Hinge CEO Justin McLeod says dating AI chatbots is 'playing with fire'

23:05Originally, it was very centrally run. There was a lot of very tight coordination, a lot of direction coming directly from me and my executive team. And then as we started to go beyond 100 people, and I would say a lot of the technology was relatively stable, like social, mobile, big data, how do we keep optimizing and iterating around this? And we became a pretty decentralized organization where we had principles around pushing decision-making down to the lowest levels possible, keeping it really on the front lines. We had pretty independent cross-functional product teams that would work on their individual little missions or surfaces.

23:50We kind of oscillated back and forth between that. People felt like autonomous, like feeling like they had a lot of autonomy and a lot of was kind of the main ethos of the company. And then I think with AI over the last couple of years and feeling like, whoa, like we really need to make a pretty big shift. Like I said, risk of disruption is high, very big opportunities to shift the product experience in a new direction. It requires now pulling decision-making back in towards the center a bit and giving a lot clearer strategic direction to the team so that we're all working together in concert towards one thing together because the whole app really has to move together.

24:32Different parts of the app have to talk to each other in ways that when we weren't going through much change, it wasn't as essential. That said, we still have very highly cross-functional product teams where product managers sit with a dedicated designer, researcher, data scientist, tech lead to attack very mission-oriented problems. And when you said surfaces and missions, is that expressed as just the tabs at the bottom of the app? Is that how it's broken down, or is it actual user journey missions? That's what I mean, surfaces versus missions. I think we've sort of gone in different directions.

25:12There's never really a clear line, one versus the other. Do you own the discover tab where you discover new people, or is your job to help people find the right person? And then you have to think sort of more cohesively about operating across different surfaces or parts of the app. And now we think about our teams operating less as individual surface units and more as part of a cohesive dating outcomes team where people feel a bit more flexible moving around to different surfaces. How do you think about assigning product managers to those teams? Because product managers, at least in my experience, they're like, I own this square, and I will mess with this square to make this number go up as much as I can.

25:52but I can mess with all the squares. It's really hard, right? Yes. That's why we have strong kind of director VP levels who oversee an overall mission, like a head of dating outcomes or a head of growth who's coordinating a set of PMs. And again, we ask our PMs, their primary identity is as a dating outcomes PM, not as a discover PM or a profile PM or something like that. And while day-to-day, most of their work may focus on the profile and identity work, they see themselves as very much operating as part of this team. That feels like something you evolved to. You're a relatively young founder.

26:35I think you found it right out of Harvard Business School. How has your decision-making framework evolved? How do you make decisions? That's also hugely evolved. I mean, everything I've learned, I've learned through doing all the wrong ways first and then eventually getting to the right way. When you're a founder and you have a small team of 10, 20, 30 people, you're kind of just making decisions by the seat of your pants. Like what feels good, what feels right. You're just using your own gut. As we started to get towards 100 people or so, what I noticed was I would be making different decisions on different days that weren't always consistent.

27:06And it was kind of based on my mood that day or whatever data was in front of me or what I just last read or whatever. And I was just finding that I was getting pretty inconsistent. And so what I started trying to do is write down my whole management algorithm. So I started putting it in a Google doc, like, here's how I make decisions. Here's what I believe is true. I started publishing that to the whole company so everyone could just read it. So we would just be very transparent about how we made decisions. It was around that time that I think I read Ray Dalio's book, Principles. And the idea of just like getting super dialed on, like, how do we make decisions?

27:41What are principles? What do we believe is true? And then I opened it up for everyone to just comment on. So we would have long debates in the margins of comments, you know, comment margins of a Google Doc form to think about everything from our like product strategy principles to whether Hinge should have like a dress code, like literally like anything, like it was all just there so that everyone could debate it. And we had hundreds of principles. And then as the company got even bigger and we got to like 300, 400 people. It's very hard for everyone to, one, you just can't have like these endless debates in Google and like Google docs anymore.

28:17And also what happened was the principle started to stabilize. There wasn't as much debate and churn anymore. And then it actually became an exercise in distilling down like what is the most essential things to communicate about what our culture is. And then I worked a couple of years ago to actually write an internal book called How We Do Things, which distilled down to four or five principles. What are the most fundamental things to understand about how we make decisions here? And then individual teams and individual projects will then write their own principles that are more specific to what they're doing at any given time.

