 Boom, what's up everyone? Welcome to Simulation. I'm your host, Alan Sakyan. We are in New York City, New York. We are going to be talking about all things conversational user interfaces. We're going to be talking about intelligent agents and artificial intelligence, saving time. And I'm really excited to have Dennis Morton is joining us on the show. Hello. Thanks so much for having me. I'm super pumped. This is going to be really exciting episode. Quick background on Dennis. Dennis is from Denmark and he's never giving up his accent. That's true. Stick into it. Stick into it. I love it. He's been five years now as the CEO and founder of X.AI. Really excited to break that down for you guys in a bit. Previously, four years as the CEO and founder of Visual Revenue, which was predictive analysis for media and that was acquired. And previous to that, six years as the COO and shareholder of Index Tools, which did enterprise web analytics and that was acquired by Yahoo in May 2008. He's also the author of Data-Driven Insights, which is about collecting and analyzing digital data. So all about data, all about AI, all about making life better for humans. I'm really excited. Let's jump in from a big history perspective on things. So we find ourselves after so many years as stewards. These human animals are now stewards of Earth. And we're at this hockey stick of exponential technology. We're at the hockey stick of population, explosion. So what are your thoughts on this current state? I think it's super easy to be pessimistic. And there's plenty of things which you can let's onto that will suggest that tomorrow won't really be better than yesterday. And you can even kind of fall asleep thinking, perhaps my kids will be okay, but thereafter I can barely see how this is gonna play out. Why? Because it doesn't require any energy. It does require some energy and some willingness to dream up what might be next? Why will we be better off? What are the things that we don't have today that could come along? And that will have to kind of not be lazy. That will force it to kind of read things for where sure, that could also happen. And I'm just in the category where I much rather invest some time and energy, try to dream up if for nothing else, just for me and my kids, some version of the future where we're gonna be way better off. And that's what I let's onto. Doesn't mean that I'm not reading the paper, not seeing the signs, there's plenty of signs, but perhaps those are just signals to then go work on it. You do your startup, clean whatever lake, clean whatever pocket of air over there, do whatever you think you can do to make tomorrow a little bit better. But that's what I'm latching onto. I'm an optimist as you can hear. Yes, very optimistic and very important. I like to dose up on the realism, which is the optimism and also the importance of mitigating the existential risks ahead of us. We had very close calls in the Cold War with mutually shared destruction. I think we're gonna have some close calls again potentially with artificial general intelligences. And so it's crucial for us to evolve as humans that you gave this really beautiful optimistic view of that if we can just really latch onto the essence of unity and of being together here on this planet, we can make sure that our kids and their kids as kids can unleash their full creativity into solving these problems. And plenty of the data we see in aggregate, suddenly on human emotion, if you will, is an aggregate that tilts towards a generation that's seen dramatic change. However, if you do segment out the most optimistic part of that population, that tend to be young people. Actually, those who got the most to lose, why is it that those who got the most to lose are the most optimistic? That should be almost reversed for where if you're 65, you could kind of give a damn, ride this one out and say, world is gonna come to an end. I had fun while here and that's that. No, the very people for where you got another 70 years kind of sees a version which is exciting. That should be encouraging. Forget about what you and I think. Somehow they see something and they're 13 years old. I'm not giving up. You can keep pushing me, but I won't give up on this fact that human ingenuity up until this point have gotten it out of plenty of interesting settings to put it polite in the past. And why is it that this one time or this particular moment is one for where we just couldn't figure it out? Yeah. I don't buy into it. Yeah. And there's plenty of reasons for why anybody who will be watching this will immediately say, damn, Dennis is naive. Perhaps, but there's certainly plenty of people who are naive about whatever venture they're working on and whatever problem they're working on. But I think it's the fact that they are naive. They come up, allows them to say, yeah, I'll lean into this for years on end because I can see a better outcome. Yeah. Now, okay, time is the most valuable commodity that we have and this is what we, we wanna eliminate drudgery from our lives. We don't want to go back and forth 10,000 times in communication with people to try and schedule meetings. We don't want to have to take extra time. Like you have this really good story that you paint out of. You land in your hotel in the next city that you're traveling to. And instead of having to go on your phone and try and locate somewhere