Is Tech the Future of Funeral Services?

Adam  00:00

Welcome to another daring dive into the depths of human technology on this Anthro Life. I'm your host, Adam Gamwell. Now today we're veering off the beaten path into a world that we think we know. But upon closer inspection reveals itself as profoundly strange. Now, it's no secret that with all the technology around artificial intelligence, sitting shelves, and people's computers and phones, that it feels like today might be a kind of runaway of technology or technology overlords are taking over. But if you look closer, and you think about what technology is actually doing, we can think about it in the ways that humans invent technology as an aid in expressing our most basic human needs. Now, this is easier to see what have you and imagine the application of technology in areas that we often don't think of as being in need of technological disruption. Now, one such example is the end of life care phase one people die around when they do, how we take care of folks and how we help that process. Think about funerals and working with funeral directors and burials and things like that. Now, Faisal Abid, it doesn't just imagine this space and thinking about how technology can work in there, he's actually pioneering this area. And so Pfizer's career has seen him sculpting the backbone of the tech world from Google to artificial intelligence. But now his current endeavor is in reshaping something that's far more intimate, an area that seems relatively untouched by technological innovation in the past 100 years. So he founded an organization called Irene cremations. And it's not just a technological disruption, it's actually a compassionate touch in the unlikeliest of places. So today, you'll accompany us as we unwrapped Fuzzles innovative ideas and peek into the future of the funeral industry, of what end of life care could be around and how technology can help us be more human, the role of AI and so much more. So buckle up for an enlightening journey, where technology and compassion intersect in surprising ways. So I say Faisal, thank you so much for joining me on the podcast today. And I'm really excited to talk with you, especially because you have a, I think, a really exciting and diverse background. And then obviously, the kind of work that you're doing today, there's been a really cool through line there. So I'd love just to kind of jump in and hear your kind of might say, superhero origin story, you know, how did you kind of get into the work of Google dev work and cloud computing and AI, and then where you're working now in end of life care? So let's kind of hear about your process and how that all came about.

Faisal  02:10

Yeah, it's very, very fascinating process. I mean, I don't I never imagined myself, starting up a dedicated company, but somehow just dive into it. And we could talk about that. But, you know, I started, I started getting into technology very early on, since the matrix came out. I guess I was like, Okay, this is great. Let's get into this. And then, over the years in high school and university, I just kept tinkering around and building stuff and started to do consulting, for companies, and what eventually, that led to was me going, hey, you know, I should be building my own stuff. And so right out of university, I went to university, I never never completed it. But the second year, I took a break a year off to see hey, can I start building a business and I was working on like an Android business, then there's Android a joke launch. So I did that. And then that didn't work out. And that eventually got me into, like, just fully immersed in the Toronto tech scene where I was either working at very interesting tech companies, just as like an engineer building software, or just collaborating with very interesting entrepreneurs, who were building really cool stuff. And I would build basically free stuff for them. Or take a very small painting, as long as I got to hang out, I hang out with them, and kind of learn how they built that business. And so that, that got me into tech. And then sort of just by a bit of planning bit of coincidence, got into AI by started working at this company that was trying to build like Chad GBT before Chad GPT like, we were trying to Bill GBT for the enterprise. And so not that we we had built like any language models, we were doing a totally different algorithm, very primitive way of doing stuff. But we tried that for a while. And that's what got me really immersed into AI, machine learning and just doing a lot of Google related stuff. I just got into that whole world.

Adam  04:14

Now. That's super interesting, too. Because even thinking about the kind of proto world of AI because obviously now it's you know, we've now been about a year with with chat GBT as a consumer product, right? But like, that's an interesting part to have. Oftentimes, when we see technology development happening, it's like, on the enterprise level first Raider comes in that arena. I'm curious just immediately on that too, as like a tech nerd, too. It's like, how were you approaching AI? Kind of before using large language models, as was this also around like conversation? Yeah. what it looked like,

Faisal 04:45

yeah, it's actually actually really good. Good question. Very interesting, because, you know, back then, the idea of NLP there was, I think attention the attention paper was written around then, and we were trying to look at okay, Well, what can we do with that? So we had a great, fantastic data scientist on our team. And that's where I've learned a lot of AI from Jay. Everything I've learned has just been through peers or being on the job and trying to figure it out. And so as we well we can build the idea was, you could say, hey, you know, I want to set up a call with Michael next week at 3pm. And the bot would go, Okay, thanks, Faisal. It will reach out to you via email and go, Hey, Michael, phase wants chat with you tomorrow, next week at 3pm. Do you want to chat with them? And in the perfect, perfect world, you would say yes, or no, people don't act perfect. And they will be like, No, actually, I'm free on the for next week. But I also have like a wedding, whatever. And that's where it would fail. So we will, we will try to do was build that NLP engineer, that was read messages and try to trigger intents, it was very intent based. And this was around the time where a company called api.ai, I believe, I had launched and Google had bought them recently. And so we were just trying to build this for one specific use case, we were very early. Because one of the one of the very interesting problems we ran into was customers, especially the enterprise. Were not ready for this at all. And the reason is, because the second did fail. They threw it away. The beauty of chat GBT and light language models is that they fail, but they fail very, very gracefully. And sometimes, that gracefully is wrong, because they hallucinate and they say a bunch of crap, that is not true. But to the user, it doesn't seem like a failure. It just seems like the language models confused. Whereas with our technology, where we were building a failed case, if you were to say, oh, sorry, next week, I'm busy. I got to take my kids to school. We the child will respond. Sorry, I don't understand. Then the second dude that, you know, it's not working like you think, you know, the illusion is gone. So we were we were very busy doing that. Ultimately, the business did not work. But yeah, that's, that's, that's what we were doing back then.

