The Woman Behind Gemini: How AI Is Actually Built and Who It Will Replace | Sneha Shah

0:00For adults with Crohn's disease or ulcerative colitis symptoms, every choice matters. Tremphia offers self-injection or intravenous infusion from the start. Tremphia is administered as injections under the skin or infusions through a vein every four weeks, followed by injections under the skin every four or eight weeks. If your doctor decides that you can self-inject Tremphia, proper training is required. Tremphia is a prescription medicine used to treat adults with moderately to severely active Crohn's disease and adults with moderately to severely active ulcerative colitis. Serious allergic reactions increased risk of infections or lower ability to fight them and liver problems may occur. Before treatment, get checked for infections and tuberculosis. Tell your doctor if you have an infection, flu-like symptoms or need a vaccine. Explore what's possible. Ask your doctor about Tremphia today. Call 1-800-526-7736 to learn more or visit TremphiaRadio.com. AI is everywhere.

1:02Everyone wants it. The challenge is unlocking its value. CDW helps organizations move from idea to implementation, with the expertise, technology and partnerships needed to make AI work in the real world. Don't just launch AI. Land it. Because when AI delivers results, it's amazing. We configure, optimize and deliver the tech that runs business. CDW make amazing happen. You've never been one to settle. Stand down or stand still. You're a lifelong learner. Energized by excellence. There's a fire inside you. You can't ignore. You've got competition to outrun, momentum to build on, and your own high standards to meet. Stop now. Not a chance. At Capella University, we help you catch what you're chasing, because you've always had the drive. Now, go earn the degree. Capella University, what can't you do? Visit Capella.edu to learn more. What is up, the entrepreneur?

2:03DNA, AI, AI. It's all anyone can talk about. This episode is going to blow your mind with the guests I have today. Now us little people, we're just getting into AI. This is a big topic for us, but for the big boys, the Amazon, the Googles. This has been around for a long time. I have the Queen Bee who has been running this for these companies for the last two decades. She is the lead AI specialist for Google and Amazon. Sneha Shah is here. How are you? I'm doing very well. Thank you, Justin, for having me and very honored to have that level of introduction. Yeah. I just told you off camera. I was like, you're like the little ninja in the back pocket of all these big, big companies. And no one knows this you like behind the scenes being the magician you are. Now you're on your own thing and we're going to get to Hazel and everything that you're up to today. But I want to take a step back. I think, you know, for most of the common people, the lay people, as we say, we look at Amazon

3:07and we say it does this shipping thing and we look at Google and we say it does this search engine optimization thing. We do it. We use it. I order my wife orders Amazon probably 10 times a day. Right? Now AI is in the search engine category. But for the most part, Google over the last decade, I go to Google to go search whatever I want. You are literally one of the founding members of how Amazon started Amazon delivery, correct? Absolutely. I was part of the team that made same day delivery happen. This is the magic where you order the previous night and it shows up at your doorstep seven in the morning. We scaled it to about 10 countries and super proud of the scale that we've been able to build this at. Yeah. You know, it's funny because I just live here. I wasn't even aware Amazon had that many. Are you a hunt?

4:08Not you, but are they 100% in most all countries at this point? Many, many countries at this point. Now walk us through what it took to build something at that scale, right? Like it wasn't like, oh, I'm just going to start in San Francisco and I'm just going to do Amazon San Francisco. Like did you start with a state, a couple states, you just go national and how did you build the scale with that? Correct. So a lot of people and a lot of team effort goes into this. I have to say, you know, many years went into building this. It is a combination of software, infrastructure, as well as the delivery team that powers this. But yes, we, Amazon has the culture of starting small. You know, they make it into a startup within the larger organization. We started the delivery in Seattle. Actually, it starts that small. Then you expand into bigger metros just to test the scale and then you expand into more countries. You know, some of the bigger challenges was countries like India, you know, the population

5:13is huge. So the volume is immense. You know, is there even the infrastructure to support it? So it is a gradual rule of it took us many years, but we were pioneers in that even a Walmart took, you know, almost half a decade to follow on and be able to meet that. So huge endeavor. Oh my gosh, were you one of those like little pods that started it in Seattle or built it out in Seattle and then let it expand? At what point were, when did you exit that? I mean, I can only imagine that was all consuming for how long? Yeah, I spent about five years building that. My team is the one that got to decide, you know, Will FedEx make this promise that we made our customer. Will UPS or USPS or whoever make that promise? Eventually, we realized that Amazon should become the transportation itself like running the last mile delivery.

