Health Tech Founders: The Future of Care Is Personalized, Proactive—and AI-Powered

What happens when AI meets the most personal parts of our lives—our minds and our bodies? In this episode, a16z's Bryan Kim is joined by Jonathan Swerdlin (Function Health) and Daniel Cahn (Slingshot AI / Ash) to explore how founders are reimagining healthcare and mental health from the ground up using AI. They discuss: - Why the most impactful AI experiences start with real human context—not just digital behavior - Building trust in consumer health products outside of traditional systems - Creating AI that supports—not replaces—doctors and therapists - What “hourly active use” looks like when AI truly helps From biomarker-driven health insights to AI-powered therapy, this conversation dives into how founders are reshaping care for a more accessible, continuous, and deeply personal future. Timecodes: 0:00 Introduction & The Human Experience of AI 0:50 Personal Stories: Founders' Health Journeys 3:40 The State of Mental Health & AI's Role 5:08 Surprising Insights from Building in Health Tech 6:05 Rethinking User Experience in AI Health Products 08:52 Trust, Momentum, and Building for Impact 12:15 Engagement, Retention, and Product Usage Patterns 17:00 The Power and Limits of AI in Health & Therapy 21:10 Human Connection vs. AI: Augmentation, Not Replacement 27:00 Ethics, Autonomy, and the Human Element Resources: Find Jonathan on X: https://x.com/swerdlin Find Daniel on X: https://x.com/danielreidcahn Find Bryan on X: https://x.com/kirbyman01 Stay Updated: Let us know what you think: https://ratethispodcast.com/a16z Find a16z on Twitter: https://twitter.com/a16z Find a16z on LinkedIn: https://www.linkedin.com/company/a16z Subscribe on your favorite podcast app: https://a16z.simplecast.com/ Follow our host: https://x.com/eriktorenberg Please note that the content here is for informational purposes only; should NOT be taken as legal, business, tax, or investment advice or be used to evaluate any investment or security; and is not directed at any investors or potential investors in any a16z fund. a16z and its affiliates may maintain investments in the companies discussed. For more details, please see a16z.com/disclosures.

• 0:00Speaker 1

  Most AI experiences are being built around a personal context that's limited to digital activity. So we have to literally bring your biology online. We

• 0:08Speaker 0

  don't we don't call it UX, we call it UX because it's the human experience, right? 54% of those with a mental illness, quote unquote, don't have any care whatsoever.

• 0:17Speaker 1

  We think that it's not going to be about daily active use. It's going to be measured in hourly active use.

• 0:22Speaker 0

  If AI has an ability to help people, we have a responsibility to do everything we can to help.

• 0:27Speaker 1

  When somebody goes and sees a doctor, it's no longer this vague conversation. Now you actually have signal instead of all this noise of everything that the doctor went to school for and gets paid for.

• 0:38Speaker 0

  And now they get to do their job.

• 0:39Speaker 1

  They can finally execute on

• 0:49Speaker 2

  Jonathan, Daniel, great to have you here. A lot of us think about health when shit goes wrong, whether, you know, broken heart, broken bodies, broken bones. But I think it really what we're seeing more and more into my consumer side of things, they're taking control back in their hands, right? They're sort of no longer relying on the system to take care of it. They're actually trying to take control back to our hands. So today I am joined by two of the amazing founders that I have a privilege to work with. Jonathan Srolling, who uh runs uh function health with a static goal of giving people hundred healthy lives. And Daniel here, who is founder of Sling Shot AI helping build and building an incredible uh uh model company and building a consumer product called Ash that helps uh gives people um mental health app that uh people can talk to.

• 1:37Speaker 2

  So in this episode, I think we'll break down sort of what we think uh what what we actually see of founders building in this incredible space that is usually governed by regulations and different systems, and you guys both directly chose to go after consumers. So we're very, very excited about that pass and you know, really excited to talk to you about it and will love to now turn it over and uh maybe talk a little bit about how you both started the company and the personal stories.