Adam  07:17

And that's interesting. I mean, because it is it is I was talking to one of the heads at hip sauce that works in ethnography the other week. And one thing that he was talking about was interesting, like this idea about how chapter GPT gracefully fails. I like the way he said that too, because there's this interesting point where, you know, and I don't know if it's just related to how people engage with technology, but, you know, you've probably heard too, like sometimes we say, Oh, we're kind of in a post truth era or have been there for a while. And part of that, you know, in the chat, GBT question is like, it fails. So gracefully. The the other way of saying that, too, is like that. It sounds it still sounds correct, even though it's wrong. Right. And that's an interesting kind of conundrum to fiddle with. So I guess I'm curious, your perspective on this, too, of like, if we're less likely to throw out the Tech because it doesn't quote unquote, work. But it looks like it's working, you know, how might we think about that? You know, like, what does that mean for us in terms of either trying to get the right answer or correct answer from something? Or is it more on devs? to kind of figure out, we should fail less gracefully, maybe give us a confidence score? And the answer or something?

Faisal 08:15

I think the very deadly thing about language models is that a lot of people are giving it a persona. A lot of people as not really the devs Sure, but also us as normal. People 1111 built it, they go, Oh, it a chatbot talking to me, verse is, Google search, comes up with a lot of cry, any search engine, who did hallucinate in the sense that I can search up something controversial, and I will find links to it, I can click on those links and read about it. And as long as I trust the source of those links, I will believe it, the language model is doing the same thing, but it's synthesizing it as a human speaking about to you. And if you start to trust that chatbot that and give it actual feelings, or this persona, then you're more likely to believe it. So I think the the education around that and the thinking processes to know that this is a tool that is returning a response, it is not someone that's actually sitting there thinking about what is returning if it's factual or not has done, quote, unquote, research. Yeah,

Adam  09:25

I think that's a great point. And I agree, too, because it's a part of it is kind of, we forget that we entered more amorphous stuff, right? And that we kind of think that it's real, right? Or can we ascribe kind of human feelings to it? And it's been funny, because I've seen more, like, not a ton, but more people are kind of having that conversation where they're saying, well, it doesn't actually like, you know, a large language model, which it doesn't actually know what it's responding, right. It's actually just statistically, you know, choosing the most likely next word that would make sense in a sentence, which is why the sentences often make sense, you know, but again, they could be wrong with what they're what they're actually saying. And that's really interesting. Yes, Basically as like we're adding more ability for more context, right? I mean, you can I think it's 120 tokens now that 120,000 tokens, right for GBTC, or turbo and Claude does 200,000. Now, I think. So it's just even these ideas of like, we're adding more context on the input side. And so it's like, I imagined to on the on the on the kind of the development side to that there's more that goes into how that answer gets made. I mean, I would think, you know, you made them more than that, I should imagine that side, you know, to what I like, Dude, you see kind of a growth on that side as well in terms of like, checks and balances of how the machine is reading the input, and like grabbing stuff from a vector database, and then maybe doing some web crawling.

Faisal 10:39

There's multiple ways. Yeah. So if you look at Google's Gemini that just launched. It also, I think can deal with up to 200 tokens, or 200,000 tokens and something like the ultra model, all of these guys. Yes, there's a lot of there's a lot of trips happening, because, and I am no language model expert. But from what I have seen and interacted with, and just my day to day, playing around with stuff as I build applications for Irene and stuff. What I realize is, there's this limit that we're reaching where, yeah, we can actually give in a lot of inputs, but it doesn't actually use all that input. So this is easily seen when I play with Claude or even GBT, if you give it and I believe even foobie quad anthropic anthropic. Yeah, I believe we even do a drop in Word, a good blog post about it, where you can pass in 100,000 characters, and change something somewhere in the middle, and ask about what was changed. And it doesn't understand it. Because for however, these language roles are working, they're not like, they're not keeping everything memorized, in a way, right? Because they're not going back. It's not like there's some mental whiteboard that they have, and they're scanning all that text, it is purely statistical. I think it's a miracle that it works the way it does. If you can figure out that giving an text you can next token predictable was there in the back, works pretty well. But it's not. It's, it's the limit that we're gonna hit. And and you hear, yeah, Lincoln, who's the head of Facebook AI talked about this a lot, where you're going to need a different architecture. And if you're going to have if you really want to mimic AGI or like the human brain, then you need a different architecture, language models definitely will reach a saturation point, a plateau where all really all you will be able to do is make it better for how much inputs it takes and make the output better, because your model is bigger, but it's not going to suddenly become a human and have his human level intelligence.

Adam  12:51

Yeah. I mean, do you do you think there's level

Faisal 12:53

reasoning? Yeah,

Adam  12:55

that's a good point. That's a good point, like, do you think about is that? Or do we see, I guess, you see those kind of conversations like is that some folks goals? You know, you know, I think, I don't know if it still is, but I remember, at least a few months ago, like on opening eyes website, they did say it's like creating, you know, AGI for, for all of humanity. Right. And there's this interesting question, too, because, you know, I was in conversations I've had with Byron Reese, who used to run Giga ohm, and now just writes a lot about kind of the philosophy behind language. And, you know, what are humans? What are organisms that work together? He said that, and he has a podcast, it's called Voices and AI. And it found that like, of 200, developers and engineers he worked with, like, maybe 5%, didn't think that humans were machines, therefore, we cannot actually make AGI right. But like most of the engineers, he found would say that we humans are machines, or we function like machines. And therefore we should be able to then code our way into something like that. So I'm just curious your thoughts about that, too? Like, how could we even know, I guess, like, Are we are we? Are we able to kind of be decoded? Do we think in like, or is that like the idea of like, we might need a different model to actually get to that human point, because I think that makes a lot of sense. Where it's like, if we have a statistical probability, we don't operate statistically. You know, there's things that we do as people I guess, but like, do we also have we have an irrationality to us? I suppose. So. I'm just curious your perspective on the right answer? Yeah, no, there's no