28:58But we do have, and one of our meta principles of our four now, is decide with principles, meaning that we don't want decisions getting made based on some random person's opinion that if tomorrow this person leaves the company, we rehire someone else, they're going to come with a completely different set of ideas about how to do something. We really try to maintain that we define our principles first, we agree on those, and then we see how our work maps to those principles. And I'm happy to talk about what the other three principles are if you want, but that's the kind of framework that we use to make decisions now.

29:32Yeah, talk on them a little bit. There's a beautiful website. We'll link to it. It says scrollytelling. It's well done. But tell people what the other three principles are. The next one is love the problem, meaning that what I would notice is we would get an inclination around a user problem, maybe not even validated 100%. Then we'd start getting feature ideas, and we'd be very attached to a feature. And sometimes the feature would even drift and not even be solving the original problem that we were trying to solve. And what I found is that if you spend much more time with the problem, really understanding what's the why behind the why behind the why, what's really going on here, can we really get deep into our users, into the data, into our user experiences?

30:19And if you want to build breakthrough innovation product features and products, you have to go to that level to get insight that just isn't available at the surface and then stay really committed to that problem. And that's what, again, allows innovation. I think a lot of Silicon Valley and other people I notice will kind of, the strategy is just to throw feature ideas against a wall and kind of see what sticks and let's see if this works. Oh, it doesn't work. Throw that out. Let's try something else. When you have a lot of deep conviction around a user problem and you really know you want to solve it, then you have the resilience to try and try and try again to solve that problem, even if your first or second iteration doesn't make it.

31:00And so one of the most foundational is love the problem. Fall in love with the problem, don't fall in love with the solution and be willing to give up the solution at any given time if it's not solving the core problem that we're trying to solve. So that is love the problem. Next is keep it simple, meaning that I think the best solutions are always the most elegant solutions that overall we want to keep the product very simplistic and minimal. Our colors are black and white. If you look at the Hinge product today, it's very clean. It's very simple. We're always stripping away features that don't make sense.

31:35And there's just recognizing that there's complexity. There's cost of complexity every time you add a feature. So even if you add a feature, if it's only marginally beneficial, the cost of the complexity and maintenance of maintaining that feature versus the marginal benefit it adds will end up gunking up the app over time and slowing you down over time, which is a hard conversation to have with product managers because they'll work for months on a feature and they'll ship it. And they'll be like, yeah, and it didn't harm the user base. And we like it. And it even moved this metric over here by 2%.

32:06And you're like, well, the cost of complexity is high. And so we need to focus on things that are actually going to have major impact. Are you all the way at two features out for every feature in? Do you think about it that way? I know some founders do. I haven't heard that before. I don't necessarily think that way. But I do just think in constantly reevaluating like what's in the app, what needs to stay and having a high bar for when you build a new feature. Does it actually accomplish what we need it to accomplish? Is it worth the complexity cost? So that's the third one now. And then the fourth, the last one is tend to trust.

32:43I just find that like trust is the lifeblood of an organization. you have to do a lot of work to proactively cultivate and tend trust by creating strong interpersonal relationships, by creating lots of opportunity for transparency at the organization. Like we have always been very, very transparent about where the organization is. So much so, in fact, that we had to make all Hinge employees match group insiders so they couldn't trade match group stock except during trading windows because we would be so transparent about where we were, what our financial position was. Everyone should know that all the way down to any position at the company.

33:21So that trust building that you create both interpersonally and from leadership down to the rest of the organization I think is absolutely essential and just saves you a lot of headache when it comes to internal politics and all those types of things. Let's put this into practice. You obviously made a big decision to refocus on AI. How did that come about? Did you wake up one day and say, oh boy, it's happening. Match Group put out a press release with OpenAI saying, we're going to work together. Did you read that and say, I got to figure this out? How did this come about? Certainly the release of whatever that version of ChatGPT that sent shockwave through the world was a pretty big wake-up call.

34:00Obviously, we'd already been using machine learning and things like that, and trust and safety, and our recommendation algorithms. But I think it was the sort of shot across the bow that came from the release of the of chat gpt's whatever it was 3.0 that really woke us up to the to the potential capabilities here and realizing that like this could be a major disruptive force in a way that we hadn't seen really since we started hinge it took a bit for us to get our strategy clear about like what our thesis is on how this is going to affect matching and dating in the future uh it wasn't immediately apparent but i think we have a pretty clear thesis now.