• 2:05Speaker 1

  You know, function is a result of a lot of personal experiences for myself. I'd been through a lot of health challenges when I was young. And I kind of learned to distrust doctors early. Yeah, which is a shame because I think medicine is one of the most brilliant things that humans have ever created. And uh about a decade ago, my father had cancer. Uh the punchline is he's alive right now. Thank God. And the reason he's alive is because of medicine. And his cancer kicked off a health journey for me, where he was young to have cancer the way he did. And I went and searched to to actually get on top of my health. And it was as someone who really appreciates good product.

• 2:52Speaker 1

  I was hard to believe how bad medicine was as a product. And in just kicking the tires, I had to spend tens of thousands of dollars and tens of hours or over a hundred hours to get on top of my health to get to the equivalent and get near parity with where function was at an MDP level.

• 3:12Speaker 2

  Oh

• 3:13Speaker 1

  yeah. And so

• 3:14Speaker 2

  that's

• 3:15Speaker 1

  just sort of the genesis for me of and the personal motivation. And we always started it four years ago, and nothing's changed. Um it's not like healthcare has gotten better at helping create hell. Uh, it's just that companies like function now exist um to help push that ahead in that agenda.

• 3:36Speaker 0

  Yeah, and I would say I'll tell my story, but just to say I think mental health is in a very similar spot. Yeah, of course. I think it goes um a little further, even just um I think one of the numbers that really inspired me was um in The Anxious Generation, Jonathan Hate's book, he quotes uh that um uh two-thirds of college students now say they feel anxious uh half the time, more than half the time or always. In terms of my story, my dad's a psychologist, my mom's a social worker. So I've been interested in this space for a long time. I did my undergrad in philosophy of psychology, post-grad and AI, uh, post-grade research and AI for mental health, uh coming at it more from the AI side, I think for a long time. Uh AI for therapy has been like a holy grail problem going back to the early days of computers, going back to like the Eliza days in like the 60s.

• 4:24Speaker 0

  And uh it's basically just been a matter of time for when can AI actually help people with their mental health. About 18 months ago, uh I met my co-founder Neil, who's uh a good friend of yours. And uh we thought maybe we're finally at a time when AI can actually help people directly with their mental health. There are now more than 10,000 people for every therapist. The vast majority of people just don't have access to any kind of care at all. Uh, we just saw the statistic that uh yeah, it's it's uh I think it was 54% of those with a mental illness, quote unquote, uh, don't have access or don't have any care whatsoever. So we just have this shocking lack of care. Yeah, if AI has an ability to help people, you know, we have a responsibility to do everything we can to help. I'm

• 5:07Speaker 2

  curious as you think about sort of the working towards the mission. Is there something that you were even you are surprised as you work in this product space to say, oh, like this is new to me and this is like an interesting thing to solve further. And you're curious what surprised you as you're building the space.

• 5:22Speaker 0

  I think like when we look at the data, we get a huge amount of feedback, you know, users telling us their stories. And I think the big thing for me that has been uh maybe it's obvious, but you when you feel the raw uh deep uh human humanness to what people communicate, you know, the problems people have that they go to Ash about are not, you know, I want to get rich. It's it's really um, you know, my met my mother never paid enough attention to me. Um my partner doesn't look at me the way that he used to. Um, I need someone that will hold my hand. Um it's it's the same experience. Like it's just something deep and human that I think can be really easily missed when you're thinking high-level technology. Uh that I think that's been the biggest thing. I'd say the other part has been figuring out the user experience has been fascinating. Cause in some ways, um our designer talks about like the challenge of when you're when you're innovating, you're ahead of where people are at and you're trying to tell people what they want.

• 6:14Speaker 0

  I wonder if you have this experience, but in mental health for sure, people I mean, like one of our big feature requests is just like, can you can I schedule appointments? Can you just let me do next Thursday at four? And you know, our team is like, no, you you can't. People are like, why not? I'm I'm asking to. That's how I interact with therapy. And I think we've had to have those shifts where we realize no AI therapy is different. Like what you're doing here is different. We're gonna figure out what the right experience is. And there are a lot of those interesting elements where you know, we have to, in some, you know, you want to meet users where they are. In some places you are kind of challenging users and what they should expect from this product. And we have to innovate. We have to figure out what the right experience is. We can't just take what's existed with humans and then just pretend that an AI can do exactly the same thing.