Faisal 14:09

right answer. I think what's very, very interesting is that if you look at what's there was a great book that I read, and it was by Max Tegmark. I think that's his last name, the famous scientist, and he was talking about consciousness. And what was really interesting about consciousness was, it's not quantifiable in the in the way that here's a red blood cell. And this is what a red blood cell is, it is comprised of a, b and c, like what is consciousness comprised of? Yes, it's in your brain. It's all it's all there. So there is this definition of consciousness that perhaps, depending on how loosely you defy that, you could say, Okay, sure. All language model is conscious. In the same way I guess. A bird is conscious, if we can even prove a bird as consciousness to some level, but a part of me feels that I will probably reach consciousness will probably reach this AGI Artificial Consciousness sooner than we think. But mainly because, and it will be by accident, I don't know, the feeling I have is we're going to get to these language models and language models plus whatever comes after this, whatever architecture is that they're going to be so complex in nature, that the feedback loops that those complexities off the chaos is just going to have this emergent phenomenon where you'll have some algorithm or whatever that is able to reason and is able to introspect on itself. I think they're all you'll have to decide, is this truly a living creature? In the sense that it's just the way a human is living inside of a machine? Or is this purely just a byproduct of a bunch of algorithms that are talking to each other? And it's simulating? Because that is, that is the scary part? What is what is how do you if a language model plus, let's call it that, if a language model plus, is simulating consciousness? How do you know it's simulating consciousness? It's not really conscious. And that's, that's the weird thing for us. I guess we'll have to decide then whenever it happens, no, do I think it's gonna happen next year, this year and road? I mean, the scientific, the Sci Fi part of me wants it to happen, ASAP. So again, the three part of this, but I hope it happens in our lifetime. I look like it well, I don't know. Yeah, I I'm not. I'm not educated in that. I'm not well informed in that social sciences and deep, deep psychological tech to understand or to make a good comment on it. But as a technologist who does a lot of work with language model and just reads up on the stuff, I guess, I know, like 0.01% of audit, I think that's, that's my answer, Dubai.

Adam  17:02

And that that's a good, that's a good way to frame that. I think, you know, I would, I would, I would suspect that you're right, though, that in terms of how quick things seem to have just developed in the past, you know, year and a half, just on the consumer side, right now, we've been developing these things for years. But like, I would be shocked if we don't see sort of massive technological change in the next 10 years. Right. And, and that that says, launch me to even say 10 years, right? For kind of how quick visor you're growing.

Faisal 17:26

But you know, what's interesting, we will see a lot of change. But look at let's look at cell phones, right? Yeah, I was there, right? I remember being in high school when the iPhone was announced. So let's, let's look, let's think of iPhone as the starting era of this new wave of smart strike. Because yeah, we had the cool little Windows devices and blackberries and stuff for iPhone. For the time, iPhone came out till today, the entire world has changed, because of how people interact with their smartphones. But fundamentally, the smartphone hasn't really changed. In the past 10 years, I would say that, like if you look at your smartphone now, and if you look at your smartphone 10 years ago, they do screen better processor, cooler ops, perhaps by what is fundamentally changed, it's just that people have found ways to adopt it. So what I think's going to happen is, language models themselves won't suddenly heap exponentially getting better. What will happen is you'll have people figuring out how to adopt them into their lives, giving the perception that wow, this technology keeps getting better. But theoretically, like, yes, they'd maybe takes more tokens and whatever. But it's not an exponential change. It's not like, it's a whole different, like the language model plus we're talking about it's not like a hold music. So I think that's what we're gonna see with phones, are we with language models, and how people interact with it, to see what we saw in advance. And you can see the same thing about laptops, the MacBook and laptops of today. And if you go back 10 years ago, they're pretty much the same with just faster CPUs. So like, there's a whole new way of doing computing. People are trying die, they know what the VR and AR and all that stuff, but that hasn't taken off. So the reason why it hasn't taken off is whether that's too soon, or maybe that's just not how people interact with it. Yeah,

Adam  19:26

no, that's a good point. I think and iPhones are a great example of, you know, where we did kind of see we, we sort of knew it at the time, right? But then you kind of see this watershed moment of like, this will Yeah, this is gonna shift how we interact with technology. And that's like a touchscreen in your pocket. A cipher. It's kind of even funny on one level to that we still call them phones right? Because it's probably the thing that people do the least on them at this point. The

Faisal 19:46

computer you shouldn't do you call if you have a computer or a laptop which comments I mean a lot of smartphone this is I bought a mobile PC or some right.