34:40We started to organize the company around that thesis. What's the thesis? That there's two main vectors that AI is going to impact dating and matchmaking. And it's essentially, I think the big story is it's going to move much closer to the experience of working with a personal matchmaking service and much less an experience of feeling like you are joining a social platform on your own, kind of trying to find your person. So what does that mean? two big pieces. One is personalized matching and the other is effective coaching. So the personalized matching front is we should be able to move much further beyond the world we are in today, which is our users speaking to us in essentially Morse code with their likes and passes, trying to communicate to us what they like and what they don't like.

35:29The idea that they would be able to speak much more directly to us about like, here's what I'm looking for. Here are my values, here's my personality, here are my interests, us being able to listen to them and their preferences, and even integrate things like relationship science to understand what types of people are compatible and what types of people are not long-term compatible, and introduce you to a much more curated, higher quality, less quantity list of people where you have much more trust that if you're introducing me to this person, then this is probably something someone that I want to go out with.

36:01We've already seen big gains, by the way, just by using the power of LLMs to drive more of our recommendation systems using the data we already have. But we just released a new algorithm a couple months ago that increased matches and dates by like 15%. That's just using the same data. But now we can start to use much more of that unstructured nuanced data, people talking to us in their own voices about who they are and what they want, that we can use very effectively. So that's the whole personalized matching front. And then there's the effective coaching front, which is a lot of our users struggle to get out on that first date and they often don't know why.

36:41And I've got, you know, even I have friends that are incredible people and they'll ask me to take a look at their Hinge profile and I'm like flabbergasted. Like this is their attempt at like putting themselves out there. So we're starting with pretty basic things. you know, Hinge has these prompts, which are short questions designed to get you in a conversation. You put them on your profile and a lot of people write great responses to prompts, but a whole lot of people write not so great, often just like one word response to these prompts that don't work. And we found it just incredibly effective to have trained a model on good prompt responses and give people feedback to ask them to, it's mostly like, can you say more about that?

37:22here's yeah don't just put no in this field yeah and be a little bit more specific and tell a little bit of a story and find and help you know good answers kind of invite another question back or get a conversation going and so we can give people those nudges so they write good prompts so that they choose good photos if we notice you know we have a team called Hinge Labs which is always looking at why some people succeed why some people don't on the app some of it is again building product features that help solve those problems but part of it is just giving guidance and notes to people about how they could be using the product better.

37:57And we traditionally have published those in date reports. We publish them to the press. We put them in the help center. But people just, for the most part, don't read those. But the idea that we could take this body of knowledge we have about how to succeed on Hinge and then look at how our users are using Hinge and then deliver the right piece of advice at the right time to the right user, I think is going to be pretty transformative for a lot of people. There's a pretty fine line between that and what I see lots of people doing all day long already, which is just talking at ChatGPT. Just hanging out.

38:27We had the CEO of Replica on the show, Eugenia. She was like, my plan is people are going to date AI bots. That will coach them up into being fully formed people and then we'll release them into the dating pool and they will have confidence and self-assuredness. Again, a fine line between we're prompting them and coaching them inside of Hinge, and we're coaching them a different way in a more self-contained ecosystem. How do you think about that? Would you launch a full-on virtual girlfriend inside of Hinge? Certainly not. I have lots of thoughts about this. I think there's actually quite a clear line between providing a tool that helps people get better at something or do something, that, and then there's a, there's a line when it becomes, this thing is trying to become your friend.

39:14It's trying to pretend like mimic, like it has emotions. It's trying to create emotional connection with you. That I think is really playing with fire. I think that we are already in a crisis of loneliness and a loneliness epidemic. And if you look at, you know, it's, it's a complex issue and it's baked into our culture and it goes back before the, you know, the internet. But if you look at just since 2000, over the last 20 years, the amount of time that people spend together in real life with their friends has dropped by 70 % young people. And it's been almost completely displaced with the time spent staring at screens.