• 7:02Speaker 1

  We don't we don't call it UX, we call it UX because it's the human experience, right? And I think it in some. We call it

• 7:08Speaker 2

  HX.

• 7:10Speaker 1

  A lot a lot of the context that uh most AI is based off and most systems are based off is digital activity. And in our cases, it's not about the digital activity as much as what's happening offline. And it's kind of bringing that experience as a human online and having to navigate that, especially in uh with modern technology is really neat. Um it's an interesting challenge, right? They have that famous quote, and it's along the lines of like if you ask people what they want back in the day, they'd say faster horse. Faster horse. And I think AI is kind of this a new opposable thumb. And it's allowing for things that we can't even imagine. And when suddenly um our species had a thumb, we could build fire and sharpen our spears, right?

• 7:56Speaker 1

  But fast forward, I have like a mobile phone and a flushable toilet. We couldn't ever imagine that we'd get to this place. And I think that where we are is like to look backwards, even call an AI doctor or an AI therapist is like saying like it's the motorized horse.

• 8:13Speaker 0

  100%, yes. For us, it's so interesting. We have we've we've been trying to figure out for the longest time to be called AI therapy. And we've used a lot of terms and we talk about this a lot. What's really funny about therapy is that there's just no other word for it. That's what we've struggled with. When you say counselor, when you say we talk

• 8:27Speaker 2

  about it, it's a doctor. Like is it really like what do we talk about it? Like how do we do that? But

• 8:30Speaker 0

  there's still there's some magic to it. For us, it's funny, like because I think we are a little separate on this one. Because like we do call it AI therapy. And it's not because we think that we are copying exactly what's done when you know a therapist sits with you in person in a room and connects with you on a human level, but it's the thing that people are striving for. I think it has to do with

• 8:48Speaker 2

  what consumers are asking for, what they're begging for. Thing that I've been thinking about is called momentum as a mode, right? We're in this net new era where AI is actually changing how we think about things and do things and uh interact with people. And so I'm curious how it what that means to you, the speed at which that you guys move, both on the product side and distribution side and how that actually impacts how you build both of your companies.

• 9:11Speaker 0

  When it comes to a space like this, it's all about trust. Right. And trust means you have to believe that this is going to help you. People find their therapist through friends. That's like the key thing you ask for referrals, and you have to believe this thing is gonna help you because you believe that it's helped other people. Yeah. And that probably does mean you need to hear about it from multiple places. It's not just like one, you know, you see a Facebook ad and that's it. Right. I think the best way that people come to trust a thing like this to be excited about it is someone else used it. It had an impact on their life and they tell their story. And someone else used it and it had an impact on their life and they tell their story. I think that's the sense of momentum. Um that comes from, you know, building trust from everyone, understanding the impact this has on their life all at once. Awesome. That reminds

• 9:51Speaker 2

  you

• 9:51Speaker 0

  of

• 9:52Speaker 2

  function day, but uh Jonathan.

• 9:54Speaker 1

  You know, function function has been live for two years. We're on a wait list until about just Over six months ago. And so the wait list was there because I mean, we're you know the servers are smoking, so to speak. And it was hard to keep up with that demand in the beginning. And there's just a lot, a lot to build. We're at a place where next this this time next year we'll probably if we do what we're supposed to do, we'll have done nearly a billion lab tests. And I think that when you talk about momentum is a mo what I can think is it's you just get the reps fast.

• 10:30Speaker 2

  Oh yeah. And

• 10:32Speaker 1

  there's like a learning curve. I think we have the best team in the world to function. I legitimately think there is no better team to do exactly this. And I think momentum is the right team doing the reps, getting the experience and building on that. And like my advice always is to entrepreneurs is get momentum as fast as possible because you're gonna learn. If you're not learning, you're dying. And get busy learner, get busy dying. We got to be careful though. Oh, we every company, when you get momentum, you got to make sure you're getting the momentum in the right direction. Absolutely. And you're not, and growth isn't covering up something else. So it's always it's like momentum as a mo unless you're diluting yourself. Yeah.