Adam  19:54

Now let me tell you on my PC and that and that now I can just FaceTime on my computer anyway. So I guess that is the same thing. Exactly. Exactly. little laptop as my phone is a lot bigger device, right? Yeah. But I think it's an interesting point to where it's like, you know, what is it? Is it Moore's Law where it's like, like the size of computing gets smaller every every whatever? Every every seven years? Yeah. And so it's like that I think that's a good way think about language models to where it's like, they'll get maybe larger capacity of tokens, but like, it's not going to fundamentally change the architecture of how you work with one. How

Faisal 20:25

you spend a lot of people hum. How many tokens would people need, like this is this this might get coded as like, you know, the worst day they'll probably receive a how many tokens does one Li? Like what Bill Gates said back in the day? Because, sure, 100 tokens 200 job gain, after a certain point, by and let's say tomorrow, Google comes out with a million tokens. How many people are cruelly utilizing the million tokens in their input? Maybe they'll keep that child going for a long, long time, which is cool, but like, how? And maybe and, you know, this means that people interact with language Bibles differently, because right now, the only the only interface that we know on how to interact with language models is an interface where you click new chat, and you talk, whether that is on Google, I mean, bar chart GPD. anthropic, all that way. And the other ways you're interacting, the language models are synthesizing existing information. Three, you brought this up before vector databases and stuff, but essentially, it's all chat. So maybe there's a different way to interact language models. I mean, I'm doing some of that with I read, and we can talk about that later. But it Yeah, I don't know. It's very interesting.

Adam  21:38

Yeah, no, I agree. I think it's like we're at we're at a very interesting time. But yeah, I mean, actually, let's let's, let's talk about everything because I think this is obviously the other really, super interesting area that you're working in. I mean, this is when I when we think about like, tech disruption, you know, my brain doesn't always jump to end of life care or cremations. But, you know, what, an interesting place to work. So I guess I mean, to move us into that space tech, you're using your tech stack, how, how that happened?

Faisal 22:01

Well think about it, what is one, what is one field and one sector that has remained untouched? Like, if he you know, if you say I mean doesn't exist right now, then, if you want to cremate somebody that will the actions you're going to take to cremate somebody are the same actions someone took 100 years ago. But all your other actions in life have changed your if you want to call a pizza, if say pizzas existed 100 years ago, I don't know if they did. You would not pick up the phone, you would probably like walk down the street and go to a pizza store. Today you want to call a pizza, you'd call Uber Eats or you pressed Domino's website. Everything in your life has changed how you interact with it except for how you take care of a loved one after they've passed away. And so, Lee, the fundamental idea here was it's very serendipity how the idea came about. And it was two parallel peoples in key about a different times. We came together where my co founder comes from a funeral home background. Her father was a funeral director. I was in Japan, and I was looking at, I met an old friend there. And he was running a Buddhist cremation as Expedia were the idea Expedia for Buddhists, cremations, the idea was, I can go on, and I can get the lowest cost Buddhist cremation service. And they had built this platform. And it was all very interesting. And that got me very excited. Because this was right that and this was right near the tail end of my adventures with the Chad GPT for enterprise company. And I was thinking about what should I do next? And this idea came about, and I was like, wow, well, there's no one Buddhist I know, in Toronto, so this idea won't work. But the idea of using technology in this space is very interesting. And so I got back and I started thinking about it. And one of the things I've learned and just in doing so many businesses of the past is ideas that weren't Waller ideas, in spaces that I'm very familiar with, or I'm working with someone familiar with. So I kind of put that idea in my back pocket. I had a pitch there. I just had that pitch deck, shoved down in my backpack. Many months later, I need my co founder talking to her. Nothing really we don't nothing clicked like, you know, there's no cool idea we want to work on together. Near the end, I asked, What is your big idea and she talks about well, doing something in Cyril spaces gonna be weird. And I said, Hold on, pull out this pitch deck. It's all crumpled off. And I'm like, Look at this. This is what they're doing me Japan, what can we do like this in Canada? And so this has dealt with the genesis of the idea of what can we do in duster. And as we looked at the industry, it became clear that the only industry that hasn't evolved is death care. You know, you cremate your loved one the exact same way as you did 100 years ago or two Number 300 years ago, going to a funeral home. So why can't we create phone corded phone lines funeral home, where we exist without a physical space, we use the same exact materials that everyone else uses. Because it's a very regulated industry, all we're doing is we're removing the physical aspect of going in person, you we actually have used AI. So much of the backend, we've created a system called funeral OS. And Cyril s lets our funeral directors be more present, and talk to our families more, and basically provide a more human experience by using AI. And the way we do that is the average Cedral arrangement. When you think of all the different steps in wall, you talk to the family, you have to fill out all these documents, you have to send out all these logistics, etc, etc. There's so much work involved there. So funeral OS, aside from funerals will never interact with the family funeral Oh, as in shots of the funeral director, the funeral director interacts with the family. So say you and I are talking and I'm a funeral director, I can talk to you, I can say you and I just talk over the phone at 3am. At night, you call me and you say my friend passed away. And I you know, as a funeral director, I talk to you. And we have say a 20 minute shots, where you tell re X over certo where your friend is where your loved one is, etc. I hang up. As soon as I hang up, zero OS kicks in fetalis analyzes our conversation, extract all the relevant details, generate all the documents needed, then ask me as a funeral director, hey, can you verify if all of this is true? If not, can you tell me what to fix? I go ahead, maybe I change the way your name was pronounced. So it Parsons weirdly, I click OK, sounds good. I hit yes, you get an email, with all the information ready to sign and you hit sign. The transfer service comes picks up your loved one takes them to a crematorium. As they're at the crematorium, you get a notification funeral OS knows when they're at a crematory and now not to be cremated, you get a notification saying your loved ones are allowed to be cremated or where your loved one is. You don't have to call think of the anxiety that songs going through. You know, I wonder where they are right now I wonder what the process is you go online, you can see every single step where your loved one is. And then we handle a rib the remains fat. So what that's done is it's the funeral director can be more president can be even more human and away because they're able to just offload all the admin and everything else to the AI. And then they can deal with the family, we can serve more families, this way we can serve wherever we're licensed, we're right now we're licensed across Canada, with the exception of a couple provinces, we can license that without. And without having this, you know, we need a space here we always sort of five meter radius, etc. So and that that's just the federal law is part for the family. When they get access to they can create a memorial on our webpage they can use we have a tool called obituary writer. And what he ritually writer does is it takes all the details about your loved one, and generate an obituary. And it's very interesting. So when we launched this, we thought, you know, this is very cool. But I wonder how many people use it? And the reason why is you might think so you might think, well, if I'm writing about my loved one, why do I need an AI tool to write the obituary for me? That is true. But what we have seen is, so many families use it because writing is hard. You can I can sit here and I can talk to you about how great my friend is. But putting all those qualities in a three paragraph obituary is really hard unless you're a copywriter. So how do you so again, we're using AI to make something more human, you're putting in all the good qualities you're putting in things they love to do the things they hated as a joke, whatever, whatever you want to put in. And then you get five, six, unlimited permutations of obituaries that you can take, put it on the memorial page and saw families love it. And it's very interesting because I remember we had a marketing PR team or something, basically be like, all you know, I think this is gonna creep people out because it's kind of saying you're like using AI to generate something like this. And I said, Okay, well let's see what happens and just the analytic and how many obituaries we generate, how many memorials regenerate? Almost all of them, are you using the AI tool? So it's helping families be more human just by by the act of Keeneland letting the AI do the admin stuff, and then they can focus on what's important. Now, that's, that's

