#193 Innovation, AI and Startups: A Conversation with Zach Rattner

0:00:02 - Mehmet
Hello and welcome back to a new episode of the CTO show with Mehmet. Today, I'm very pleased to have with me Zach, joining me from San Diego. Zach, I like to leave it to the guests to introduce themselves, so all the floor is yours to tell us about yourself and what you do. 

0:00:20 - Zach
Sure. Thank you for having me, mehmet. My name is Zach Rattner and I am the chief technology officer and co-founder at an AI company called Yembo. I'm an engineer by training and I work on leading the engineering team now. So what we do at Yembo is we build computer vision algorithms to identify items inside a person's home and we provide services to companies that provide home service jobs, so primarily property insurance companies, moving companies In all these home service spaces. It's very difficult to provide an accurate estimate because it takes trained expertise to do it, and we provide computer vision based tools to allow these companies to do things in a more reliable, scalable, fast way. 

0:01:05 - Mehmet
That's cool. Thank you again for being on the show today. So, zach, I liked a little bit prepared before the episode and I was going through your experience with scaling flashcards, studying tools to over 3 million users to shape product development at Yembo. So can you tell us a little bit about this experience and what actually brought you in this space in the first place? 

0:01:35 - Zach
Sure, yeah you did your homework. I guess I would say I've always been a bit of a nerd. One thing I always asked for as a kid for my birthday was we had a shared family computer so I had to share time with my parents and my siblings. I'd always asked for more time, and if I didn't have the ability to make games and these kinds of things as a 12 year old, I ended up buying a legal pad paper at the local dollar store and I would write my code on the paper, waiting for my sister to get off the computer and then as soon as I got in, I'd type everything in and see if my code worked. So it sort of felt like a blessing when I realized later in high school and I was looking at colleges that this is actually like a profession, it's not just like a hobby, and I feel like that arc has been consistent throughout my career. 

Personally, where this flashcard tool, I made something for myself. It was kind of difficult to organize all your different courses and build a schedule. At Virginia Tech they would have these huge entry level like chemistry 101 kind of courses and they're very competitive to get in the 10 am class. No one really wanted to get in the 8 am class, and so I built this tool for just mostly for myself, to help. I'm a morning person, want to pack my schedule in the morning and then be done with classes by like 11 or so. And when I built it I realized that there was this local startup that was building academic services for college students and I just I wasn't like trying to sell it to them, I just said, hey, I know you're in the space. I did this. It helped me. I don't really know what to do next. Like should we talk? And then an internship came out of it. 

And then that's when I was able to build that flashcard tool and then had an opportunity to join a large telecom company named Qualcomm right out of college, and that was great Kind of got to. I felt like I grew up super fast there because what I used to do at a smaller startup, like an intern, can just push code on production there's, you move really fast, but at the expense of maybe some oversight. And when you go to the larger company it's like so many, so much scale that like literally billions of handsets using your code. So if there's a bug, somebody's going to find it, and that was great learning experience. 

Great time was right when 4G was coming out and then, after a few years of doing that kind of left started doing some consulting work. And then the idea for Yembo came and Yembo started as an expensive hobby. It was like we had the idea. We started working on it. I wasn't really sure, like, if it was going to. We had some things that we needed to prove around the technology and we didn't know if it was going to be viable for the current state of the art at the time. 

And kind of one thing led to another. We had smaller successes, that kind of snowballed into larger ones. And at some point my wife told me, like I don't think this Yembo thing is like a hobby anymore. Kind of just looks like work, Like it takes your day from nine to five and you talk about a dinner table and it's kind of like grew up on its own. 

0:04:44 - Mehmet
Yeah, that's nice, and I think because when you become passionate about an idea, it goes out from a hobby, I believe, like it becomes like feeling your whole, I mean life. 

Now, one thing you know, and this was interesting for me so you have 18 granted us patents. So what inspires your innovative thinking? And I'm asking you this question because sometimes I hear from people you know we're not out of ideas. You know we don't know how to innovate, and you know some people out of, I would say maybe. I don't want to call it laziness, but you know we are all lazy by nature as humans, so we try to copy something instead of, you know, trying to get something out, you know, from, from nowhere, to something that might become, you know, a viable product or so. So what inspires you, zach, to really get these ideas and try to test them and build them? 