to, you know, get your drink that you would like. Instead, you just say, Alexa, order me a flavored water or whatever. And then, boom, just like that, it arrives at your door five minutes later. That eliminates 10 steps of drudgery of trying to look things up on your phone, trying to maybe pick up a phone and call someone and whatever, deal with delivery. The robot will come and deliver you the flavored water. So tell us about the importance of saving time and how these intelligent agents can do this for us. Again, in my naive, optimistic view of the future, we really want everybody to be part of. This is not just for me. This is not for the select few. I think this is for everybody. So we've had this long-lasting promise of us having to work less. Keen suggested at some point we'll be working 10, 15 hours. And I think we are now so far into that promise that we can barely see when and or if that's gonna happen. But I think at this moment, we might actually finally have reached some inflection point in technology for some of the very chores that comes along with pretty much any job could disappear. Now, then we can talk about how much less we will end up working because of that. But I think there's certainly a step one for where you can be an account manager in customer success, in sales, in engineer, a recruiter, it doesn't matter. If you go look at that set of bullets which they hired you on, hey, I would like you to be our recruiter over at Spotify. Here are the seven things that you're gonna be responsible for. Now go look at your inbox. How many of those emails over the last two days can you let on to those seven bullets? Half? If half, that means half your life is shit you just have to deal with to do the fun part for where you find it exciting to be in a recruiter over at Spotify. But they're right over here by the way. And in that regard, I think that 50% is about to disappear. Those are those little annoying chores that are not hard or difficult or require true human ingenuity. It just requires you to do it. You wanna be reimbursed for that bagel. You gotta do those five little tasks. If you wanna meet up with me next Wednesday at one we have to do a little bit of ping pong. And that is just those parts that comes along with every job that I think in phase one will disappear. Now, when they disappear, there's the idea again, if you're lazy that okay, that means then there'll be a pool of people who lose their job because they're not having two people. I can have one person. That suggests you either assume we couldn't come up with other things where we want those people to do or that the very same recruiter just made sure that we ended up with better candidates or spoke to more candidates or any number of 100,000 other things that you can imagine, not just less people. Less people is just such an easy thing to say because you don't have to think about what are the 40 things that could happen now for the recruiter. So again, I think that is phase one. That is about less chores, perhaps the same amount of time spent at the office. Now, this next phase what I would like to see and perhaps now I'm going all in. So there's certainly some setting right for where my dad would take on a job and pretty much stay in the same location, do that job for decades at a time and that's kind of how work or jobs worked at that moment in time. And I think on that vector, we've reached this point for where people in your generation will change jobs every two, three years and it will be normal, not this loyal, just normal. And it's also become normal for where you might not even need to come into the office. The whole idea of the fully depicted team is not exotic, that is becoming the norm, not yet fully in the norm. But if that is a vector, what is the most extreme version of that? Because it looks like we are working less and less time for the same company, we're coming less and less often into the office. So perhaps the future, and this is gonna sound crazy, is really just one for where you and me are more like professional footballers. We join a club, we play some good games, we play multiple tournaments at the same time, I might join another club and it's all gonna be extremely flexible on our end. And I think we might start to not want to separate work and life into these two buckets for where we've been told, for years on end, if you can't separate those two, you might not be in a healthy position. Perhaps we should blend them so well. So when I ask you, you just talk about a life well lived. That's right. That's right. Sorry, that turned into a mini rant there. I love that. The optimistic rant. I love that. I'm a big fan of what you just said, and I think that Richard Branson has also made it very clear that there is no such thing as a work-like balance. You gotta love what you do. You be so obsessed with what you do that you breathe, eat, sleep, love it every moment. It's part of your essence. And so you talk about blending that line, you also talk about how these kind of come in, these next, I like the footballer analogy. That's a good one. When you are going from a big part of future is gig economy. So as you go and you work for a couple years, couple matches, couple tournaments, add a certain workplace and then you go and move on to another one and help them out for a couple of tournaments. Maybe win a championship with them, et cetera. That's a