Adam  30:08

a super fast I mean business actually, that's cool to think about how it's got both the, you know, funeral director interacting with the family side, and then the family side, you know, products like working with the obituaries. So, yeah, that's I mean, that's, it's a cool point to you, you know, because as we walked into this part of the conversation, where it's like, what's in need of disruption, and it's like, this is obviously a very good, a good example of that, where it's like, especially the question of transparency, and there's a lot of anxiety in terms of if you're going with cremation to know where your loved one is, right? over is even having like, things like that, like we're at, you can get a textbook that or, you know, however that might work with pinging someone that like, that's, it's a kind of empowering, right for the family to be able to have a sense of both closure and know what's going on. Exactly. Because that's like, the hardest part is to say that if you just let go, then what's what's happening, right, and we don't really know. So that's the interesting, the interesting thing to think about these, there's kind of two pieces that that are jumping out to me and one is like, the the right kind of work helps us automate the drudgery or the the admin stuff so that we can actually be more human with each other. And that's a super important piece. And the second thing I was thinking of, is this idea of transparency. Right? And that do we know what's happening during the process? And across it and like? So I'm curious, like, as you think about these, these two pieces to like, I think we, you know, we've seen like the, that is kind of a, the idea of like, kind of knocking out some of the drudgery work. And I'm curious, like how your funeral directors feel about this, like, do they find that they have more time that they have more, you know, ability to kind of spend time with families? I mean, it sounds like that that is the case. I think that that's, that's a really great outcome. But I'm curious, like, have you heard from folks like, what their experiences are with that, like, I feel like I've just been able to kind of drop it and just be with the family is

Faisal 31:47

that our field directors absolutely love it. Because for them, they got into this business idea. Funeral Director is the most kindness, empathetic human you'll ever meet. And they got into this business, because they want to be kind and empathetic towards the families. They don't want to do manual admin stuff. Right? That means that is the that is where all innovation has come from and technology, if there's something repetitive, we automate it, and thanks. So that is the goal at Ireen, where, as soon as we see a funeral director doing repetitive stuff, we automate it. So the funeral can spend more time with the family can spend more time with more families, right? Because if you have one field doctor, and it takes six hours, at a normal funeral home to make an arrangement, then how many families can that funeral director actually serve? Right, two families in a day, we serve sometimes 1020 30 families a day? Well, it's because Sure, let the AI do everything right, you do what you're good at? Let us be more human. Let ai do all the admin stuff.

Adam  32:55

Yeah. And that's, I mean, this is a really good example to where it's, you know, there's, there's also this kind of other conversation happening, you know, in a lot of industry space, where people are like, concerned that AI is going to take jobs. And this is a good example, where it's like, no, this is actually this is the stuff you want AI to take. So we can actually be more human, right, because it's the other thing we was talking about, like tech, theoretically is there to help us be more human, especially when it comes to automation, alpha. That's a kind of a good model to help show other industries of like, this is what this can look like, where it's like when you actually put humans in the forefront. And you're knocking out some of the admin work as actual work is relationships. Right?

Faisal 33:28

Exactly. I don't think a and that's the key. The actual workers relation is very humans were built on relationships, the companies that are kind of jumping the gun, and going, we're gonna fire entire customer support team because we have AI is gonna die. My opinion, that's gonna go down a terrible, terrible path. Why? Because any customer service, people have wide variety of emotion. And the AIS as of today, and perhaps ever forever, might not be able to understand will not be able to understand human emotion, right? Human frustration, human anxiety, etc. And center over chat the we humans can. And so replacing that, look, you know, I saw someone recently tweet something about talking to Lenovo or one of these computer companies and computer manufacturers where they're like, my charger is bustling. And it's just like going in a loop so that you can charge GBG in the background is very clear. But it's like, Hey, have you thought about doing this? No. Are you thought about doing this? Like if this was a human, it would not go down that loop. It will say okay, clearly something's busted. Let me set up an order for you. So I don't think customer facing chatbots are the right approach when it comes to customer service or, like things like that. Yeah, customer service chat. Bots are great if you want to just like you know, retrieve information. Like, you know, perplexity is a pretty cool search engine. Well, they're gonna please Google. I don't know blindly. You know, that stuff is kind of neat. Yeah.