39:55And as a result of that, we've seen massive increases in mental health issues and people's loneliness and anxiety and depression. And I think Mark Zuckerberg was just quoted about this, like most people don't have enough friends, but you know, we're going to give them AI chatbots that believe that AI chatbots can become your friend. I think that's honestly an extraordinarily reductive view of what a friendship is, is like someone being there to say all the right things to you at the right moment. The most rewarding parts of being in a friendship are being able to be there for someone else to risk and be vulnerable to like actually just share experiences with other conscious entities so i think that while it will feel good in the moment like junk food basically to experience someone who says all the right things and is available at the right time it ultimately just like junk food will make people feel over time like less healthy more drained and will displace human relationships that they should be out cultivating in the real world.

41:01How do you compete with that? Because that is the other thing that is happening. It is happening, whether it's good or bad. And you've got to hinge is, you know, you're offering a harder path. And so, you know, you say, like, we got to get people out on dates. Like, I honestly wonder, based on the younger folks I know, sometimes they're just like, I just want to leave the house. Like, I'd rather just talk to this computer, right? I have too much social pressure to just leave the house in this way. That's what Hinge is promising to do. How do you compete with that? Do you take it head on? Are you marketing that directly?

41:32I'm starting to think about it very much about taking it head on. I mean, we want to continue at Hinge to champion human relationships, real human-to-human, in-real-life relationships because I think they are an essential part of the human experience and they're essential to our mental health. It's not just because I run a dating app and, you know, obviously it's important that people continue to meet. It really is like a deep personal mission of mine. And I think it's absolutely critical that someone is out there championing this because it's always easier to race to the bottom of the brainstem and offer people junk products that maybe sell in the moment, but leave people worse off.

42:06That's the entire model that we've seen from what happened with social media. And I think AI chatbots could be, frankly, much more dangerous in that respect. So what we can do is to become more and more effective and help support people more and more to make it as easeful as possible to go out and do the harder thing and the riskier thing, which is to go out and form real relationships with real people who can let you down and may not always be there for you because it ultimately is a much more nourishing and enriching experience for people. And we can also champion and raise awareness as much as we can.

42:45I mean, that's another reason why I'm here today talking with you is because I think it's important for people to be putting the counter perspective that we don't just reflexively think AI chatbots can be your friend and not think too deeply about what that really implies and what that really means. We keep going back to junk food, but people had to start waking up to the fact that this was harmful. We had to do a lot of campaigns to educate people that drinking Coca-Cola and eating fast food is detrimental to your health over the long term. And then people, as people became more aware about that, a whole personal wellness industry started to grow up.

43:22And now that's obviously like a huge industry and people spend a lot of time focusing on their diet and nutrition and other types of, and their mental health and all these other things. And I think similarly, we need a, like social wellness needs to become a category. Thinking about not just how do I get this kind of junk social experience of social media where I get fed outrage news and celebrity gossip and all that stuff. But I actually start, how do I start building a sense of social wellness where I can build enriching, intimate connection with important people in my lives.

43:59We need to take another quick break. We'll be right back.

44:13We're back with Hinge CEO Justin McLeod. Before the break, we were discussing the role AI is now playing in dating apps and what that role might look like in the future as more people turn to chatbots for companionship. Now I wanted to ask Justin about some of the ways AI might be misused in online dating and what plans Hinge has to stop it. The connection between the wellness industry and the rise of social media is a whole other podcast and maybe a PhD thesis. There's a whole lot there to unpack. I take your point though, right, that maybe using our phones in healthier ways is the future. It will make us better and that will be a reaction to the negativity that we see from phones today.

44:49Literally as we speak, I'm sure Elon Musk and Donald Trump are continuing to tweet at each other in an unhealthy way or maybe the future of the entire planet. But you've got to use AI today, right? You have prompt feedback running in the app today. You're helping people pick better photos. The flip side of that is, well, they might just use AI to generate the content. Can you detect it if your prompt feedback says, hey, that's not a good answer, and someone runs away to Gemini or ChatGPT and comes up with a better answer that doesn't actually reflect them? I think about this kind of like extreme photo filters, which used to be popular in Instagram back in the day, which is that ultimately you are going to have to go meet up with this person on a real date.

45:31And so you want yourself to come through as best that it can, because obviously you're not going to bring ChatGPT on your date with you. I'm worried about this. I want to say that I'm worried about this. And it's not a winning strategy. That said, do people ask advice and little tweaks? They do it today. They ask their friends, like, how should I respond to this text message or whatever. So in some sense, I don't see it that differently because you will have to meet up with this person eventually and show up as the real you. That kind of feedback inside of HandChange obviously has messaging features.