good analogy. I like that a lot. That's a huge part of our future because your creativity can be used in different tournaments. And so that's your job is to go into those different places, use your creativity and then swap, let other people have their creative shots. There's a bunch of effort in wanting to decrease the time vortexes. Like you were describing, there's the calendar, there's the email time vortex, the texting time vortex, the social media time vortex. And so we do things like you and I were just talking beforehand about how we both always have our notifications off and our voice on silent, the phone noise ringer on silent. And that itself frees us up to be able to creatively do what we want and when we need to use the device, we access the device and go outward. So we're trying to eliminate the time vortexes and there's so many of them. There's, you said, 10 billion formal meetings scheduled in the United States every single year. 90 million professionals are scheduling these meetings and there's so many back and forths that take days, you wait a day, they get back to you, you have to look at your calendar, et cetera, et cetera. So, you know, companies like yours at Text.ai, so you're just, boom, you're CCing and a virtual assistant, a personal virtual assistant, an intelligent agent that is able to just super simply and we'll get to the tech stack, but they'll send over a matrix of calendar options to the other person and then you'll get to more easily book a meeting, save time, it's a huge part of our lives. It's just, again, one of those things for where you and me could meet, choose day at four, choose day at six, win day at 11, they're all equally good to me. What's not cool though, is the back and forth trying to figure it out because all I have to decide is, oh, this is gonna be fun, this is gonna be awesome. I do wanna meet with you, but once I've made that decision, I don't wanna participate. I said, making the decision, that's where I add value, figure out one time over another as long as it's within my preferences, just figure it out. And that is where software in so many places at this very moment can take over. I said, I don't need to participate in that. And there's certainly some technological challenges, some social challenges that we need to overcome when we turn a human process into a software process and that's fine. Yeah, yeah. There's something really interesting about how you're starting to describe these tasks as though I'm the gatekeeper of making the decision. Once I make the decision I wanna meet with you, I don't wanna do the busy work, I don't wanna do the drudgery. And that same thing can be applied to all of these decisions that, okay, well I decided that I want to eat a specific style of food. Okay, great, go and find me the best style of food that I selected that I want to eat in my local area. Thank you. There's so many of these. If you're out of ideas of what to do in life or you're working your little list of startups to do come next year, you need to go no further than your inbox. Just look at each one of those emails for where you kinda think a little because you have to reply to it, that's an opportunity because that email shouldn't be in your inbox to begin with or if in your inbox you should have some agent just taking care of that. You might wanna say, oh, but I need to inject one input of mine, oh, then just allow the agent to act on your behalf with that one input. Put, yes. Good, and off you go. Yes, yes. And so then there are ways to design this tech stack that give you more control of the valves of letting the agent run with your calendar, run with your decisions, et cetera. You can turn off some of those valves and gain more control yourself. It's like the access control keyboard. So I wanna talk about the tech stack behind this technology. So run us through this. It seems as though there's a lot of natural language processing and generation that happens here because you add Amy to the email where you want to schedule with Alan and you say schedule with Alan for the first week of March at 200 Broadway and then maybe you'll give an afternoon. Maybe you'll say an afternoon. Maybe you'll say something like that. So then Amy will parse that and figure out, okay, he wants the first week of March and afternoon at 200 Broadway and then Amy will send me the generated language of the options. So you'll generate the email text body saying hi Alan, I'm Amy and then you'll give a matrix of time examples. And so teach us about kind of how the technology works. So I think one way to kind of craft a backdrop for really any agent in the not too distant future is that the agent needs to be able to understand the request and given language is not a solved science, there's no such thing as just writing freely. You can certainly carve out one corner of the universe trying to understand that to the best of your ability and sometimes you would even need to kind of rely on the customer to just enhance that if you didn't understand it in full. Like having to ask Alexa one more time or telling Amy that you meant at the 200 Broadway office not the 15 Broadway office. And if you do understand it then the next kind of step is some sort of reasoning engine for where if I do understand you emailing Amy saying that you're gonna be running five minutes late that's all good but what do I do with it? I said what