Adam  34:59

But then it's a good point. I mean, it raises an important question around trust, right? And why would I? Why would I either trust an organization that I'm working with or a service that I want I want to use versus another one. And so I'm curious to think about that in the context of Irene to that. Did you? Did you have any kind of challenges? Or maybe not like, like getting people to trust a more automated system? I guess, no, the families don't see that that kind of side of funeral Alaska, like, I'm just gonna be curious how that came into play, or the funeral directors that were skeptical at first, or something that

Faisal 35:31

we never want the family to ever deal with technology we want to be invisible, of the technology isn't visible in the background, the funeral directors are working with Sheila Lowe is the family's sure they can come onto the website. And they can fill out all their information if they want, which a lot of families do, but a lot of families color. And a lot of families want to talk to a humor. And this is where funeral originally came from, because originally our hypothesis was when we started was, hey, he really wants to phone call right now, because we're weren't millennials, like, you know, do you actually make a phone call you just text? But so we thought, hey, this is the hypothesis. Clearly it was wrong on like, day two, we realized it, but a lot of people started calling us. And so we said, okay, a lot of people are pulling us we need to figure out how to automate the phone calls, the phone call information collection, because funeral director was sitting there and making notes. And, you know, there was doing a lot of stuff. And we were getting more and more phone calls every month that we grew. So that's where this entire idea of fear of loss came in, where, Okay, why don't we build this operating system that just runs the entire business. But at the funeral directors do what they do best, which is talk to the families. And so the families will never interpret feeling well, I still interact with the documents, and the byproduct of a funeral was generates, but that is always vetted by the funeral director. And this was to your point where the field director is hesitant. That they would be if we said, there's no inter interaction by you. But that's not the way to treat AI, maybe in 10 years, 15 years. But we know AI hallucinates, we know, AI is just gradual, it's bullshit and bullshit out. And so if we always like, at every step in the process of funeral directors signs off on what the AI has produced, and what the AI is doing. And so there is human accountability, and there's always human intervention, all it's done is sort of taking you two hours to do is taking two minutes, and you can then be 30 minutes, looking over and going okay, everything looks good.

Adam  37:47

Yeah, that's great, too, because I think that that's a key piece where it's like the idea of human in the loop when it comes to working with technological systems. Right? Exactly. We don't step out of it. Like, it's really smart perspective. I'm curious to think about this, I think, on the one hand, because there is like such a perhaps, like more enhanced interaction between funeral directors and families like curious like, how you think about like, as a business and as working with technology to like, how do we ensure that we kind of come across as caring, like when we're, you know, as a human right, when we're like, yep, taking a technological approach to things like this? Yeah. Well, what's

Faisal 38:22

interesting is look at our reviews, families, know that we are one of the most caring fewer homes out there. And that's because the funeral director has time, feel that he knows they can spend 1015 minutes on his call, because the AI is dealing with everything. And so, because of that, you can call us like I said to me, I am well reason why I said that, as you can call us 24/7. We're online, right? Like, we weren't like you have a death happens at 4am. You can give us a call. If at 4am, you have anxiety, and you're unclear about something about what just happened. And you want to verify where your loved one is, or something to do with anything. You can actually call us, you will have a human pick up a human answer. And if somehow a funeral director was not available at that time, the funeral director will call you back ASAP. Right. So there is always someone ready to chat with you ready to talk to you and answer your question. So that's, that's what technology enables you to do that. Geology enables you to be more human. If it's applied in the right way. A physical funeral home. I can't go to a physical funeral at 4am and knock on their door and be like, Hey, I have a question. Can you answer that for me? Yeah, I have to wait. I have to sit in my anxiety till like 9am in the morning, or sometimes, if on the weekend, some funeral homes might be close. So I have to wait till Monday, which is not good for families.

Adam  39:53

Yeah, that that's a great point too, in terms of like, especially when something is anxiety provoking, especially with the death of a left on are kind of trying to find out what those next steps are, like, being able to have that point of connection is actually quite important. Like what's really interesting about this too, is that I saw on your website too, that that stat I didn't know is that 75% of Canadians opt for cremation. So that that's like a quite a high percentage that are that are choose cremation. And so in just kind of thinking about that the idea of kind of culture of what we do after somebody passes away, like what what's your kind of perspective on the relationship between like how technology might change the culture of end of life care, you know, so in this case, I think one piece we're kind of talking about, it's like adding, I think more human in the loop as it were right? Between funeral directors and families. But I'm curious like other, have other things kind of gotten on your radar, your attention in terms of like, how we might shift culture and other directions as well around this area?

Faisal 40:44

Well, what for what's interesting is there. There's a lot and like, yes, 75% of people just cremation. Why is that? Because a lot of there's a lot of population density in cities. If you want to air your loved one in downtown Toronto, where are you going to go? Right, you might go really far. Plus, it's very expensive. And so a lot of people prefer cremation. Now. When they prefer cremation. There's some opponents of DOD going well, you're not going to have like a celebration, right? Like a whole funeral. That's why like, a lot of year early on, when we launched Irie, investors, and when some people might say, well, people prefer to get together and have like a funeral, like the whole service. And our our our belief was, that's not what everyone wants, I'm sure there's a bunch of people want that. But what a lot of people were seeing ones is they want to have a celebration of life, or wherever the college, they're honoring, in their safe space on their own time, that is more meaningful. If you lost a friend, would you rather go to a funeral home and invite everyone there? Or would you rather go to a cottage where you and your friends hung out and have a whole day right? Or even at this year condo where your friend always came by like, whatever, something more meaningful to you, and your and your loved one. And so more and more families are doing that? What I am, and this is like, the you know, this is just wild ideas, what I think will happen as we age, and Gen Z and whatever all sorts of new generations get older and older. What I think is very interesting is we might actually see a lot of like, this idea of even getting together might happen in a virtual space. It might happen in maybe if VR or AR VR gets bigger, it might happen in VR, or it might just happen on you know, oh my then do some social next big social laughter happens where Tao will be more meaningful to them to have a memory online versus meeting everyone meeting together. Plus, we're becoming such a society where my friends might live in New York, or around the world. Right? They might live in India or somewhere else, not everyone might be able to come together. So the one spot that they can all come together is online. And the reason why I have friends all over the world is because of the introduction, the fact the people you've never met. So I think as these generations get older, it might be a different behavior. And, you know, I read whether it really is around 100 years, I'm sure we'll be involved

