate content?
Karen Su
I feel more pragmatic about it now. I think a couple of things have changed. One thing is recognizing that we have limited capacity and that we need to prioritize. Of course, prioritization is a subjective exercise. I’ll just speak to how things are today.
Spotify has over 7 million creators in our system. But less than 1% of them take up the majority of streams. Think of Ariana Grande or Drake, who have millions of monthly listeners. Then think of me or somebody uploading music from their bedrooms. The vast majority of creators on Spotify have very few listens. They’re more like me than Drake.
Before joining Spotify, I used to think of platforms in the abstract. Like, ‘Of course it can be an equalizing force, since every artist will now have equal opportunities to distribute their work.’ I think that’s still true to some extent. But I’ve also learned that platform products, like all technology, are tools that amplify existing dynamics.
Think about the pre-streaming days. If two creators had the exact same song but only one was backed by a label that had hundreds of people working on marketing and promotion, which would have their music heard more? Probably the one with hundreds of people working to promote them. A radio station host might be more inclined to play that song, so lots more people hear of that artist. From a platform perspective, you see two songs, one immediately has a lot of listeners, and one has none. You’re going to take that as a signal that more people might like the first song. You don’t know whether that popularity was caused by a marketing team or organic popularity. But given that signal, you might be more likely to recommend the first song. Just like a radio station host, our recommendations algorithms can amplify a small initial difference into a large eventual difference.
Here’s a more concrete example. Many teams at Spotify do behind-the-scenes work to keep things functional. My team, for example, deals with the problem where albums are placed on the wrong artist page because they were delivered without any unique identifiers. Like I’ll try to upload music under the name “Karen” and it will go on some other Karen’s page. When this happens to a member of BTS, hundreds of their fans will organize to bombard the Spotify social media accounts to get it fixed quickly. When this happens to an artist that no one listens to, it might go unnoticed for a long time. Our long-term goal is to eliminate this problem for everyone. But we’d start with the BTS use case because it’s an easier problem and more people care. We wouldn’t want to leave hundreds of angry BTS fans hanging while we go look for potential problems that no one has reported.
This is not exactly translatable to the content moderation space. But given the reality of finite resources and seemingly infinite problems, I’m guessing those teams might share similar approaches around prioritization. I think we’re in the early stages of platforms taking social responsibility. I still think there's a lot of potential. I think two years at Spotify have given me a reality check on how long change can take and how many people it’ll take. I'm in burnout mode right now, so I am focused first on capacity and pragmatism.
Xiaowei Wang
I think it's always inspiring when people are real about their capacity and where they're at.
Karen Su
Yeah, I’m trying to rest up so I can keep going later. Just trying to be realistic is the biggest thing I'm still learning.
Xiaowei Wang
Here’s your TED Talk moment — what would you like to invite readers to join you in on?
Karen Su
There’s a MLK quote I really like: “Power without love is reckless and abusive, and love without power is sentimental and anemic.” I’d invite readers to reflect on how they balance the two.
Technology is no doubt powerful. If we’re in any position to wield this power, we have a responsibility to be careful. We can’t always predict the nth order effects, but we can listen to—and try to proactively seek—feedback from people we might impact.
On the other extreme, don’t paralyze yourself with unrealistic expectations like I did. Then you’re really not doing any good, especially for yourself and your loved ones. When your love is overflowing but your actions don’t measure up, take a pause and focus on impact over intention.
Good luck out there!
Karen Su (she/her) is a third-culture kid who grew up in Beijing, Hong Kong, and Boston. She is an artist and product manager currently based in Brooklyn.
Find Karen at https://www.instagram.com/karkesu/