0:05:52 - Zach
I would say having diverse experiences that are not directly related to the area you're working in is usually where that creative spark comes from for me. So the interesting part of you, if you zoom into those 18 patents, they're all over the place. There's a few wireless cellular ones from my time at Qualcomm. There's a few beauty supply ones, which I had an opportunity to work with a beauty supply company. So I'm actually even though my hair has looked like this since I was like a teenager I'm actually named a dryer patent, and then more recently, some in computer vision. But I feel like the the most interesting pieces are looking at how other Areas have solved problems and use this concept called first principles to kind of just figure out for yourself if something's going to work or not. And I think the best example of this is the underlying technology that powers the, the nacho heating elements like the melted cheese. 

If you study that, I had a good reason to study that at a point in my career and that ended up serving as the innovative spark that kind of made a battery powered hairdryer possible. But we were looking at how everyone else was heating hairdryers at the time and Everyone's kind of doing the same thing, and if you wanted to build something based on battery, you really need to innovate around power. And this was one of the the projects kind of in my in my days before Yembo really took off and it was just interesting that, like I started studying, it started doing this first principles thing and then ended up literally talking to a Nacho cart vendor around power requirements and that was like the, the spark that made it all work. So am I an expert in hairdryers? Not really am I an expert in nachos. I mean, I eat them, but that's about it. But, like, finding the intersection of these different areas always seems to open up kind of Innovative solutions that people haven't thought of before. 

But if you try to do exactly where everyone else has been doing, it's not gonna end up being super innovative. 

0:07:50 - Mehmet
Yeah. So I believe this is the you know part of many people like yourself, zach, who are always on the hunt of Solving problems, whatever it is, and you try to see if you can apply your knowledge to solve this, and you know innovate around. So, which is which is a great way, because you know we repeated on the show multiple times Like the best way to come up with a good idea is to go find problems actually, exactly, yeah, so, and it looks like you did it fantastically well now, and you mentioned something when you were talking about your, your current startup, and you said, like you were trying to, to, to build, and then you get some some you know Use cases which were which were like, yeah, but it seems something that is Needed outside there. Oh, what methods you know, do you employ to validate a product quickly While minimizing both financial and time risks? 

0:08:53 - Zach
So I think I've done this the right way in the wrong way before. So if I can be of any help and maybe sharing some, some ideas of what to avoid and what to do, that'd be great. The key that I found is you mentioned this before you hit the nail on the head. As you mentioned, you're solving a problem, so there's usually I would say always should be an idea in your mind of what is that problem You're trying to solve, because we don't want to just talk about like Cool things. Wouldn't it be cool if we could do this? Could we do that? Because then, because then you don't know if it's actually going to be practical or not. So the framework I like to do is start with a hypothesis and, if you remember back to like middle school science, you have a, an explanation of how you think things are going to work that you're looking to test. 

So when we were starting Yembo, the actual first hypothesis had nothing to do with AI. We weren't sure would people be willing to click a link and record videos? Because if you aren't willing to do that, then it doesn't really matter how cool the AI is. You didn't get to that point, so that was like hypothesis number one and what that led us to do was it led us to test it without even building any AI. So we just started and said found a moving company and said, look, I can save you the drive time. The product wasn't super valuable there because it was like something I had hacked together in a couple of days, but we were not looking to test the AI yet. We were looking to test will somebody be willing to go through this process and is it Compelling or is it like tedious and you can kind of convince them under the right conditions. Then then, when we found we're saving the drive time where you don't have to do a non-site visit, we were saving convenience if you can record at your own time, you don't have to schedule a call and we found like people overwhelmingly not just were willing to do it, but preferred it to the previous version of setting things up in your home. 

And and If we had just been like, hey, we want to start an AI company, let's go build AI it would have cost millions of dollars more. We probably would have run, had a money and wouldn't have even validated if this was like a, a market worth doing. So I would say always have that hypothesis and Don't take your ideas personally. You're inviting criticism. At that point in time I was looking for a reason for it to fail, because I wanted to be convinced that this was something that was worth spending the next. I guess I'm seven years into this now, so I didn't want to lightly go and make this decision to spend a bunch of effort on building something. So we put down like what are all these things that would have been fatal if they don't work, and we just tested them up front in in very low stakes environments. And then you kind of know you're working on the right things then, because now you you've understood that you've kind of de-risked the, the most important parts. 