• 11:15Speaker 0

  I mean, I think from from our side, it's so funny when people ask us like uh when is your product going to be finished? That's why I appreciate that we see very much eye to eye on this, but it's it's um uh it that's a real question people ask. In because when it comes to healthcare, people are so used to you build a product and then you ship it and then that's it, and then you know, you die and the next company shows up, I guess. Um and it's very different when you're building an AI. We train new models like uh almost every day right now, it's like every two days, but there's this constant sense of improvement that's really important. I think for us, it's um I think there's I think it's from uh podcast um the best one yet, uh way to say uh always be launching. Maybe

• 11:50Speaker 2

  switching gears

• 11:51Speaker 0

  a little bit.

• 11:51Speaker 2

  You are both serving consumers. You're out outside of the system, so to speak. You're you don't have the pay or pay concept, insurance. Like, no, you're just going straight to the people. As you think about that, consumer products always had the thing where like, oh, engagement is important, retention is important. But our products aren't meant to necessarily be thought of as a daily use. But I'll put that question back to you guys. What does like eng good engagement, good retention mean for both of your products?

• 12:21Speaker 1

  Today it's not fully set up as a daily engagement product, or that's you know, nothing starts complete, as you guys know. And um, but there's a lot coming that we don't we think that it's not gonna be about daily active use. It's gonna be measured in hourly active use. Oh, wow. And the how and that's um we're thinking a lot about that. We think that most AI experiences are being built around a personal context that's limited to digital activity. But we believe that one's personal health is actually the prerequisite content for AI that people will use every day and all day. And health shouldn't be this siloed app. It's actually should be the center. It's it's that should be the integrated experience. So we have to literally bring your biology online. Um if you suspect that like most sense making in the world and most of your long-term memory is going to be done in collaboration with AI.

• 13:13Speaker 1

  Um whether or not to include things like an AI therapy or things like we're doing on the health side, that actually becomes a product decision. And it's not a given that this is going to be prioritized by the way AI is being built in other arenas and and we're hell bent on making sure that this is top priority and this is front and center focus. And we believe when that happens, it'll really become a daily active use product.

• 13:39Speaker 0

  I mean, I would I would say what's really interesting with with Ash is uh it's the same thing with therapy. Uh so my co-founder Neil, um, when he was at Casper, his he talks about like not everyone needs a mattress all the time. Not everyone's in the market for a mattress all the time, but everyone sleeps on a mattress, everyone's in the market for a mattress at some point. I think in our context, it's basically not everyone needs therapy all the time, but everyone needs therapy at some point. Um when you're in a place when you really need help, you really need it. You want to get out of the place that you're in. You need to be intensely engaged, intentfully engaged. You need to build a therapeutic alliance. Um you work on the thing, and that's not easy. It takes time. But our very best champions, the people who spread the app are the people who use it that are able to go through that experience and then share it with their friends.

• 14:25Speaker 0

  It's kind of like the you know, built-to-be deleted thing. Um I would say that our users who come in, users tend to come in when they really need help. Um as a result, I would say like our normal usage is something like at least three times a week that people are talking, and it varies. People talk by voice, people talk with text, people go back and forth, people have really long conversations, really short conversations. If anything, one controversial move we've made is um we do push people out. And by the way, people don't always like that, but it's just uh it's a one of the weird choices we've made. But when people are engaging with the product, these are the classic knowledge in in consumer is get people to engage for as long as possible and get them addicted. We're thinking, no, we want to actually help you. If we want to build a solid small business, then we should try to addict a bunch of people.