6:13We call it like the trucks or the small cars to go and deliver it. And, you know, all of these different programs come together over many, many years to make that promise as what we are today. And therefore, you know, yes, I saw it over five years, grew it to like I said, you know, 10 countries. And then it comes to a point where you're like, you know what, I could do this for all my life. I could become a transportation expert for all my life, but I decided to pick another challenge then. Yeah, now just quickly to stay on topic, I, when you make this Amazon specifically, when they think through this idea, like they have to look at the cost of what is the truck the people, the salaries, the lost goods, where you are part of trying to figure out the model, like it just sounds so. If you were to say, hey, we want to go start this today, I would think about all the costs that it takes to lift that vertical up.

7:14Was that a losing model for a while or was it always profitable? It's Amazon delivery has always been a losing model. It is all about the experience that you provide to your customer that brings them back. Shopping and that's taking us that it created. Of course, there were micromodils that were created as to, you know, let's not ship something over air for tomorrow morning 7 a.m. You know, you know, from United States to Europe, please don't do that. But of course, there were ceilings created. But for most part, the whole value of Amazon back then was, you know, make this the world's most, you know, customer friendly side and we lived and breathed by that. So, you know, so it, do you think today is still lose? I mean, it's so, it's every, I'm not joking when I think I get 10 deliveries a day. I'm not joking about that. Is it still losing because the cost of operation is so high and then Amazon monetizes in other

8:17verticals, but it keeps such a great stickiness to their client? Correct. The imagines are pretty low on the retail business. No one thinks about that. I almost guarantee people are mine. Like, I know I'm blown away. That is that low. But it does make sense to say, okay, you got to offer something amazing, essentially for free to be able to give them the thing that you can make some money on. Exactly. Now, I know this was in your department, but where does Amazon make the most of its money? Is it the AWP? AWS, 100. AWS? Yes. Yes. We got to software margins. That's where the scale came in. We also, you know, Amazon, how does number one for many, many years took on? So certainly we made most of the money there. Yeah. Now, were you over at Google before Amazon or were you at Amazon before Google? I started my career in Amazon straight out of Carnegie Mellon. State that for six years grew this move to Google because the opportunity of, you know,

9:17building cloud came in and became real at that point, you know, compute and storage just started off. I got the opportunity to become founding engineer of a pioneer database, like a global scale database. I was like, this sounds amazing. I did physical delivery at Amazon. Now, I will get to do bits and bobs like data delivery across the boom. So super exciting opportunity and I moved to Google for that. That is great. I would assume if we define AI, were you playing with any use of more logistics over at Amazon or was there AI layered into it or did you not really go all in on AI until you got over to Google? There were very early versions of AI or machine learning built into all of these platforms. Yeah. At the scale and the speed that you want to operate at, some of these dishes, decisions are baked in with a lot of machine learning that went in early on. In Amazon. What would be, you know, when was this?

10:21What years were these? Because I think again, I'm so excited about this episode because for the commoner, we're like, oh, AI has been around for a year and a half, two years. It's just not the case, right? Absolutely. When were you kind of working with some of the fundamental AI components in over in Amazon? What years were these? This is early, you know, 2012, 2014. Right. Right. It's called machine learning back then, you know, machine learning dates even before that. It just became mainstream, you know, in the early 2000s, where we started seeing value generated very quickly out of it. Yeah. Now social media has made it the biggest thing on the planet. Absolutely. I mean, every day, you're scrolling on LinkedIn TikToks of the world as a wall based on that. AI. Now, you get over to Google and you got the privilege to lead up a pod and what was your focus there? So I was the founding engineer of a database team.