46:06Are you going to add a little coach into the messaging feature to say, hey, don't be a dick? Well, that already exists. That's called Are You Sure? that's AI driven to make sure that people send inappropriate messages. But yeah, again, the right nudge at the right time, because I think if we build the right tools within Hinge that are appropriate for dating, then people will then use it less to like run out to chat GPT to use it in ways that probably are less appropriate. Nudging people to say, hey, you guys have been chatting for a bit. Did you know that most people after they exchange this many messages, usually just go on and move to a date?

46:39Or, hey, it seems like the conversation has died. Here's like something interesting that you may not have noticed on their profile that you can ask about. Little things like that, certainly. One of the interesting dynamics here is you'll add more and more AI to the digital experience people have with each other to make them perform better or act better or be more interesting, whatever it is. And then they'll go on a date, and then they might leave your platform, right? They might switch to iMessage or call each other on the phone. I don't think Gen Z is calling each other. They'll do something else.

47:07They'll go on Discord. How do you bring that experience along for the ride to say, we're going to continue to stay here and help mediate you and coach you through this relationship? We're not there yet. We still have a lot of work to do just to get people out to the first date. And at the same time, I do think there's actually a lot of opportunity to help coach people through that experience. How to show up on a first date, what to talk about on a first date, how to build intimacy over time, how to ask about the right things to determine compatibility. So I think there are definitely opportunities for that.

47:37It's not on the 2025 roadmap, but certainly something I'm thinking about. Do you worry that people are going to just upload full AI avatars on the hinge and catfish each other to death? We have a very robust trust and safety team that is thinking two, three steps ahead about how to mitigate things like that. I have spent too much time talking about watermarks and AI and Synth ID. There's lots of episodes of the show that are deep in the AI watermarks game. and it has effectively come to nothing so far. There's just a lot of problems there. Are you in it to say, okay, we can detect a full AI photo here?

48:15There's so many signals when it comes to creating a dating profile from the phone number you use and the email you use, your IP address, all these things like that, that we have a very multifaceted way of determining the authenticity of profiles. I'll say that. Running these models is costly. There's lots and lots of different kinds of models you can run at different costs. Are you using lots of models? Are you sending everything to GPT-4? How does this work for you? We use different models for different things. Sometimes we build them internally, completely. As you said, public, that we have a relationship with OpenAI.

48:46It's just we use different things and are always balancing cost and performance and ability to build in-house versus not. Do you see that trend shifting over time? I'm very curious about what the frontier models can do versus what the cheaper, more efficient models can do. Have you seen that shift over time as you've started to deploy these tools? In one, we've seen the cost of the frontier models just decline precipitously, which is pretty interesting to watch. But I'll say that there are models, even the prompt feedback model is a very, very specific and discrete model that we can essentially mostly build internally to understand.

49:21Do you run that in your cloud, in your data center? Are you running that on people's phones? I don't actually know. I think that's on our data. That's in the cloud. I'm almost positive that's in the cloud. that one. The reason I'm asking is to do any of this well, you need more and more data from people, right? And you're kind of asking them to generate more and more data. Like that's not a good answer. Tell me more about yourself is more data. And it's data that you're now storing. It's particularly data about gender, sexuality, dating preferences. That stuff, like the government suddenly has like a very unusual and somewhat threatening interest in.

49:53Are you worried about that, that the Trump administration or some future administration would show up and say, tell me all of the transgender people on your platform. Obviously, we have very, very sensitive data that we have very, very clear protections around. And we haven't seen anything like that. So you haven't had any of those incoming requests yet? No. The Trump administration has also said they're going to start scanning social media profiles for references to Palestine, for comments about Trump himself. When you talk about matching people and values, those things come up. Has there been any request for hinge profiles from DHS or ICE or any of these other parts of the Trump administration that are doing this kind of social media scanning?

50:35No. The reason I'm asking is the amount of data you might collect that is very, very personal seems like a rich target. Have you thought about the planning for how big of a target this might become as you prompt people to input more and more data with AI? Certainly. I think we'll have to handle those things as they come. We're obviously in a very uncertain time right now, but I will say that primarily we are a platform about creating intimate one-to-one connection. And that's people's very private lives where they should be able to express themselves in the way that they see fit and describe themselves and their own sexuality and their own gender and the way that they need to do that.