Adam  43:36

with the new language, LM plus right at that point. Exactly. But I think that's a great point to where, you know, it's it's not so much about, like, what are we implementing, but it's more like how do we get out of your way to help you do what's meaningful to you? Right? And I kind of helped facilitate that.

Faisal  43:52

Exactly. And that's why we were not the most important kind in this one right? When the loved one passes away, why do you have to center everything around the funeral hall, right driving to a funeral home shopping a bunch of places you know sitting there and being told about all these different things you should make one phone call well let the feed let the funeral director take all the details over the phone or you go online whatever you want to do and then go back to doing what's actually important which is talking to your loved one which is you know your family like supporting everybody because that is what's important shall we're letting you be more humans versus being this blur state of like shopping around and a bunch of stuff

Adam  44:38

Yeah, on autopilot mode that doesn't feel good anymore the autopilot not exactly yeah. Do I'm curious like the the the options that you have you got cremation you also have Aqua Meishan which is interesting process. I'm curious like how people think about these how the Is this like a rising practice that people are doing more often and I know because it has it's like a has a sustainable At option two, I guess, because if they're not using, you know, fuels to burn anything, you're using games and kind of a water process, but they actually was fueled don't know what can you explain what that is. And then like, Are people choosing that you

Faisal  45:10

can read all about the technical details on our website, you search up propagation bug, essentially, in a high level, it's 90%. Or, or high 95% mix of water, and a five to 10% Lie, which is essentially, disinfectant. But what it does is it does like the slow, gentle water that breaks down the organic compounds over a course or a couple hours. What that is, it's a very interesting, it's a very interesting way of word cremation, I guess. good cooperation, because what's interesting about it is a lot of people when when I see people choosing here, yes, there's this sustainable angle to it, but they choose for a lot of people or they choose or they're like, I didn't like fire. Right? Like it doesn't actually they, they've passed away. But But like, when sometimes when they prearranged with us, they'll say I'm choosing affirmation as a pre arranged read, because I don't like fire. I want water I like to junkie for water. i That's how I warm to the saw, I want to go. And so we're seeing more and more families use it. Yes, it's still not legal in a lot of places. Um, so more, or we get it, we do get a lot of families in places where it's not legal, calling us and talking to us about it and asking us, hey, you know, what can we do to tell the government to make it legal, but I think this was just like even even launching Irene to Gus almost a year and a half, simply because we had to get licensed and in order to get license. It the regulation, like I talked to you about earlier 100 years, it's been the same? Yeah, I think no, I think the legislation on funerals had been changed for like 20 years. And I saw a big change. So they didn't understand what I really wasn't how it fit in the in the model, it fits, because it's not doing anything different. It did pay me in terms of how a body is cremated, and Trump picked up everything. But it's a different way of presentation, right at using Nina Ricci technology, all this stuff. And so they just didn't understand that they took the baby doll a long time to actually convince them. And David Elgar.

Adam  47:22

Yeah, that's a good point to where it's like, even as we think about when we are innovating with technology, that, obviously you're gonna face regulation and certain and certain things right, in this case to just like, how do you broach that? If it's never been broached before? Right, then, you know, to the legislators even have an awareness of like, here's an alternative method. And we could use, yeah, yeah.

Faisal 47:40

I didn't mean to no fault. They're all because they're doing their job. Or they need to make sure that, you know, everyone, like there's no, you know, the families are taken care of, there's no games, etc, etc. And then you present them and somebody new, it's human nature to be afraid of what's new, but when you are able to break it down and explain that this is why it's beneficial for the consumer. In fact, here's all the checks and balances,

Adam  48:04

then then they understand. I think a big question here, then, too, is what are their ways to actually help automate some government processes? Because obviously, bureaucracy is even slower machine than collecting paperwork? Right? Is that Is that something that we could ever look forward to? I think, in terms of helping increase the speed of that legislation even do like I mean, could could things like that? Do you envision like technology being able to help in arenas like that, that we'd also don't often think about.

Faisal  48:28

I think needs to be a dramatic shift in the government. For that, if you think about all the governments in Leeds, I'm just talking about knows Canada and America, they've grown up in an era that did not, they did not grow up with technology. They had a relationship with technology of law, law, a lot of politicians who are like, in power with Prime Ministers and big ministers, they haven't grown up with technology you starting to see in the past 10 years, all the people that have grown up with technology come into power, but they're not at the top. They're not like President, right, the President is like 80 years old. So but but I think eventually, when you do have these people graduate up and start taking positions of power, that's when technology will become more and more immersed in the government, in turn, be more easier for just regular folks like you and me to entrap the government, because they have enough to six cents on technology. They understand it deeply because they grew up with it. Right? It's not this, like, let me understand how the internet works. It's just like, yeah, what it is how stuff works.