11:24It was called Cloud Spanner. It is one in the only database that allows you to spread your data across the globe and still be accessible at, you know, millisecond speed to say it very simply. Think of the Amazon's the world, the world that's served a lot of customers across the globe. Imagine every time you had to buy something, the data had to come all the way back to you the US to verify that you're a user and then go back to your country in India. Hopefully, which is literally diagonally across. And what tends to happen is it makes it slow. The experience is bad. And then you know, people just wane off eventually. You want to be able to make it fast. You want to be able to make it easy. You want to be able to have your customers use more of what they're doing. And so Spanner allows that data to be distributed where your user is. Very simply put. Yeah, but not simple to do.

12:27Not simple to do. Yeah, this this group like, you know, we were four when we started this group into maybe a 400% team. We had the likes of the biggest banks in Europe. The biggest gaming companies. You know, I mean, some of these names are public. So you know, Deutsche Bank would use. Yeah. And then Tando would use a Spanner. These are, you know, think of when Pokemon Go was released at a point, there were about 10,000 users playing the game at the same second. Wow. We'll be able to quickly move the data around have them. The the experience that they have was very critical. And that's what we followed. Something amazing is happening in data management. Even though storage needs are constantly changing, companies are only paying for what they need now. That's because CDW customized an ever pure evergreen one SLA driven storage as a service solution delivering hybrid cloud storage that's easy to manage, efficient and continuously

13:29evergreen. Simplify storage administration with ever pure and CDW make amazing happen. Find out more at CDW dot com slash ever pure. You've never been one to settle. Stand down or stand still. You're a lifelong learner. You're a huge ice by excellence. There's a fire inside you. You can't ignore. You've got competition to outrun, momentum to build on and your own high standards to meet. Stop now. Not a chance. At Capella University, we help you catch what you're chasing because you've always had the drive. Now go earn the degree. Capella University. What can't you do? Visit Capella.edu to learn more. Hey, it's Kelly Rowland. You may not know this, but I have XM. So I get how it can steal your time. But why let XM take over when you can talk to your doctor about Ebglyce? Ebglyce, lubricism app, LBKZ, a 250 milligram per 2 milliliter injection is a prescription medicine used to treat adults and children 12 years of age and older, who weigh at least 88 pounds

14:32or 40 kilograms with moderate to severe axima. Also called a topic dermatitis that is not well controlled with prescription therapies used on the skin or topicals or who cannot use topical therapies. Ebglyce can be used with or without topical corticosteroids. Let's use if you are allergic to Ebglyce. A allergic reaction can occur that can be severe. Eye problems can occur. Tell your doctor if you have newer, worsening eye problems. You should not receive a live vaccine when treated with Ebglyce. Before starting Ebglyce, tell your doctor if you have a parasitic infection. Paid partnership with Lily. Respect your time. Ask your doctor about Ebglyce and visit Ebglyce.com or call 1-800-LilyRX or 1-800-545-59-79. So in use gaming, I'm not much of a gamer, but like all the streaming and all that kind of stuff, is this streaming? Is that what basically talking about? To be able to have a million people streaming a game all at one time to be able to have all that data going around at that lightning speed? At that lightning speed. Exactly. Streaming, things like the leaderboard is a classic example.

15:34Millions of people playing this together. Nobody likes to be one second related to be at the leaderboard, right? You want to know your top players. That is powered by Spellar. That's amazing. And so you took out now how long were you over at Google? Almost a decade. I moved, I got to scale many products at Google. You know, I took Spanner to over a hundred million, realized that, you know, this can keep growing. The team had become pretty big and you know, there were enough people to move the scale beyond that. And so I got the opportunity to move to vertex AI. I built that like the flagship AI platform for Google, scaled that into Gemini Enterprise, you know, their agentic platform. And so you had the honor to be able to scale what was the name before it was called Gemini? Good. Gemini Enterprise. But before Gemini, what was it called? Vortex AI. Vortex AI. Vertex.