0:11:28 - Mehmet
I think you know also you nailed it, zach about the hypothesis. You know, like you, it's assumption. Like you do, you don't Know. If you don't actually go and test it out, you will not know what the result is. And I'm repeating this couple of times on the show now, and Because you know I work with a lot of startup founders. 

Some of them they have ideas, some of them they are still, you know, in the ideation phase, and I always tell them guys, okay, why you are trying to find a problem for the distribution you have, you should do it the other way around, because otherwise you know, like, as you said, like your ego, accepted or not, like would say no, no, no, this is a good idea. I cannot do it this way. I'm putting AI in this, so it will be more fancy. And you know what, like it doesn't have to become always you know fancy thing to become Successful, which is 100% now In a, say, startup environment like Yimbo. What specific skills? Because you know, like you are the CTO, but what other specific skills you believe are essential For collaboration and how do you foster that with with the team? 

0:12:42 - Zach
I think the most important part about having Good collaboration on your team is great communication skills. I can be the most brilliant engineer in the world, but if I can't clearly communicate my thinking and my thought process to other people, then it's not going to work that well. So we hire. In our interview process, we screen for communication. We generally have the candidate present something small, something that they're familiar with. We're not quizzing them at that point on like the specific details. We're just trying to get a sense of their collaboration skills and their communication skills, because I think the best AI products are also humanities. 

Products Like these are tools that are used by somebody, and if you look at our team structure, we don't just have a bunch of software engineers. I joke that if we wanted to go into like an easy part of software development, we picked the absolute wrong industry. Because to do well at Yenbo, you need to have cutting edge AI. You need to make it accessible to an audience that hasn't historically used it before, so you need good user interface design and user studies and user flows need to be like factoring all these different kinds of things in. And then, because it's a web product, thanks to these tech giants like Google and Facebook and Netflix, people expect it to always work and never go down. So you need really good software engineer infrastructure. Any one of those things you take out, it just doesn't work. So we look for people with diverse backgrounds. We have lots of people who have transitioned into software engineering, so they had some career beforehand that they can fall back to, which I found is really interesting when you're debating, like what defaults should be, what the user interface should be like, because we don't just have like a bunch of computer science graduates who have the same amount of work experience and all that kind of stuff coming in. 

And I think the next part is hard to kind of engineer. It just kind of happens. If you build your team, right is you need this concept of psychological safety where you don't feel like attacked if you suggest something that gets told no, because when you're brainstorming, right, you want a lot of ideas, you want to consider a lot of options. I say a lot of things that get rejected, but you want to push those boundaries, you want to consider more extreme versions of the things you end up building, because that's how you kind of find that sweet spot in the middle, right. So I found that being able to have a lot of different ideas with people with very different backgrounds from yourselves, have the safety to be able to critique each other's ideas and not have that get misinterpreted. Like you can critique my idea without telling me that I'm failing in some way and then having everybody convinced. 

Then, when it comes time to actually implement, like everybody knows that there's not like a big unforeseen thing that we like had in the back of our minds that didn't speak up. Like people are more committed when they're doing it. So like there's a lot of debates in the early days when we're starting something, and I think that's healthy. If you have like one person who has all the ideas and everyone else is just kind of like sitting tight, I think that's not really what's going on. I think that's it means people are probably not willing to speak up. I think it's unlike that. People are just nobody has any thoughts. So it looks a little bit messy if you're an outsider and you walk in and see what's going on, because it's like why are these people debating so much? But I think you have to because we're building things that are at very big scale and people are going to be using it in a lot of different environments, a lot of different situations. So you need to think about all these different considerations when you're starting out. 

0:16:22 - Mehmet
Yeah, perfect, I think you know this is also. We see it in large. I mean, once they were startups, actually, and I think they kept this. So, you know, keep it open for ideas. Company like large companies. They can afford even trying to build, you know these ideas and they might fail. This is what you know. Google, they did it, amazon, they did it. 

You know, and they keep it in this culture of, okay, there's no right or wrong until you know we try actually to achieve it, to see is it like something logical, is it something that really can do an impact on the company or impact on our customers. You know like, so, yeah, like, and I love you know this kind of cultures. But the question here for you is, like, talking about culture Now, yourself, you are a CTO and you know how do you think with. You know it's critical to keep the culture, and how can you actually keep it while growing the team? Because we know when the team grows, it becomes like a little bit. You know you have, like now, managers and maybe you have to delegate to other team members. So how do you keep this balance, I would say, between keeping the culture and delegation and keeping also what we were just discussing. I mean, everyone can share their opinions without being afraid of being hammered, as you said. 

