How do you know you’re in a filter bubble? Perhaps you’ve noticed you’re increasingly out of touch with friends on the other end of the political spectrum. Maybe you’ve lost your ability to tolerate opposing views. Whatever your reason, you’re not alone in wondering whether your world view is increasingly being shaped by the content you consume.
Filter bubbles happen when algorithms tailor your online experience. They screen out opposing information and make you less likely to see content you disagree with. On a broad scale, they contribute to misinformation and make society more polarized and divided.
Governments around the world are trying to figure out how to tackle filter bubbles and other online harms, as are large platforms like Meta, Google, and Twitter. Their committee hearings, regulatory proposals, and product updates routinely make headlines, but we hear less about the many small companies and startups that are tackling online harms in creative ways, on a problem-by-problem basis.
Photo by Mohammad Metri on Unsplash
In a previous post, I wrote about Block Party, which allows Twitter users to filter out harassment and other unwanted content from their feeds. This blog features Zameer Masjedee, co-founder of LISN. LISN is a Canadian startup that breaks down filter bubbles in the podcast space. We chat about recommendation algorithms, the value-add curators provide, and what it’s like being a startup working on online harms.
To begin, can you tell us a bit more about your work?
At LISN, we’re building an audio platform that makes it easy to access a wide range of opinions on topics that you care about, which isn’t something most platforms do. Most platforms optimize for things like engagement, creating echo chambers that reinforce people’s existing beliefs. This was highlighted really well in Tristan Harris’s documentary The Social Dilemma: people are left feeling like their perspective is the only—or only correct—perspective, and when this happens en masse, it leads to polarization. It leads to an “us versus them mentality.” I think a lot of us have experienced this firsthand when talking to people over the last few years, and I think a lot of us realize we need to address this problem, because if we continue down this path, we’re not going to end up in the best place. At LISN, we want to make it easy for people to co-locate and consume diverse perspectives. Podcasts are a very long format by design. In a single three-hour episode, you might talk about three entirely different topics, but if you break down those episodes into smaller five-minute clips, you end up with atomic media components that are a lot easier to surface in a way that’s organized for users. Down the line, LISN will likely include some AI [artificial intelligence] elements to help surface clips, but the real value is going to come from manual curation and users being able to identify curators they like and subscribe to them.
I’ve heard you say that “content curation is broken.” Can you expand on that, and how does this relate to the problem LISN is solving?
All the content you access online is typically either recommended to you or curated. Recommendations are predominately determined by algorithms, whereas curated content is hand-picked (though sometimes with the help of algorithms). On most platforms, users interact with some kind of recommendation algorithm, and there’s a reason for this. Think about the scale most platforms operate at and how diverse their content offerings are. Most platforms are way better suited to automation than a manual system with lots of overhead and bottlenecks. Unfortunately, automation generally leads to more of the same, or similar is good. If the algorithm knows I enjoy a topic, it’s going to give me more of that topic. Curation, on the other hand, is much more intentional.
Another problem is there’s so much content out there that people often default to convenience. If they know they like a certain vlogger or influencer, they’re going to consume a lot of what that person shares because looking for new content takes too much work. We’ve engaged tons of users on this. What we found was that most podcast consumers will consume about five to six shows, and a lot of them feel like that’s sufficient because it takes too much effort to identify and keep up with a new show. But what if we could offer listeners the same level of convenience by connecting them with curators who are experts in the topics they’re interested in? People would no longer be limited to the same handful of sources they typically engage with. They’d no longer have to choose between being fed more of the same content by algorithms or spending too much time finding content they like.
There’s lots of polarizing content all over the place—why did you choose to focus on podcasts specifically?
Podcasting lends itself so well to this problem because of how it’s set up: anyone with an idea can pick up a mic, record something, and find a hosting service that’s willing to publish and promote their audio file anywhere around the world. And then you have all these different applications that will subscribe to that podcast and distribute it. What this structure does is decentralize the responsibility. The opposite of this is a platform like Facebook, where users create, distribute, and consume content all in one place, meaning one sole entity is responsible for moderating the entire supply chain. In podcasting, those things are decoupled in such a way that motivates more people to share their ideas, because, regardless of what “side” of an argument they’re on, they know they’ll be able to distribute those ideas without being shut down by one person or organization in particular. I think for that reason, podcasting is really great. It’s the freest media, and as a result of that, has the most diverse set of opinions, which aligns well with the problem we’re trying to solve.
What else can we do to improve the quality of online conversations?
I think first and foremost, making the co-location of diverse perspectives convenient. But another challenge is when you come across an idea on the internet, you abstract the humanity from that idea. It’s very easy to just interact with an idea at surface level instead of trying to understand where the idea comes from and why that perspective exists. Have that same conversation in person, and your experience will be very different. You’ll come from a place of understanding because you know the person, you understand your commonalities with them, and so you’re more willing to engage with them.
So, I think the question to ask is, how can we help people better understand what aspects of an individual’s life contribute to them having certain perspectives? How can we make people less dismissive of others? I think one of the things we need to get right on this front is the balance between free speech and censorship. One of my takes on this is that we shouldn’t just ignore that certain ideas exist in the world, because there’s a lot more value in being aware of their existence and how they came to be than categorically dismissing them in a way that alienates people. We see this a lot from mainstream media, and I think it’s one of the reasons why this whole “us” versus “theme” narrative perpetuates. Instead of muting certain opinions, we should work on presenting them in a way that better conveys why they exist, why we should be aware that they exist, and how they might influence someone’s worldview.