Adam  49:33

This is how I Yes, I've always used this technology, right. Yeah, exactly. It's yeah, but I think it's interesting point too, because it's like we're kind of at that precipice, you know, where we're not totally there yet, but like, we are seeing more younger people entering into government and so we're seeing we're seeing that's gonna take place. I look forward.

Faisal  49:53

I look forward to when I am 60 or 50, that we have a president that was born in the same year. I was hoping that would be very, very cool. Because that person or even someone younger, because that person, you know, would have gone, he or he or she would have lived through all these generations have seen language models would have seen all these interesting stuff happened. And then when they're in that position of power, how they're able to, you know, approving and reject laws and privacy laws, that's going to be very interesting.

Adam  50:27

Yeah, I agree to nothing. I think the thing Thanks for Thanks for for the chat. So far, there's been a really early, fascinating conversation, and I appreciate you kind of rabbit holing with me around the different arenas. Is there any like you know, as we as we kind of think, as a close out here, like are there, I guess other areas that gets you really excited, you know, if you're thinking about where we're seeing technological innovation, or there's like a new thing that's coming up kind of either through Irene or just like other things that are catching your radar that we should be thinking about, like in terms of the humans and technology, and where we might be going?

Faisal 50:54

Well, I've been working on a lot of cool stuff this year, which will, which we'll announce, but you know, what I what I think is very interesting is vision models, and how visual q&a reasoning, how powerful it's becoming. I remember a couple years ago, I worked on a model, I work on something for Boston Dynamics, I work on the spot robot. And what was interesting, the there was a whole I mean, without giving you a NDAs. That was like the idea was Spa has to recognize something. And it needs to read read after recognizing, so read values. And I remember going through that exercise, and I remember how difficult it was we got to work it. But the amount of like, computer models, everything that was required to do it. And then I didn't look out. So who think like Gemini, they Google came out with it, I posted an image. And I go, What is this? And it worked. So well. I like bang, like, Damn, that is amazing. So I am very, very interested in how we can utilize the visual language models or whatever the hell they call much more loving more, more than like language models, just like text input array, and then how you're able to go, how many objects do you see? Well, these are all cool. But like I remember reading an extent when I sent a crowd, and I said, How many people are wearing a red shirt, and it got it really well. And that that specific model is very probable is hard to do a lot of times, so I'm excited about that. I am actually trying to figure out where I can use that. And Irene, I have a couple of ideas, fides, nothing, nothing, nothing on the roadmap yet. Like that's always after like 6pm. That's always my r&d. My r&d time is 6pm time. Okay, great. Let's just tried cool stuff and see what we can come up with. Yeah,

Adam  52:49

no, that's, that's a cool idea. I love that too. I mean, what they just came to mind is that you put me in brainstorming mode. Now too. With that, it's like, if people submit pictures of their loved ones, it may be able to kind of say, well, here's some ideas, we talked about personalities or ways we could talk about them in the obituary. To get some discrepancy.

Faisal  53:04

Just lay that out. That's very interesting. If that happens, I will. I'll leave a little link to you,

Adam  53:11

though, as we discussed here on this podcast right. Now, some of the best ideas, you know, hey, Faisal, it's been really great to talk with you. And I appreciate you hopping on the pod excited to get folks to check out the organization. And both your work and also, you know, if they're in Canada, check in Iran also. And we got we have a good good, good of our listeners in Canada, actually. So this this, listen,

Faisal  53:30

we're making big announcements. I'm not gonna say but you know, we're coming, we're coming to the USA, where will mostly

Adam  53:39

go? Well, maybe I'll see you over in the side, too, then. Well, thanks so much. As we wrap up our conversation with Faisal ibid. It's clear that technology isn't just about innovating and pushing boundaries for the sake of it. It's about how it allows us to remain fundamentally human, especially in fields where empathy and human connection are paramount. The transformative impact of innovations like Irene and funeral OS showcase how technology and artificial intelligence are being leveraged to replace drudgery tasks, allowing funeral directors to connect more closely with grieving families, providing them with more effective care, and caring support. Furthermore, our discussion into evolving language models and AI that the way that these influence our interactions, understanding and trust in technology left me with a lot to think about, you know, as accessible AI tools, like GPT become increasingly immersive and integrated into everyday lives. I think it's essential for us to engage in continuous critical conversations that ask deep questions about the implications of these technologies for our society. You know, it's our shared human experiences and relationships that really, truly resonate in our lives and across those others. And it's refreshing to see technology that is aimed at facilitating this rather than obstructing it. But as we innovate, it's pertinent to ask this as technology allows us to remain and feel more human is that redefining what it means to be human in a tech integrated society. So I want to extend a deep appreciation to Faisal for joining me on the podcast today. really had a great conference. Jason with you and of course to all of you for tuning in, and being a part of this meaningful conversation and adventure into the human experience. If you'd like to deep dive into the subject matter more, you can check out links to Phaselis company and work in the show notes below. And as always, I'm eager to hear your feedback, thoughts, questions on these topics, so please don't hesitate to reach out on YouTube, on LinkedIn on social media on the website. And if you haven't already, please subscribe to this Anthro life so you don't miss a nother episode. And if you found this one valuable, go ahead and share it with somebody that you think would find it valuable as well. It's one of the best ways to help grow the podcast and build an organic community, which is really what you care about, in this case, finding like minded individuals that like a good challenge here and they're helping us find ways to expand our thinking unknown. And as well you can of course, check out the Anthro curious substack blog for more content, deep dives from myself as well as the Anthro curious community. And if you have something that you want to write about, feel free to reach out to me about that as well and see if we can get you over there on our substack thanks again for joining me. Stay curious, stay open and stay inspired. I'm your host Adam Gamwell and you're listening to This Anthro Life. We'll see you next time.