set of actions do I take? And that depends on the type of agent. If five minutes, perhaps nothing. If 10 minutes, perhaps notify Dennis. If 30 minutes, probably reschedule. And whatever set of scenarios that you can imagine you need some reasoning engine and that will be the same for any autonomous agent. If the car can see that there's a pedestrian in the road you need now some sort of decision framework for where you should probably stop the car or that would be a good idea. Now if you do then have a reasoning engine that can take a set of decisions or kind of fully navigate this universe then the output will obviously be some sort of computational outcome you need to kind of turn that back into language so you can communicate with your constituents. The host, the customer and multiple guests and for that you need some sort of MLG engine for where you can write back in a language and or design that you can understand and that is I think the three primary components. Now the reasoning engine should not just be one for where this is what you said but it should obviously be a reasoning engine that builds on what I've told you before where it says I don't really like 5 a.m. meetings. No I'd rather not meet for three hours straight. I do need a little bit of breather room in between my meetings. When I say the office I mean this particular office. When you send an invite please send it to this particular calendar. All types of things I would want to kind of say once but only once and then there could obviously be certain constraints that arrived from the guest which is I'm actually only in Manhattan next Thursday so I know Dennis said next week but we just gonna collapse that into just Thursday because that's the only day I'm gonna be in the city. So that is I think three components. Anybody who does anything would need to think about it. Now what typically happens here given the complexities around language understanding is that you end up probably over investing in that end and that is perhaps a mistake for where you should figure out exactly how much you want to invest in that so you don't starve your reasoning engine so I think that is probably where we can invest the most right now and gets the biggest bang for the buck and you can certainly not starve the energy because then you end up with really just a bag of templates and that doesn't really do justice to your agent and I think plenty of people have seen this kind of very sparse bag of templates coming straight out of the engineer on one end very little reasoning just some basic logic put in place and then given language is so hard all hands on deck let's work on that. So I wanna give the viewers an example. So here I am using my phone and here on our phone as you'll see okay cool. So here's our message started with Dennis so I'm gonna jump into the text message and when you see the text message you see the options that pop up that is called an n-gram at the top I'm, I, and we right up here and those that is a predictive algorithm for what you would want to start your sentence with and every time that you click on one of those not only are you training the algorithm every time you click it but you're saving time that's the most important thing and so we will get to voice in a bit I wanna get to voice but I wanna go through some examples you know you talked about the reasoning engine so that's the processing of the language that I would give to you in the conversation I would actually and that's the geek in me now I think it's probably important to separate the actual ability to understand language and see that as a separate NLU challenge when you do understand it then you need some ability to reason about what you just understood as in I have a funny Dennis accent but if I tell you I have to kinda end this at six o'clock you now need to figure out what do I do with this information Dennis said he had to run up to a meeting at six but what is your next action you did understand it but you might actually not know what to do with it interesting so this is broken down it's too very kinda separate kinda challenges so we have a natural language processing then from the processing we have to understand the information and understandings like you just said that I know that at six but then the reasoning kicks in which is okay do we need to speed things up how do we need to move next can we push it back 10 minutes whatever we need to do and so then those are the next options and that's right now a really tough place for this sort of reasoning engine this is difficult to build and figure out how to best move forward when you're kinda playing with a different situation so here's a couple things I wanna maybe hit back at you so you give these examples of people sending very weirdly worded messages cause I work a lot at one or two a.m. and I always say like if I'm meeting with them that day I say today because I send it at two a.m. and so when they open it they'll see it as today but some people will say tomorrow and that will throw things off so there's these weird quirks that we have and also there's a you gave this really good understanding that you can say that you always wanna meet at 200 Broadway you always want to never meet at five a.m. so you can give these sort of basic levels of code that then the algorithm can understand you can add input to the algorithm so it can better understand what