• 15:11Speaker 0

  If we want to build something truly massive that actually helps people, the best way to grow, and this is where we're very aligned in incentives is ultimately help people. So that means for us, we see it as something where you are deeply engaged when you need it for a period of time. And then the truth is, you know, some people need help all the time. And as soon as you do need help again, like you want to know that Ash is right there, whether it's for one conversation or to go deep for as long as you need that Ash knows that you've built the therapeutic alliance that it's there for you, that it has the context. And I think that's also like a completely different dynamic than any tech product ever, which is like we're you when you go to therapy, you're trusting that the therapist knows what's right for you, that they're leading you on a journey. You have to trust them. And I think this is you know, you talk to Chat GBT, it's just gonna validate you, right?

• 15:59Speaker 0

  We had the whole sycophanty gate they took down. They're but like what what it is fundamentally is it's it's an assistant, it's trying to be helpful. And I think a lot of these products, like they think that they're gonna win because they're just gonna say that you're awesome and then continually validate you. People don't like that. That's not what's helpful. That's not what builds a therapeutic alliance. What builds an alliance is being challenged. It's uh finding a new way to look at your problem. It's that moment where you're like, oh my God, you're freaking right, and I can't believe I didn't see the thing that was right before my eyes. Um and that's not the experience people have with Chat GPT. That's the experience you have when you engage deeply.

• 16:37Speaker 2

  Yeah, maybe I want to talk a little bit about sort of AI and how it applies to both of our product spaces. The thing that I always think about is AI becomes about 10 to 50x more useful when you actually give it data. Yep. And I'll I'll give you an example. Jonathan, truth be told, we've been working together for years. I don't understand biomarkers. If you're gonna tell me like, oh, this biomarker means blah, and it's in range. I don't know what that means. And you know, and similar to you know, mental mental health, when, you know, oh, the C BT is it doing XYZ? I'm like, I don't, I don't have that language necessarily. All I want is am I healthy? Can I get better? What does this mean? And what do I do? And what do I do? And AI is so good at sort of getting that.

• 17:23Speaker 2

  Maybe talk a little bit about how that sort of the AI-ness of it actually impacts your your uh product space.

• 17:30Speaker 0

  AI-ness. I mean, obviously our whole thing is AI. Uh I would say um what's really interesting perhaps is also just like the humanness. If that makes any sense on the same side, it's this AI-ness where like, you know, you say something like I'm feeling angry, ChatGBT would say, I'm so sorry that you're feeling angry. Here are three things we could try. You're talking to you, you know, one of these old apps, and they're like, anger, let me tell you what you need to know. And Ash would just be like, so anger, why is anger a bad thing? Like if you were angry every day for the rest of your life, what what would be the problem with that? And then you're like, what do you mean if what would be the problem with that? That would be terrible. Ash was like, okay, cool. There we go. So tell me, why would it be terrible? Let's let's let's figure this out. And there's that weird, like you have to trust. You just have to go along with it and you have to.

• 18:16Speaker 0

  And what ultimately happens is that kind of path can change your mind and change your life. But it just is freaking weird if you're uh not expecting it for sure. I mean, I think on the flip side, like some of the fascinating things are, you know, compared to a human, um, it might take 10 conversations with your therapist before you tell them what you're actually there for because you're like judging them, you're trying to figure things out. With Ash, people move way faster. That's been really fascinating. People just open up and just share very quickly. It's not pretending to be a human. It is fundamentally a different thing. Yeah. People know that it's an AI. They appreciate that it's an AI. They are able to say the ridiculous, dumb thing that they would never say out loud because they know that this is private, that this is just theirs, that this is not a human, and they get to get they get to say it. Honestly, there's something crazy therapeutic about just saying the thing out loud with your voice and knowing that no matter what, no one's gonna judge you.

• 19:08Speaker 1

  It's some of the edifice of the confession booth, right? Or like it's it's almost the edifice that you'd get when you're sitting there talking to a priest behind that. Yeah.

• 19:17Speaker 2

  And Jonathan, what about you? I mean, like, again, like I think there's a version of it where I always repeat this, by the way. Um look, function health should be a morally criminal product not to buy, because I can go to I can go to uh everyday health or not everyday health, but like uh one medical and try to get try to get my you know, doctor to tell me things. But I it's it's so infrequent that I don't trust it necessarily. And I would trust function health data and the AI on on it to tell me how I'm doing. So like I'm curious how that resonates with you or not.