0:17:47 - Zach
Yeah, I think things will change. I mean, a company of 30 people can't operate like a company of five. But to your point, as in leadership, you need to be responsible for the culture and it's something that you have to work on, just like anything else you don't. Maybe in the early days you kind of get it for free because the company is so small and everyone is just kind of on the same page. So I don't know if I really buy that, but if you were to make that argument I'd give it to you. But certainly when you, when you have more people than you can talk to in a day at the company, it's something that you need to work on and there are a few things that we've done to try to make it work out faster. But I think at the end of the day it comes down to the people, right Like you can't change somebody's personality when they join, so it more comes down to finding the right kind of person, which means articulating what that right kind of person is and maybe having one of those debates we're talking about with your team, because that also changes. I found in the early days you want like a pioneer, an innovator, who's okay to maybe not have a ton of infrastructure, not to have a ton of guidance and maybe a hacker type in the engineering world, but then as the tool grows, people start relying on it. You need more stability, you need more reliability, like that. Personality of the kind of people that you need will will change and I think that's healthy and that's okay. But I feel like the key is getting to that psychological safety where the team feels safe and able to collaborate as quickly as you can. So we have a general guideline for all managers on our team where we say, when you hire somebody, try to have their first win within 30 days. So for an engineer, that can be code in production so they can actually see the results of I came in, I did X and now, like customer Y can do something new they couldn't do before. For someone in the design team, we aim to have their their like workflow accepted into the, accepted into the R use stigma. 

But the idea is like nobody really likes to be the new person, right To have to be like least less experience than everybody else, and when you join you kind of have to be the new person. So what we do is we kind of go out of our way to minimize the time you're in that spot, because what we realized is we spend all this time and energy recruiting the right kind of people. So there's some really awesome people who join and then when they join, you don't want to like waste their talent by having them sit around and like read a bunch of onboard documentation for 90 days and then like not really have contributed anything. That's on you as the employer, not the employee. If you're not giving them like a clear path to kind of establishing their credibility with the team and we found this works pretty well is what happens is 30 days isn't a lot of time. 

So if you're a manager tasked with this, what are you going to do? You need to find a project that is uniquely tied into that person's skill sets, because you don't have time to train for anything, but normally you've hired somebody because they have that skill set. So what you're doing is you're forcing, thinking through how can I most effectively plug this person into the existing company and then have it have a result, and then people on the team can see oh, this person's an expert, they just did this, they just did that, like that's, they've got some wind under their belt and we found that that makes the next project that you do just go that much easier, because now the person is a known quantity, they've contributed something. People know that that area that wouldn't have gotten done otherwise if they hadn't been there. So again, you can't. You can't change people's personalities, nor should I, nor should you try. But I think you can go out of your way to make an environment where they're able to become productive as quickly as you can. 

0:21:27 - Mehmet
That's great, great Zack. Now let's talk a little bit about, you know, yembo, right, and let's talk about technology a little bit. Sure, now the issue nowadays that when we mentioned the old AI, you guessed what people think. Right, so they think chat, gpt. But I think there's a lot of areas like actually it existed, but now, because all people you know, they shifted their focus on large language models, you know, and everyone started to talk about this. But let us remind you know folks out there, even technical folks, by the way, not only like normal people like what first tell me you know what kind of AI you are using, like on a high level, of course, right. And then you know where do you see, you know like this can be applied in maybe some other areas that you know like no one is talking about, maybe now? 

0:22:27 - Zach
Yeah for sure. I feel like when I started Yembo, when you said AI, everyone thought about Terminator and everyone thinks about chat, gpt. I don't know, maybe that's a step in the right direction, I don't know. I think that at a super high level, yembo is more focused on perception tasks than generative tasks. So things like chat, gpt or stable diffusion where you enter in a prompt and you get something back, where the AI is creating something, that I mean it's a powerful use case with a lot of applications, but in our industry specifically, where we want to reliably detect what's here, it's not super valuable in its current form right now. So we focus more on things like object detection and tracking and localization, being able to identify. If you were to scan a video of this room, like how many boxes would I need to pack the books there? That's a question that, like some people spend 10 minutes on over the phone. If you're a move estimator and I'm moving and you can't see because we're just talking on the phone and you're asking like well, how many shelves are there? 