51:11I view that as absolutely sacred and fundamental to our mission. And people feeling safe to do that is absolutely critical. So those would be our very highest priorities. And I also imagine not top priority for relative to social media people are posting and blasting thousands, millions of people with ideas like this. Our platform is not about one-to-many posting and conversation. It's about intimate one-to-one connection and one-to-one conversations. I think I would warn you that having a data pool of that kind might make you a target. I'm curious how that plays out over time, particularly in this administration.

51:54There's some platform dynamics here as well. Like iOS and Android exist. They are platforms. They are also themselves rich targets for the government. Overall, there's a push for the platforms to do age verification themselves. There's laws now in certain states. There's laws in other countries. The Apples and the Googles of world have pushed back against this in various ways. do you think they need to do it? Do you think that it's at the iOS and Android level that need to do the age verification? Because this is a core component of bringing people onto Hinge. It's not for children. It's certainly not.

52:23Yeah, we're 18 plus, and we have our own age verification methods. But yeah, we've been pushing for these platforms to do age verification themselves because they have even more robust ways to do it. Their arguments in response, when you listen to Apple and Google push back against these laws, is that it would be too hard, They would create some sort of censorship regime that the app vendors need to be liable for this. Have you seen any movement in that dynamic? I mean this is – I think at the highest level, this is one of the big dynamics of how we might regulate platforms in the future. I'm staying much closer to the product development and the – where we're going with AI right now than I am to that.

53:04Your monetization is obviously a hinge premium. I think it's$55 a month. There's another tier that's$45 a month. The big news in platform world is that Apple is no longer allowed to prevent alternative payment systems. Match Group, in particular, has been leading this fight. It's in all the press releases. Has that changed the dynamics of Hinge for you? I don't know if it changes the dynamics, but it's certainly going to, you know, having more flexibility and giving users options to be able to pay different ways I think is good for everybody, for sure. Have you launched an alternative payment service yet?

53:39we have not. Are you going to? Yes. How soon? Certainly by the end of the year. Is that going to be a match group payment service or a hinge payment service? How do you think about that? These are things that we're figuring out, but most likely hinge. That would return somewhere on the order of 15 or 30%, depending on how the billing works for you at your scale and recurring subscriptions and all that. Is that just going to be pure margin? You're just going to get the money back? Oh, I think it changes the equation on many fronts, right? Like it allows us to invest more in the company. It changes how we would price.

54:13So no, I think it could result in lower prices. It could result in more investment in the company or it could result in more margin. It's probably like some combination of all three of those. Just walk me through the, I mean, Match Group has been doing this fight for a long time and you're already describing how you might change pricing or the lifetime value of customers. The decision gets, Epic fights this fight for like five years. At the very end, the judge says, I'm very mad at you. you can't do this anymore. Did you immediately start making plans that day or were you like, this is going to get appealed, we have to wait?

54:43There's been back and forth and appeals and stays and things like that. I think just a couple of days ago, the appeal was denied. So I think that made it pretty real. And just to put the decision-making into practice, did you say that day we need an alternative payment system? Certainly the day that the original ruling came out, we started to plan. What does that planning look like? Is it, I'm going to call it Stripe? Put us in your shoes. That happens, someone comes to you, and you say, okay, we need to start to plan. Walk us through that moment. I mean, just like anything else at Hinge, I think that we stay grounded in our principles.

55:20We look at the big picture. We look at the teams and the roadmaps and the things they're focused on right now. And we think, does this new information change? and as we look at our growth team, does it make sense to build the next monetization or expansion feature? Does it make sense to pivot resources over to this thing? And given, as you said, the 15 % to 30 % gain that's on the table, it's pretty high priority. There's an ecosystem of companies that might be building this stuff more centrally, that might be charging different rates. I'm kind of excited about that. It's wonky and boring. There's a reason we're ending the episode.

55:53Yeah, a whole new industry, I think, will emerge. Well, not an industry, but certainly a suite of services will emerge around this to allow people to manage subscription payments, cancellations. It's certainly nuanced. But at the end of that, what you want is rates to come down. Where do you think the rates should be? Because I know no one has ever thought they should be at 15 % or 30%. Where do you think they should be with a little more market competition? When you stack the credit card payment processing fees on top of the fees around customer service and all the sort of nuance of managing this, but I do think it comes down to like the 5 % to 10 % range.