your preferences are and so there's this really interesting interplay then between the amount of control you have with the scheduling, with the intelligent assistant and also the amount that the intelligent assistant needs to reason and understand the situation and this adds a very interesting degree of learning that's going on with these machine intelligences. It's a super interesting time right now for where we've been able to get access to ever more data and some of the models that we weren't able to really get to perform at what we had hoped for perhaps 10, 15 years ago are now performing at a very impressive level and that allows us to kind of truly train on things and see systems become way smarter tomorrow versus yesterday through its sheer use of the system that is certainly extremely exciting and something that we are not betting the farm on but building a company on which is that the more meetings we schedule I think the smarter we really end up being at scheduling meetings so whatever version of Amy that you might have touched in the past if you touch Amy or Andrew tomorrow it's really two different people or entities. Yes and this is much more cost effective too we're making things cheaper we're making them faster and better and so you rather than needing a $45,000 of a year personal assistant you pay eight bucks a month to use something like x.ai. And here I think again if you wanna bucket things there's a whole set of agents for where it seems obvious that they might replace a function that a human might have done previously but then there's also in the other bucket a whole host of functions for where we take something that used to be a luxury and democratize it. So here's the thing that 10 billion formal meetings being set up in the US every year less than 0.1% are being set up by admin staff as in that idea of the personal assistant you know, because you've seen it on the telly but they don't really exist in real life sure, yes in Manhattan, in finance here and there you spot them but they don't really exist as in almost all of us set up our own meetings. That's the norm. That's not about replacing something that used to cost 45K that is about you being punished in such a sense where you would have to sit at 11 PM in your underwear at home fiddling with your meetings and now you don't have to do that. And I think those are two very different things as in that self-driving car doesn't replace a chauffeur that I used to have hanging around my apartment. No, I used to just drive myself, right? So in that regard, I think we need to kind of separate those two, if that makes any sense. Yes, absolutely. And you gave a really good understanding there just the percentage then it's 0.01% it's super low that the actual people that have a personal assistant how many people have a personal chauffeur how many people have personal cook personal cleaning staff those are just luxuries for where I certainly know why they exist I can certainly imagine me having it but I can't imagine me paying 60K a year to have some personal chef hanging around the apartment pouring cereal for me now I'll figure that out myself but there might be some machinery for me. You know what? For me to get this juice type drink I'll buy a little machine for that then I don't have to walk down to the street and I can have a slightly automated setting and tomorrow it'll be a little bit better. The whole kitchen, you get to pick salad pasta, whatever you like and then boom, boom, boom, boom. The kitchen is such an exciting place right now but we haven't really seen much change we've seen a few machines a few machines enter but they've kind of stayed the same they're slightly sexier but they kind of stayed the same you and me should start a venture on that. I have two interesting questions that I think you'll so when first, first is do we is it as simple as just syncing with my Google Calendar and then it's red? Yeah, so and those are my availability we schedule meetings for a living that's what we do no more, no less the way it works is that you sign For now, for now what's the next endeavor? We'll figure out soon. Well, we could move that over but we're really a hyper focus on meetings I just want to make sure that if you pull out your phone right now look at the events on your calendar and I didn't put them there then I got work to do that's my goal in life now you sign up you connect your calendar so that we have visibility into it when you're free when you're busy and so on and so forth because if you don't then you hired a blind assistant that wouldn't work and the way we get access is through basic Google and or Microsoft authentication we do not ask for access to your inbox that means we can't see anything which you write to anybody about but when you do want to wake up the agent you simply CC in the agent as an email address there's nothing you've installed there's no plug-in or extension or add-on it's just an email address like you hired somebody and if you want to hire Tom Tom doesn't come with an app Tom is just Tom and you write Tom and tell him hey Tom I want you to do three things for me and it's the same with Amy and Andrew you just write hey Amy can you get me and Alan together on 200 Broadway for half an hour come first week of February and you click send now the agent will then read and understand what did you just ask it to do read that back to you