Okay, how long is a shelf? So those are the kinds of things that we focus on at Yembo, and we achieve it using more detection tasks than generative tasks. Now, there's a lot of interesting things going out there, I think. Things like a lot of use cases around by getting ideas for marketing, copy or customer support tools. So I mean, it's always like fun to experiment and play with these things, but my view with them is, as long as the AI makes mistakes, then we'll still need to have people involved to do these kinds of things, because the thing to remember about all of these algorithms is, I would say, they're biologically inspired, but they're not biology. So the way they work is like roughly mimicking what neuron and brains do and things like that, but they are not human brain. 

So I had a friend send me I forget the name of the tool she used, but she used one of those AI image generators and sent me a picture of herself and said hey, can you update your contact photo in your in your phone of this? And I did it, and then she texted me the next day and I was like I'm sorry, I can't, I have to have to change it back. You have three front teeth and that's all I'm looking at. I can't see anything else. It's like the AI it doesn't. It doesn't inherently know like a person would, so it makes mistakes that a person wouldn't, because it's not really reasoning, it doesn't have intuition, it's just modeling based off of what it's seen before, and that it's not really like a bug. You can fix that easily. I think that's kind of a fundamental property of the way that these models work and that makes it applicable in a lot of spaces, but also makes it not so applicable in others. So I feel like people's imaginations run wild, they try to do things with it and then they have some mismatched expectations sometimes in this generative AI space. 

But I think overall it works great as a tool to in use cases where you can afford to make some mistakes, like if I'm having a, if I'm having a block and I want to say, hey, let's get 10 logo ideas and like it's okay to just be like rough sketches. I think that kind of stuff, sure, if you're like a designer and you want to get over the creative hump, then that that makes total sense. But to just like I just tried, I was curious. I wanted to force myself to like, imagine I can't talk to a designer and I want to just draw. 

I think it was some simple icon. I was trying and after like 45 minutes of trying to draw like a little icon of the front of a house, I just gave up. I was trying all these different prompts just to to see like if this is debuggable. And I think what I've concluded is for the state of the art right now, and the space is changing really quickly, so maybe the state of the art now is out of date by the time this podcast airs, but for the state of where things stand right now, it's great for these tasks where you're earlier on in the process. If you want 10 ideas for blog headline, sure, but if you want to say like, write me a really engaging blog, copy paste. I don't. I'm not sure we'll ever get there. 

0:26:31 - Mehmet
Yeah, for me, like I would say, it helps me in automation. You know, for example, if I, if I, if I'm doing something manually that I can I know that I'm a fee can do it, so I can give it to AI because I know it's going to give me the same thing that I say. For example, just to give an example, the other day I was preparing kind of a spreadsheet and I know that, exactly the format that I want, but I don't want to copy paste. You know the field and so on. So I just I said, okay, I need this and I need you to every time rename it in a certain way. So basically it's like kind of programming, but on a high level. So I think and you know, the other day also we discussed, like chat GPT is like you know, think about it, like people who doesn't know how to to code, for example, maybe it's like same as the. There's a sampler and compiler and you know all these. So now you have you have chat GPT, which is like a layer between human and the machine, you know, so it can turn things. So I'm using it this way and although, like I'm a technically guy, of course, but I mean it. It helps me in automating few things. 

Text generation, I agree with you, you need. You need to correct a lot of mistakes. Sometimes it hallucinates. Sometimes Art perspective I cannot judge, but I think I would agree with you because you know, I think you know even one of the pretty much guys who you know generate a lot of money actually out of it, because he was using one of the engines and he was, you know, sharing how he was trying to fix the hands, because sometimes you see like six fingers instead of five. 

Yeah, so 100%, but what you know. Coming back to what you do, like at Yembo, so can we say and this is also because I'm curious about the technology, because you gave me a use case for example, someone want to help you in moving house and then you start to take, maybe, pictures of your rooms, and you know your living room, your bedroom, you know all this. And then they will come up using your technology and say, hey, you would need, for example, one truck and we need 20 boxes, right? So can you blend this also with something like, I don't know, like augmented reality, virtual reality? Is there anything that comes into the play here? 