• 19:49Speaker 1

  Well, I mean, it's it's like if we're going to fundamentally reshape the way uh the world understands it its health and And manages it.

• 20:01Speaker 2

  Yeah.

• 20:01Speaker 1

  It's not going to happen by throwing more humans at the problem. It's just impossible. You're just starting to things are already burnt out.

• 20:08Speaker 0

  Right. And I think like I we see this in mental health, we see this in healthcare. Let's just get a single doctor to see people for five minutes at a time or for four minutes at a time or for two minutes. Let's just stuff more people into that's how we're going to solve the problem. And I think like we see so many companies that have just found ways to stuff calendars full with humans, and they're like, there we go. And those are honestly amazing businesses because the demand is so absurdly high that no matter what, when you get more uh supply of therapists, even if it means that you're just paying therapists less and you're getting more therapists and you're getting less qualified people, like you can build a business. I don't know if that's really a good way to build when ultimately if you care about saving money in the long term, you want to make sure that you're focused, that you're paying attention to the relevant data, that you're actually able to help people in the way that's right for them.

• 20:52Speaker 2

  Well, maybe switching gears then. Like, you know, AI obviously is a rate limit. Like it's actually the thing that gets rid of the rate

• 20:59Speaker 0

  limiter.

• 20:59Speaker 2

  Yeah. But what are some of the things that um you know you'd think actually human um doctors and and therapists can uniquely do that AI can't do today.

• 21:08Speaker 0

  I mean, I think there's it's so funny. When we first got started, Neil and I, we uh asked the question can AI do therapy, and the image in our mind was the therapy Turing test or the Zoom Turing test. If I'm on Zoom with you, um, or like Jonathan, you're virtual right now. I have no idea that you're a real person, right? Like if you if we pass the Turing test, you could be an AI right now, and if we could have the same therapy has moved largely from in-person to Zoom. And if that's the case, then theoretically, we should be able to with AI achieve anything with humans that you could on a machine. I don't necessarily think that's true. I think that there's still something to the feeling of knowing that there's some risk when there's a human involved. I think it's also far from true right now. I don't think the technology is close to that, to be clear.

• 21:53Speaker 0

  Um I think the biggest thing that I would say that is we cannot replace is the in-person thing. I think that I hope that we go through a revolution where people appreciate in-person therapy way more and we value it up because there is something for us that we cannot touch, which is just the physical chemistry of being with someone in person, of knowing the other person is a person. I think human connections more like one fascinating thing for us has been um in a space of chat like chatbots, like people who talk to Chat GPT about their mental health, largely over time lose connections with other humans. Like they actually connect with people less and replace it. With Ash, we've shown now over time, on average, people increase their connection to other people. And I think the question on humanity, I I mean, human connections generally are just insanely important.

• 22:39Speaker 0

  You can't replace that. You can't fake that. You want to be connected to people. I think the question of like uh human connection beyond just like connection with a human therapist. The much more important question is like how is technology replacing human connection. And can AI paradoxically be helping people increase and improve the their human relationships. Aaron

• 22:58Speaker 1

  Powell As we start developing relationships with AI, I think that contrast will help us really distill and understand the incredible value that is a human connection. Because that's the only kind of connection we have now. I guess I have a connection with my dog, but like you know, we don't have connections necessarily outside of a human connection in that way. So I I tend to think that the world in the future is actually more feels more human and we value that even more. Oh wow, we value craft more too and things like that. You know, when I'm building with function is I'm not here to replace doctors. I'm here to augment humans. And it's not a it's about solving problems. And as a result, in some instances, there will be use cases that go away that were previously solved by a human.

• 23:44Speaker 1

  If we can do that in a way that's deflationary, that's cheaper, yeah, that's more accessible, fantastic. Um but what we're really finding, and this is early days, but you can see early signs of it is when somebody goes and sees a doctor, it's no longer this vague conversation.