16:35Vertex AI. Vertex AI. Vertex AI. Now is Gemini, which half the world probably uses at this point if not more. And you were able to build that out. Yes. What is it take, you know, I'm old enough to know that like coding, right, was like my generation, like the internet launching and coding. When you're learning how to do what you do, are you coding? What is going on behind the scenes of AI, right? As effectively you built vertex AI, which now Gemini, right? So let's just say Gemini moving forward, you built it. How the hell do you build it? What are you doing? So a day in the life of, you know, a senior architect is what I do. It goes anywhere from understanding what the research landscape is doing, right? Understanding where the customer finds the value. We need to bridge that gap. That's my role. There's a lot of value that can be created with, you know, LLM's, which is the hype now.

17:41But bridging that gap for a Lehman is important. So going from understanding business use cases, going from actually prototyping some of this, you know, the coding that you mentioned. Showing the value to your leadership, right? There are a lot of bets they can make. And why is your bet important is it was one of my very critical drones to be able to influence them and show them the strategic value for our customers. And then of course, we're reading a very large team to make sure these things work in reality in production at scale. And that's provide value to your end customer, you know, one of our very early customer is a beloved social media website. And they do a lot of search. They wanted to scale their search on Gemini Enterprise for billions of users. And to be able to have that get unlocked within matter of months is the magic that I create.

18:42That's wild. I just, it's just so foreign to me, right? Like I'll go to Claude or I'll go to Gemini. I'll use it. How the hell was it built? That's insane. So to build it, what is what is tactically like what is actually happening to build it? Is it code? Is that is an easier way of explaining? It's just a bunch of code. Yes, it's a bunch of code and a bunch of very smart minds that come together to define what that code would look like. Define the boundaries, define the constraint, define the usability. So there's actually people behind AI. So everyone's talking about machine learning and we'll get there because we'll get there. But to start to build a Gemini, it is very smart people essentially coding, building in constraints, building it, and I don't even know the right terminology. Like building it. Like when someone asks a question, this is what we wanted to be able to spit out.

19:43And this is correct. A lot of data went in. So there are teams that just focused on data. What type of answers can we get? There is a lot of news on books. Models have eaten up all the public books alive. Now they can collect that data and feed it into the system so that we can answer very, very, very, you know, we can get into the details of some of these books. There were teams that actually took some of these questions and answers and iteratively test how good are we to answer these questions. It took a very long time for LLMs to say 2 plus 2 is equal to 4. Really? Yes. From time on, very long time to get there because they are probabilistic models to get them to a deterministic answer was hard. And therefore, a lot of teams and efforts go into making some of these technology hard into four real use cases. Wow. And today, you know, the whole who applies around outcomes.

20:46Like, how do we make sure that this actually creates business value? Not just 2 plus 2, 4. Everybody knows that. Not just solving math problems that kids are solving. The business needs to run million and trillion dollar scale. How do we enable that? Yeah. Well, I mean, that's the other thing that, you know, I think, ChatGitBT knows ads, right? So because you go, there's got to be a cost to all this. And so who's covering the cost to all this? Because it's got to be a revenue model. Our ads going to be the revenue model is going to turn into Facebook. Essentially, your Google and just ads become a huge revenue piece of how AI exists. Or is there more than that? There certainly be an ordinary consumer side that is much harder to put a value to this. And therefore, ads or targeted tools while you are doing your task will be suggested. But in the enterprise world where I primarily focus on, we definitely believe that models will get highly commoditized.

21:54You've seen the recent news by OpenAI just because open source models, one times the cost they're doing has very open AI has considered lower their price as well. Right. And now, how can enterprise use this to generate real world outcomes is the question. Yeah. Yeah, that's, well, let's move into what you're doing now, right? I could ask you a million questions about AI and I think everyone could, right? I mean, you're, you're one of, I don't know, maybe a founding queen of this whole thing. Who knows where, where you sit into this bigger race. But I want to talk about Hazel, but I also want to talk the machine learning part about this. You mentioned it kind of briefly. As it's being used, are we at a place where a lot of this is going to be able to like, you already said they've consumed all the public books, right? So they already have the wisdom and knowledge that comes from, you know, 2000 years of book writing or whatever books started being written.