0:28:59 - Zach
Yeah, I mean, we're building a lot of different things like this, so we are experimenting with all of these. The key for us and this is kind of an unusual use case from what I've seen out there is, in a sense, we're kind of like the opposite of a sticky app like Netflix or Facebook. People don't use our software, fall in love and then decide to move every day because of it. So we need to design for a first time user who is doesn't have context, who's at a point in time in their life where they're really busy, like no one's bored, two weeks before they move right, they're always got to, they're always have other things on their mind and they need to understand it well enough to get the job done, which is actually a really tough, tough thing to do. And we call ourselves we have this phrase at our company we call them geek users and I call myself one. Like I've yumbled this room hundreds of times so I don't trust myself anymore and like oh yeah, that makes sense Because I have the benefit of being sticky right, I use it all the time, but most people don't. So we experiment a lot and we do a lot of user studies, but we also discount ourselves, which is kind of interesting. 

And we have to go and look at first-time users and there are platforms and tools that you can use that help you go find a first-time user that's not particularly conversant in your tool and that's what we really use. 

To judge is we measure things that are quantitative, like how long did it take to go through, how many times they have to click the back button or call up for help and then also some things that are more qualitative, like asking how was your experience, how did it compare to what your expectations were, all that kind of stuff. 

But the whiz-bang crazy over the top like AR features or things that are really technically complicated, oftentimes don't work out that well and if you think about it, it's like the person who's using it it's a tool for a job that needs to get done and they don't always want vibration and sounds, playing and lights and all these crazy things going on at once and even though from an engineering standpoint, it's really cool to be able to coordinate and unsynchronize and do all these things. So it's interesting. I feel like my intuition has gotten maybe better, but I'm still about I don't know 50% right when I'm thinking through something because you need to put yourself in the mind of the actual person using it. And we always joke like, hey, there's 10 or so of us working on this and there's going to be millions of people using it, so which one do you want to optimize for? The 10 or the millions? So you have to keep on reminding yourself that, at least I do. 

0:31:38 - Mehmet
Yeah, I think user experience is something also underrated, I would say. I see a lot of apps sometimes and it's not because I want to criticize the app developers or whoever, but I just, you know, instead of putting, for example, one store or two stores on the app store you know so what I do, usually, you know, I tell them guys, I'm not criticizing, but I think, for example, your onboarding was like five steps and I think you can do it in two steps. 

0:32:06 - Zach
Right. 

0:32:07 - Mehmet
Or it took me, let's say, 10, I'm not exaggerating Like it took me 30 seconds to move, you know, from the first step to the second step, and I think you can optimize something in the back. You know, I don't know, it seems you are using maybe a slow or outdated infrastructure. So, yeah, like this is something. Also because, coming back and allow me, zach, to add this, because I talk with a lot of founders One of the things is when you put also a lot of features, and these features, maybe the user, they don't want them. And I ask you specifically about the AR VR things because I want to convey a message Again because, back to the AI conversation, don't overcomplicate your system, because the user will use it, not you only. 

And you know, for me, as a normal user, I like the apps that are like very straightforward Maybe two clicks, three clicks maximum, and I'm done, I'm out of there, right? So which is? It's an underrated. I'm trying actually to get someone we had one episode before about user experience, but I'm trying also to get another person on the show specifically to design user experience, because I think it's very, very, very underrated, especially within startups. Now, is that moving to something which is. I am lucky to know that the launch is at the time of recording this episode, which launching your book. Grow up fast. So what can you share with us and I know it's like about your transition from an employee to a founder and the mindset so but what you can share about the book with us? 

0:33:44 - Zach
Sure. So this has been a fun project I've been working on for about a year now and the premise of grow up fast is I wanted to consolidate down the lessons that I've learned in the past seven years on starting an AI company. So everything from evaluating, is this technology applicable in this space? Because at the time, ai for moving like I've literally been laughed at before in prospective investor meetings we were explaining what you're trying to work on, it's like we're gonna take everything that was involved in putting man on the moon and it works in self-driving cars and drones and we're gonna sell it in this insurance industry and moving companies and people just like don't always wrap their head around it, but like we knew something, like we knew that there was something valuable there. So we talk about that, about how did I convince myself and what do we do along the way. So we chatted a bit about that in the early days around. Can I even get somebody to click a link and go through and do this? And then we became a remote company during COVID. So managing a team has changed and what we need to do like for two people, for five people, for 50 people changes. So they've got some lessons along the way there and I tried to kind of digest things down and make it move quickly. 