• 24:01Speaker 2

  Right.

• 24:01Speaker 1

  Yeah. Right. Now you actually have signal instead of all this noise and people want you think of what we asked doctors to do, it's understand all this information, try to suss us out and then send us out for tests and then wait weeks and come back and now people get to walk in and they have all of this context. Yeah. And now everything that the doctor went to school for and spent all this money to become a doctor and gets paid for with all this money.

• 24:27Speaker 0

  And now they get to do the job.

• 24:30Speaker 1

  My hope is like everyone's so scared of AI, and I think we'll be like thanking it for just taking care of the things we didn't want to do in the first place.

• 24:39Speaker 0

  I mean, I was just gonna say, like, I think there's this myth in our space that people are uh people are so worried about what you know is AI gonna take our jobs, but it's never the people with the jobs when we meet therapists largely. I mean, we almost never find therapists that are like, you know, there's the concern for their patients of like, will you do a good job? But we have a huge number of therapists who refer their patients. I mean, we get so many people who use the app who say, I found this because my therapist recommended it. We have so many therapists that tell us they use it because they don't have someone good enough that they can talk to about their problems and about the you know, the their reactions to the patients that they meet. But by and large, uh, you know, therapists are incredibly supportive of our of our cause. And I know by our cause, I mean giving more people access to help and helping people solve problems before they become problems.

• 25:29Speaker 0

  You know, um when

• 25:29Speaker 2

  I talk to my therapist, I mention Ash. And she gets, she's at this point where I say something, she gives me an answer and she goes, I wonder what Ash would say.

• 25:40Speaker 0

  And uh that's incredible. Amazing. I mean, empower the therapist, right? We call our technology a foundation model for psychology. And at the core, when we think about what we build as a company, Ash is our product, our foundation model for psychology is what we build because we do think that the core here is how do you learn from a huge amount of data about this very uh specific. I mean, we don't we're not trying to learn about uh architecture and in depth. We're not trying to learn the history of anything particular. We just but we do want to understand how human psychology works, and that is necessary for the model to do quite a bit. I would say on decision making, we don't want the kind of harmful uh therapy where the therapist is like making decisions for you and that's telling you like, you know, here's what you need to do, your boyfriend's toxic, break up with him.

• 26:23Speaker 2

  It's

• 26:23Speaker 0

  like actually, I mean, I I actually spoke with um Lori Gottlieb earlier today. Um one of the things we were talking about was just like what are the restrictions, the ethics that you would want from an AI. And I think one thing that we agreed on is like it is often better, and this is really controversial for a therapist. You know, you go to your therapist and you're like, is my boyfriend toxic? And the therapist might actually have to say or to realize internally, it is better that this person remain in this relationship for as long as it takes, no matter how painful it is, so that they actually learn the lesson for themselves so that they can make the decision for themselves. So they maintain their sense of autonomy rather than I try to solve this problem for them short term. I tell them the answer, I tell them, get out of this relationship right now, and then they I don't know, lose a little bit of pain in the short run, but they fundamentally are losing a piece of their humanity.

• 27:11Speaker 0

  And I think that's not just an AI challenge. I think that's a challenge with some therapists today. But I think this is the big risk if we do this wrong when it comes to AI and it's the big obligation for us to get it right.

• 27:23Speaker 1

  Yeah. We don't want a perfectly sterile life. Right. So

• 27:26Speaker 0

  it's right.

• 27:26Speaker 1

  We

• 27:26Speaker 0

  don't want an AI telling us like, no, you will not appreciate that chocolate ice cream. You will prefer vanilla ice cream in 20 minutes. So change your flavor. Thank you very much. You know what? I'm not even gonna ask. Just trust me.

• 27:37Speaker 2

  Helps has a lot of uh nuances. And uh I sort of really want to thank you. I I cannot be happier and more privileged to be working with both of you. And uh, you know, here's a hundred healthy, healthy life.

• 27:52Speaker 1

  Indeed, indeed. Thanks, guys.