22:58But wouldn't that have all been housed in Google with, so this is maybe my naive, but where did AI even go find it? Wouldn't have come from Google in the first place? There are two parts to that answer. I think the first part that you're asking is this information that was publicly available on Google or the internet is what has been consumed already by the models to be able to give you an answer. No, right? Google search was about I don't understand your intent exactly, but I can give you a choices from which you can find your answer. Models are a little more at one step forward where they are like, okay, you know, we can understand your intent behind this. And I present you with an answer instead of just giving you suggestions or links, the very very good find your answers. The other part is that, you know, model providers spent a lot of time actually physically looking into things that were not on the internet.

23:58So they took physical books and fed it into the system. You know, there's a lot of talk around a very niche models, you know, from India who took scriptures that were really physical. They've never been put on the internet. Languages that are not very prolific on the internet. Yeah, regional content that is not prolific was all fed into it to make that information available. But the second part that you said is is very true. It has the knowledge, but I I soak off this agree with the wisdom. There's no judgment in the. Yeah, sure. It is very black and white. It's very, I mean, yes, it has, it is if search was black and white, this is just multiple shades of gray, but a human judgment or a wisdom. It completely lacks like it does not understand the environment. It does not know why you're asking what you're asking, how you're asking.

25:00So, but it's good enough to be able to at least give you the direction in which you want to go. Yeah, and I use it for a lot of, you know, thinking through my thoughts and reiterating my thoughts and using it for content is, you know, obviously my podcast then well. So for that practice, it's great. I mean, I like Claude probably the best in that practice of just, you know, IDAing and coming up with ideas and refining the idea and understanding how to explain the idea, but there's so much more to it, right? Whether it's be Grocker. Again, Chad is the first, you know, I feel like that's the starter level and then there's all these other ones that were aware of Gemini, obviously is there. Hazel AI, you are now, so you left the Titans, you've left the big boys and you say, I have so much experience. I'm going to go do this better. Talk to me about Hazel AI. Hazel AI becomes the engine behind the world's largest economy of making technology.

26:02So today's system integrators, you know, the Accenture Infosys TCS is of the world are the largest technology economy. So, you know, the biggest thing that I've ever seen is the technology that is, they serve all the global 2000s to do become tech native cloud native and now they will make them AI native and Hazel becomes the power engine or the technology engine right behind them. So, what I did this is all the learnings and the wisdom that I had over the years building some of these products from 0 to you know millions of dollars, I encoded them into the Hazel platform. So, now I was sweet of agents with that same level of wisdom, the judgment that it takes so that the technology providers can go faster, have higher throughput and ROI. So, you know, the average, you know, models or AI to generate real value. Now, so what whose Hazel AI is ideal client like who would use Hazel AI.

27:07Let's work backwards of how this economy works, right. I'm going to take an example of a very large enterprise today like a Coca Cola or an Amtrak. These run on technology or ERPs, let's take one example as SAP Coca Cola must run a very large ERP system across the finance procurement inventory pricing. What's ERP. ERP becomes the software that runs all of these modules of functions like say Coca Cola factory needs to produce you know a million bottles to be shipped to Chicago. All of this work is getting tracked in an ERP. And that work that becomes a software layer for all of their teams to look into so that procurement team is saying, I need you know Chicago's procurement team will say I need a million bottles. The factory on the other side is saying I need to produce these million bottles. I'm going to create this order over here.

28:11I need to the shipment team is saying I need to box this and send this to Chicago. So, they are tracking this work in the ERP. So, SAP is one of the largest ERPs you know many billion dollar time. And what tends to happen is these are extremely complex systems that need to be wired together. Now, a TCS or an Infosys or an Accenture comes in and helps Coca Cola set this up to do the upgrades to maintain this to make sure that you know this becomes a I native now. And what even an Accenture or a TCS would put about 100 to 200 people to maintain this on a regular basis just for Coca Cola. In all those different there would be 100 to 200 people doing all those different things. And what I'm hearing you say is Hazel can come in consolidate all that work in one place. Maybe there's 10 per I don't know how many people work you know work on Hazel but you can solidate all those departments, all those different things in one place.