So writing a book is a lot like writing software, in the sense that you know you're actually done when the end user is able to achieve their goals and enjoy what they were trying to do from your app or from your book. So I had never done this before, so I treated it sort of like a product launch. So I thought I was done. I had an editor, so we go back and forth, so I had somebody I could debate with. And when I thought I was done, we did a user study. So I found some beta readers that were in the target market and then I gave them free copy of the book. I gave them a feedback form. And the last question like you were talking about App Store reviews, so it's if this book were on Amazon today, how would you rate what you just read? So you can't factor in like potential changes down the road, it's just based on what you read. 

And I was so happy to be finally through all the end of it. I was expecting to get like awesome marks and 2.8 out of five came back and I was like all right. I got some work cut out for me so that was about six months ago and then kind of dug in and learned people wanted to hear more of the human aspects of it, like it's not just like a, not just focusing on technicalities. So we had to go back and you had to have humility, kind of come in and say, okay, this thing that I thought was awesome, I built it for myself, maybe. So we did a couple more rounds of that and then when it got up to 4.8, I made my piece with the fact that you can't please everybody. 

I'll never probably hit five and that's the copy that's live today on Amazon. 

0:36:37 - Mehmet
Wow, so it's available on Amazon. On the time of recording this episode. 

0:36:42 - Zach
At the time of recording. Yep, if you're watching this, it's live. Yeah, it's called Grow Up Fast. Lessons for Men AI Startup. 

0:36:47 - Mehmet
That's amazing and actually just to add my two cents here. I think, zach, I'm not surprised that people wanted to hear from you the human aspect, because I think what's happening, especially in the tech world maybe I repeated this three times, four times today but we are trying to make the technical aspects so shiny and people they get bored after a while. I mean, I'm seeing this and the other day I noticed something, for example, myself being in tech for a long time and used to be working with startups and vendor side, so what I see people and I took LinkedIn. Linkedin is because for me it's a place where I can test and I have audience since long time but I noticed that when someone shares too much technicalities, when they share, it's kind of a rigid thing cool technology people. 

They don't react, and I think people are waiting for more human blended with technology, maybe something like this. Because I think yeah, because we know for a fact that, yeah, technology is evolving. We are now ready to have the new breakthrough product anytime, and things can change very fast. So people are not anymore like wow when they see some oh, wow, look at this, it's very cool. So they need to understand where the idea came from, like in your case, for example. They want to understand okay, what was your inspiration when you started your company? How did you manage, for example, the stress with your colleagues? 

People are interested to learn more about these things, more, and this is why, by the way, I had to pivot my podcast because of this, because for the first 30 episodes or so, I do daily, by the way. So the first 30 episodes or so, they were short. You can listen in 10 minutes, 15 minutes. I was just giving either explanation about a certain tech or maybe a breakthrough event that happened. Yeah, it was going fine. I didn't expect that time if someone would listen to it. 

But the moment I start to change the theme a little bit and get a guest like yourself and not only discuss technology, discuss some other aspects, things change dramatically for me and I listened to my audience and said, okay, got you, I understand I should not be only like making it hugely tech. And I said, should I change the name CTO Show? Let me think about it. I said no, I don't have to change because CTO is something, it's technology, cto is startup, cto is entrepreneurship and we need to learn a lot of things. And we discussed about communication, marketing, sales, all this stuff. So I said, yeah, let me keep it this way. 

So it was a good reminder also, zach, for myself and for the audience, maybe about the human aspect. Well, as we come to the end, I have a final question. Okay, what is the question that you wished I had asked you and how you've answered it? 

0:40:10 - Zach
This is great. You're questioning your assumptions and you're open to feedback. I like it. You're following all the right principles. 

0:40:18 - Mehmet
Yeah. 

0:40:21 - Zach
I think we could have talked maybe a little bit more about some of the challenges that we ran into with and if I can leave you with maybe a quick parting story, I can talk about a fun challenge. So the Yenbo product. We allow sales agents to close more jobs because they're not asking how many boxes does your living room fit in? They're able to kind of get that visual documentation and build a report with the customer, focus on all the things that they're good at and then kind of move on to the next one. But sometimes people don't always know how to wrap their head around it. 