56:32When you think about recovering that, you know, up to 20%, are you thinking, okay, I can use this to lower prices and grow, or I have to build many, many more AI features to compete against the coming onslaught of AI chatbots? I think we're very, very focused on innovating for the future. Like I said, it changes the equation. So it's on all three fronts. It's lower prices, it's higher margin, and it's more investment in the company. But it certainly gives us major opportunities to invest in the core product experience at a time when there's massive disruption. So it's a particularly critical time to be doing that.

57:11There's a lot of talk about platform shifts. You've talked about platform shifts here. People might be using Hinge differently because they have AI tools or because the AI tools are helping them find one another more efficiently or better. A lot of the platform shift I hear about is, oh, we're going to have new devices. We're going to have new form factors. They are just going to talk to ChatGPT in the bar. We are. Maybe we'll just have agents that represent us and they'll go on dates for a while and come back and say you should go on this date with this other person we found on Hinge because the agents have fallen in love and now you just have to not screw it up.

57:44That kind of takes the screen away. That takes your surfaces and your missions and puts them in a totally different place. How are you thinking about that level of shift? Is it even on your radar? Yeah, I'm thinking about it right now. I think we overuse the form factor of our mobile devices right now for all kinds of things that it doesn't need to be used for. And I think a lot of those will be siphoned off into some other form factor, especially I think audio and voice is going to be a very big piece of it. But I don't think that means that the form factor completely goes away. There are things that you need visual cues and references for where a screen is still going to be the dominant form factor.

58:17And at least a piece of the dating equation is going to be that for sure. Do you think we'll get to a place where people's agents are just dating each other and then reporting back? No, I don't really think so. Like, I think there are much better... Isn't that what's already kind of happening in the matchmaking algorithm? Like, in a very reductive way, isn't that sort of what's going on? I think in a very, like, abstractive, reductive way, you could say that. But, like, that's not really what's happening. We are not trying to, we're not simulating dates. I think it's like a very expensive and inefficient way to do something that actually is much more straightforward.

58:51There's just a part of me that says like some, you're going to have some competitor that attempts to this to happen and we will all have to contend with it. I just think that that's a bit of a red herring for, you know, trying to map someone's psyche and guess how they're going to be. That adds a lot of complexity when actually you can just talk to people very directly about who they are and what they're looking for and what matters to them and compare that against someone else and what they describe and actually make a lot of good connections and clear understandings to figure out who should match with who.

59:23This has been great. Tell people what's next for Hange. What should they be looking for? It really is the evolution of the product. I think that the shift to AI is going to be bigger than the shift to mobile for the industry. I think that we're going to be able to do things that, you know, if you think big picture about what mobile did, it just made the process more approachable, faster, more fun, easier, more approachable. But it was still the same fundamental experience of just like cruising for people, trying to find someone that, you know, based on very limited information, matching with them, trying to figure it out, going on a date, realizing this is not your person, trying to find the next date.

59:56moving much closer to a world of really deep understanding of compatibility, being able to zero people in on the right person very quickly. It's going to be a very transformative experience that I think is going to very much change people's understanding and perception of the industry. So I'm really looking forward to the next couple of years because I think that we will see more change than we've seen ever in the industry before. We'll have to have you back to check in on how it's going. Thanks so much for being on Decoder. Yeah, yeah, thank you.

1:00:27I'd like to thank Justin McLeod for taking time to speak with me on Decoder, and thank you for listening. I hope you enjoyed it. If you'd like to let us know what you thought about this episode, or really anything else, drop us a line. You can email us at decoderatheverge.com. We really do read all the emails. You can also hit me up directly on Threads or Blue Sky, and we have a TikTok and an Instagram. They're both at decodarpod. They're a lot of fun. If you like Decoder, please share it with your friends and subscribe to your podcast. Decoder is a production of The Verge and part of the Vox Media Podcast Network.

1:00:49Our producers are Kate Cox and Nick Statt. Our editor is Ursa Wright. The Decoder music is by Breakmaster Cylinder. We'll see you next time. Thank you.