29:15And it's doing it all for for Coca Cola. It will do it all for the TCS and Infosys today. So that the the expertise of a TCS person is no longer around. You know, how do I move the needle from X to Y like the coding part is taken care by Hazel agents. Yeah. What they are bringing to the table is expertise. Right. So all the repetitive task of. Can I can I write the this as a design you know just like how you idea it. Can I you know write the code for it can I go and you know test it can I go and you know deploy it can I scale this all of the engineering goodness is done by Hazel agents. While the the the folks from TCS and info the architects on that team now become the layer that just validates and pushes it. This not only brings them the speed but it also brings them the efficiency like now they can take not only a Coca Cola but the same architect can now work on.

30:17Epsico on free to liaison everything right he's not just constrained by the time it takes to do this his work. When you partner with CDW you get more from your devices with solutions that take modern work to the next level. CDW experts are delivering powerful productivity with Lenovo AIPCs helping users block out distractions access virtual support anytime anywhere and share content seamlessly between devices. Make amazing happen learn more at CDW dot com slash Lenovo. There's a fire inside you you can't ignore stand still not a chance you're a lifelong learner who's come this far now we're here to help you keep going further. Capella University what can't you do visit Capella dot E.D.U. to learn more. Hey it's Kelly Roland you may not know this but I have Xima so I get how it can still your time but why let Xima take over when you can talk to your doctor about Eppglus.

31:21Eppglus lubricism ab LBKZ a 250 milligram per 2 milliliter injection is a prescription medicine used to treat adults and children 12 years of age and older who weigh at least 88 pounds or 40 kilograms with moderate to severe axima. Also called a topic dermatitis that is not well controlled with prescription therapies used on the skin or topicals or who cannot use topical therapies. Eppglus can be used with or without topical corticosteroids. Don't use if you are allergic to Eppglus allergic reactions can occur that can be severe. I problems can occur tell your doctor if you have newer worsening eye problems you should not receive a live vaccine when treated with Eppglus before starting Eppglus tell your doctor if you have a parasitic infection. Pay partnership with Lily respect your time ask your doctor about Eppglus and visit Eppglus dot com or call 1 800 Lily RX or 1 800 545 5979. So is it easily said you know you hear a lot of these experts who kind of try to de-bunk this I don't want to say myth but this idea that like AI is just going to take everyone's job.

32:25I do believe there's a component of that right I mean you are Hazel AI is reducing 200 people into how many 10 15 people doing the same thing. I don't know I don't want to make up numbers for you but you're taking 200 jobs and you're saying really you'll need 10 people to do what Hazel can do right so that is a reality but you still need that those 10 people to come in with what you consider wisdom or to come in with expertise in a way like what what I want I'm hearing it and how I want to say it is is. If the weather is changing the AI won't be able to understand the weather is changing which may change certain data points that has to have the human behind it say based around the weather changing is going to change how many bottles we order because we're not going to ship to whatever right you still need that human to run that part. So there are two parts to the the myth that you call right AI is not going to this AI is definitely displacing jobs that means that it's not going to just take away the jobs from the economy they're going to become new jobs in the economy a developer no longer needs to just do his task sitting in a in a small cubicle in a chair but now they can get to influencing and decision making right instead of them just.

33:53typing on the keyboard now they can go into meetings and change the course of the things that they are building. So is it what is it still fair and you know and I'm not trying to be combative would it be fair to say it will remove the lower level jobs that you know you do really need 200 people right. I feel like that's the fair statement to say the lower level jobs that you don't need the cost of the human anymore I think that will go away. So entry level jobs who were just sitting and doing tiny repetitive tasks will no longer be required right but that said those same people will now start at a ladder above which will be that if they're smart enough to if they're smart enough to. Yes I mean that that that happened when you know cloud came out there were a lot of database administrative people yeah just you know like kept the lights on on a database when I came in in spanner almost all the jobs went away because