so you know that it is all good and kosher and then continue by removing you from the conversation reach out to one or more participants negotiate either through some visual UIs text language you're clicking onto a website and picking your time you're connecting your calendar there's multiple ways but that doesn't really matter what matters is that we are kind and polite and efficient towards the guest and that you don't have to be involved in it and that typically takes if it's two humans about 1.9 day on average not because it takes that long to write an email but because I send it you're working you write something back I'm off working on something and that ping pong takes shorter two days here though you click send then you click archive why? because this is not your job anymore and that I think is a really important kind of distinction here there's plenty of software that will augment your work make you a little bit faster a little bit more accurate but you're still in charge it's still your job so if you do some sort of financial model for your startup for the year 2020 that's your job and you'll certainly be using Excel for it but at no point when you present that to your board will you say hey me and my pal Excel have been working on this for the last three weeks no you did it and this type of software and there'll be many more of us is software where you kind of describe the objective don't get the system in the task describe the objective hand it over and say you do it yeah if there's some sort of obstacle let me know I'll help out but I want to make sure we agree on who owns this and you do Amy or Andrew and that I find really interesting and that is certainly software that kind of excites me that I can start to begin to describe things in objectives and not little features yeah they're very easy to use and time saving this is the kind of stuff that is big part of our future now what happens if I also have x.ai and you have x.ai then can they go back and forth so this is what we talk about around the office as scheduling nirvana writes I've just ceased in Amy asked her to put something on my calendar for first week of fit between you and I you reply back saying ah awesome I've ceased in Andrew on my end he can hash it out with Amy but if Andrew is Amy who the hell is she talking to what we don't do which will be funny ha ha perhaps is sending emails back and forth between ourselves what we do is some sort of internal machine language and it turns into a preference negotiation where we maximize the happiness on both ends so we both adhere to your likings and my likings so in that little Venn diagram of one or more people we find that kind of optimal spot for where ah yeah next Thursday four o'clock meeting room 9F that's what we're gonna do so that is where this particular type of product many similar products have real massive network effects for where the more people using it the better is for the whole population yes yes and then do you have specific staffers that will train the machine learning algorithms to deal better with the replies that you get so let's say you get the very obscure replies from me and then you have to now try and train Amy to get better that I wanna oh well I replied saying I want 60 minutes instead of 30 and I replied saying that I need a well lit quiet space and then etc so tell us about that so we have about 70 people right now that does nothing but data labeling and that means the more data we have the better Amy and Andrew and the whole system generally performs so we invest heavily in that seven zero seven zero yes doing data labeling that is crazy data labeling only yeah that's crazy and that just if anything sends a signal about the value of data in this day and age yeah that's cool so then you are you are training the algorithms through this data labeling and then it's better understanding as the information comes in how to best reply and then on that now you could turn this into a four hour kind of mini seminar that's the that's the very kind of part for where you want the most robust data set possible so that you can make the highest or most accurate predictions possible but that's one part we're also in somewhat uncharted waters right now for where the conversational UI is still so new that we don't have ten years of best practices to how best put together an agent so if you and me kind of got together this weekend to hack together some app for some funny thing that we came up with we'll have some idea of both who of our friends to call how to put it together so it looks like an app for where if you open it up you can almost immediately know how to use it that's not the case with the conversational UI we don't have really decades of experience we've had some voice systems in the past that I'm not so sure fully carry over to this new conversational UI paradigm that we're in right now with Alexa the Google's and the Cortana's and so on and so forth and that has been super interesting and we've made a ton of mistakes but I think we are slowly starting to figure out exactly how to best design an application within that UI framework awesome segue into the future of voice so teach us about what your thoughts are as we enter into this realm of having an emotional connection with software of all different sorts that makes our lives easier so I suddenly think it's important again to make a distinction