Like I always like to say, we're not selling outlook to Gmail users. You're making something new, often in an industry where people have been doing things a certain way for decades. And even if you were to look at a screenshot of our tool now, our design team has put together a bunch of research around this and the way we ended up. Now we have a lot of very simple single stroke with icons, printed corners, kind of small, fun, playful animations, and that all came from like we learned that customers would buy the product but the people who buy aren't actually the users. It's like the owner might buy and then, like their team, is expected to use it. So we realized we have this gulf where the value prop that sold us to the company you still have to like win over the individual users and make sure that they can work it into their workflows. 

And we found that just by beating our chest around AI, we were intimidating and alienating a lot of people because, again, they think of Terminator and they don't think of being more productive and being able to close more commissions and things like that. So we had to kind of go out of our way to make this technology accessible. So, like a traditional object detector, if you read academic papers, you normally draw the bounding box around it. It looks very like a very terminatory, but in our world, like we put a circle on it on the center, it's like not really a technical reason why it just it looks less scary, there's no sharp corners and that actually like moved the needle in terms of if you looked at you did a training session how many people actually logged in before, as people need to be able to see themselves using it, saying, okay, that that's like a future that I can, that I can understand, and by building and kind of a maybe obsessing over these details around, like how do we make it easier to use and more inviting, is something that we we learned by doing it the wrong way and then adapting quick enough that we were able to kind of write this course in time. 

But I've just been kind of impressed at how there's very few just like pure technical problems that we faced. There's always like a mix of its parts technical parts, software, design parts, humanities and I feel like that's where the real, like rubber meets the road is how do you balance all these sometimes competing interests and then make something new out of it. So it's been. It's been a fun journey. I'm excited for the next few years, but I feel like nothing comes for free. You have to think through all these different things and then think through ways that that you can build something that's valuable to your end user. 

0:43:35 - Mehmet
Yeah, great. Actually I was reading something today very similar about you know when companies decide to do a change and then you know how the insiders, they would be afraid of losing their jobs or maybe they will be politics sometimes also as well. And you know the book describe actually the book is the blue ocean strategy. So so you know they advise like you need to do seeing is believing. So you need to show them actually the benefits so they understand they can visualize it better. So, 100%, exactly On that point. One thing I did ask you, zach, and now I remember have you expanded the? You know, you know, did you have presence other than the US? 

0:44:17 - Zach
Yeah, we do. We. We didn't do it on purpose, but moving is very networked. So if you're moving like New York to London, most moving companies aren't in both. So you, they partner with the firm in the other country. 

And we had a customer in the US and same thing happened. They did an international move and then on the receiving end that company got our visual inventory. So that's most moving companies traditionally have like just a written list of all the different items that are in the move. But the Yembo difference is we put the actual pictures of the items there. So it's not just like any TV, it's this TV and here the here's the size, like that. So they got this and they're kind of shocked like who has the time to go through and put like 300 little pictures of all these different items in here? Like it's great for me, but how did it? How did somebody put this together? And so they reached out to us and one thing kind of led to another. So we're in about, I think, 20 countries or so now. So most of our business is in the US, but it we do. We do have, we have a data center in Europe as well as in Singapore, and we serve clients all over the world. 

0:45:21 - Mehmet
That's great to hear. Well, zach, thank you very much for being with me today. I really enjoyed the discussion. I think you, you know, put the light on a lot of topics you know from technology perspective and from management style perspective, and also overcoming the challenges. So thank you very much for coming and sharing your experience and reminder that the book I hope by the time that we air this, I would be able to put a link to the book. Anyway, I will. I will leave your LinkedIn profile and your website, the MBO website also as well in the episode description. So if anyone is interested, whatever you are in the world, to explore the technology and maybe connect with Zach, so feel free to do this. 

And, as usual, when we come to the end of the episode, this is what I repeat If you have any feedback about this episode or the show in general, please don't hesitate. You can show me on purpose. I know that it's very noisy up there, but whatever platform you use, you know this, why I put all the handles. So reach out to me guys, I would love to hear your feedback about the show. If you are interested to be also on the show, same as Zach was yesterday, don't be shy. Please come in. I'm having guests from all around the world. Wherever you are US, new Zealand you know. Opposite direction, japan, north Pole, south Pole so you can be anywhere in the world. I can arrange the time zones, no problem, and thank you very much for tuning in and we'll meet again next episode. Thank you, bye, bye. 

0:46:52 - Zach
Thank you so much for having me, mehmet, my pleasure, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you, thank you. 

Transcribed by https://hello.podium.page/?via=mehmet