The online harms space is heavily populated with academics, governments, and large, private sector players. What’s it like being a startup trying to make a difference in this space? Are there any unique challenges? Advantages?
Platforms in this space have to make a lot of big decisions that are hard and don’t have a golden answer. From our perspective, if we can build something that’s incrementally better than what already exists today, then that’s a win. If we can positively impact things that are meaningful to us as a company (like whether people from different schools of thought are able to communicate with and understand one another), then that’s great. One of the advantages of being a startup in this space is we don’t face the same critiques as some of the larger institutions. We also don’t have the same media or public attention, which allows us to experiment and try new things from a place of great intentions and without fear of major backlash. I don’t think Meta could just roll out an update that significantly impacts recommendations or misinformation without a ton of criticism from people on all sides of the argument. We also have the ability to build intimate relationships with core users to understand how they feel. You are able to do that a lot better as a startup than a large corporation. At the same time, you don’t have access to the resources large institutions do—some have hundreds of people working on these problems from a data science perspective, or a psychology perspective, and they have mass amounts of data to analyze.
There’s plenty of skepticism, and perhaps even frustration, about the algorithms that shape content suggestions, or dictate what’s “up next.” What’s your take? Is this something that can be avoided? Are LISN playlists algorithm driven? Curated?
I think the final answer for LISN will be a hybrid model that allows users to efficiently analyze content at scale, but then curate it more intentionally. There’s so much podcasting content out there that you can’t reasonably expect individuals to be able to find all the relevant ones. You can incorporate algorithms into your application in a way that is intentional and—for us right now—that has a lot to do with data organization. We want to know if we can use AI to help us understand what topics are being covered in a podcast episode and what the general sentiment toward that topic is. Is it positive? Is it negative? There are tools for this today that do a really good job at assessing these things. And once we have those pieces in place, we can make it easy for consumers to navigate content and give them more time to work on the curation themselves. I think that’s the winning formula. Users maintain control over finding the content they need, and they also get to determine which content will have the largest impact and promote the greatest understanding of diverse opinions. A great example of this hybrid approach is Spotify playlists. When they first launched playlists on their app, RapCaviar, which is a manually curated playlist, was one of their most popular. Spotify had tracked down an individual that was highly educated on rap music and was able to stay up to date with new releases. Because of its quality, RapCaviar quickly became one of the most popular playlists and attracted a significant number of users from competitors like Apple Music. People valued being able to defer their music choices to someone who they knew was more educated in the space—in other words, a curator.
Content distribution platforms, and social media networks are dominated by a handful of large players (think Meta, TikTok, LinkedIn, YouTube, and Spotify). Because of this, there is little incentive outside of government regulation for large players to respond to issues like privacy, disinformation, and online harms. At the same time, network effects (which dictate that a product or service’s value increases with the number of users) and switching costs (the costs a consumer faces when switching from one product or service provider to another) make it difficult for users to try out new platforms. Given these constraints, how might new players go about disrupting this market?
On your last point—how might new players go about disrupting this market—the one philosophy that always just rings true is build a really great product that people want to use. We’re obviously going to be focusing on that from a product perspective, but I think beyond that, it’s possible to take a philosophical stance on where society is going with respect to content consumption. One example that parallels super closely with what we’re doing is privacy. Apple took a stance on protecting consumer privacy in 2010, at a time where no one was really following that narrative, and now privacy is a core pillar of tech products. I think we’re in a similar place right now in terms of how conscious people are of the content they consume and how that impacts how they think and how they experience the world. People are becoming more aware of how content is surfaced to them, and if you accept that, the next logical step is that consumers are going to want solutions that don’t make assumptions on their behalf and give them more control over the content they consume. I think that’s how new players can disrupt the market: not by following the standard of making content really easy to consume, but by recognizing consumers don’t just want convenience.
What’s next for LISN? Where are things heading?
We’ve been working on LISN part time for three years, and it’s been a lot of fun because we love this space, we love the concept, and we want to see this product exist in the world. But we’re now at a point where we want to figure out if this is something the market is interested in. The first step in that is promoting what we’ve built, what we think the problem is, and why we think this solution is meaningful, and for us, that looks like a Product Hunt launch on March 31st. Product Hunt is a platform where technology companies can talk about what they’ve built to investors, others in the industry, and people who are interested in the up-and-coming technology scene. Our hope is to get a large base of new users so we can understand how people feel about the problem we’re solving and so we can begin iterating on what we’ve built. Based on the momentum we develop, we’ll be able to validate if the timing is right for the vision we have, and if so, then we’d love to raise our first round of venture capital and continue working on bringing this to life.
Zameer Masjedee: Zameer is co-founder of LISN, a platform that helps people consume more diverse opinions through podcast clips. During the day, Zameer works as a Staff Solutions Engineer at Shopify, where he helps the world’s largest brands push boundaries in what’s possible with e-commerce. Zameer lives in Vancouver, BC. In the winter, you can find him skiing Whistler. In the summer, you’ll find him overlooking the Howe Sound from the summit of a nearby hike.