between the agent and the UI where the agent might live so people might mix these two together so the conversational UI is just a UI just like the command line is a UI or the graphical user interface or perhaps the mobile UI is a distinct UI in its own right but the conversational UI is just another way to interact with your user and there's plenty of things that can happen that UI that does not have to be intelligent or does not have to be an agent and it could be as simple as a city bank sending me a text message asking me whether I did this transaction yes or no it only takes two inputs one for yes two for no and that's that and it's an awesome use of that it's only way better than trying to call me like they did five years ago so that is a distinct UI now the reason that we tend to kind of collapse those two is just that the agent has a very good home within the conversational UI because when you want to describe objectives describing objectives is kind of cumbersome in drop downs and check boxes and radio buttons but it's very easy in language hey set something up for some afternoon last week of January at my office okay exactly what drop down some check boxes that I just described there but it was very easy for me to kind of just say out loud so that's why the agent has a good home there now the agents that we will see in this space are kind of right now all over the place from little kind of toy things that shouldn't even be an agent or shouldn't even be within the kind of conversational UI to more kind of ambitious settings for me and my 100 propeller head friends have been working on this problem for five years right and everything in between right yeah there's some voices such a big part of the future to almost to the extent where we might not even learn typing likely in the future this is crazy that we're gonna be interacting with these spatial computing technologies and that I want to just facilitate most of the things that are drudgery in my life through these artificial intelligent assistance you we were talking just as we as we entered in here about this journey of entrepreneurship you just mentioned the five years of hard work with your team the you have these peaks of really excitement you have these lows of like of sadness and anxiety etc. Now there's really no like I'm good in the like neutral state of entrepreneur it's like riding these crazy waves and I'm I'm glad that you're that you're building this I want to get a couple quick principles on the way out entrepreneurship is obviously super important go and execute always okay what's a core driving principle of your life be happy when I get to the end it sounds sad and it isn't and I sit and look back with my kids and grandkids all I really want to be able to say is that was fun wasn't that fun we had fun didn't we doesn't matter what car what apartment how much money I made did we have fun yeah that was a good ride and then I can go to sleep and say it was all worth it that's my life that's all I'm driving towards I love it I love it and then how about we ask you if you could rebuild civilization how would you design it so I'm Danish or as my American friends call me the socialist so immediately I run away screaming right now there's plenty of things that did not work in that system which is why I'm over here such as the celebration of entrepreneurship risk and reward which works extremely well in the US but the decency and safety that comes along with certainly what we built back in Scandinavia isn't as dangerous as some people might want to make it and there'll be plenty of things which I will be taking from that version but I would try to pair it with the lust for risk reward and the celebration of entrepreneurship and joy of other people's success that you really see over here and if there's some sort of amalgamation of those two things that'll be my version what a pairing I love it and then the last question is what do you think is the most beautiful thing in the world young people dreaming of a better future that gets me excited even silly things or things I seem barely understand as in why is that important but I can see somehow it's super important to them that's something we just shouldn't kill we should make sure that there's room for them to dream big and dream about tomorrow and anything which we can do to support that lean in very powerful answers Dennis thank you again so much for coming onto the show and talking to us I'm really grateful that we had you we should do another I know make a repeat of that we'll do it once a year we will yeah we will you're welcome in our studio in San Francisco anytime I'll let you know when I'm coming to New York next everyone thank you so much for tuning in we greatly appreciate you let us know your thoughts in the comments below we'll be unpacking all of the nuance around these intelligent agents more as we keep doing shows and also go and check out x.ai the links in the bio go and check them out they have a 14 day free trial give it a go and remember the more people that are using this the easier it is for us to schedule meetings with each other and lastly is continue supporting great organizations like Dennis's and ours and join us links below as well so we can continue sustaining this and growing it and go and build the future everyone manifest your destiny into the world much love and we'll see you soon cheers cheers