34:57they were just spanner provided that out of the box but it they become they became sequel exports yes it takes some time to ramp up and you know get into that level but not nonetheless it wasn't a fancy job to anyways keep the lights on on a database it's not like they were the happiest people doing that. Yeah it's it has been the same way like farmers one of the happiest people being in the sun you know removing. If you've gone out of the the the maze right but you know the tractors made it easier now can they expand into other fields yes 100% can they walk to the market and sell at a higher price yes 100% yeah that's what everybody needs to do. That's fair that's fair well listen I'm with you on that girl like everyone needs a level up like they got to find their lane if I was coming to take your job find where you fit like. You know I'm definitely not a victim and at all is it's hey it's here it's not going anywhere right like so where do you fit in figure that out you know I'm contemplating you know I myself I'm always trying to better myself I believe in the law of the lid meaning my businesses my income.

36:10It can only go as high as I can go exactly right and so if you're not improving yourself there's a lid you can't earn more you're not deserving of it and when you try you will likely fail and I've done that by the way where I was way outside my lid and it didn't go so well for me so now he's. I want the things that I wanted to bring back is that we also need to know that the opportunities are not finite with AI the size of opportunity has increased so now even where like just in my my world of just development you see a lot of coding has picked up like the number of apps have exploded you know 10x 100x yeah enterprises are taking far. This is taking far more max just because AI has made it cheap and dirty sure right and so the there are a ton more opportunities than the previously and so the level to go off or sideways has also increased and most people need to get into that that discomfort zone as I call it yeah we're able to find their own lane.

37:23I think that if no one understood any of what we're talking about that is probably your biggest takeaway right is that the opportunities are there now where do you see you know where we're going now hazel specifically like who do you work with are you able to say like who you work with and where you're integrated or that you can't mention those names very confidential we'll have a lot of news coming out and PR around it let's call. Yes, congratulations. I try to make a ton of announcements this summer good and some really big names we work with five of the top 10 GSIs you know we serve over 10 billion dollars in business or some of these companies we the end customers are anywhere from retail giants down to manufacturing to agriculture so super proud where the impact I can make. You should be very proud of yourself very proud of yourself and in them very honored to be able to have your phone number and be able to call you because I don't know how to work any of it so I want to bring something to light here you spent 20 years at two of the largest companies in the whole planet Amazon and Google.

38:38And you're talking about now building a company that is serving what is a five billion dollars of enterprise value or 10 billion dollars in it is your company it's yours you are the CEO snail is the CEO and I want to highlight something here. And I don't know what that means in revenue and all that kind of stuff because when you talk about five billion in enterprise value and 10 billion like I'm not I don't know how that trickles down to hazelay I but while tell you is what I heard here is you put in the repetition to deserve to be in the spot you're at today. And a lot of people aren't they aren't they want to be snail hog because they say oh I don't need to go to college I'm going to go build my own AI well how I'm going to ask you this is the questions where the questions coming how did you get into the doors of these very large companies.

39:38Did you maybe reference where you worked for 20 years it's definitely my credibility. It's me building Gemini enterprise may being able to scale a lot of these businesses from from you know zero. I'm going to ask you to pause respectfully I want everyone to hear what you just said. And I don't know how old you are 45 and I look at younger and when I say younger not a lot younger like in the 30s. And they just think like oh I'm going to start this AI company and I'm going to be the new hottest AI company and they're they don't even have a quarter of the resume you do nothing but they believe they are going to start the biggest new AI thing. And I just I want to applaud you and I hope that these listeners and viewers understand what her and I are with the question I just asked was sometimes to build greatness you need to go learn from people who did it first or do it on someone else's budget right initially your budget was zero because Google paid for it or Amazon paid for it and you got to learn off of their budget and you got paid to learn it.

40:48And I just wanted to pause you there respectfully because that's so valuable. Oh absolutely I mean there are two types of people one who don't mind. Slogging it out to learn at their jobs and a lot of part that I do at Hazel today which is sales marketing even this podcast is way beyond my comfort zones are we learning. should be. I mean, if you're a teacher, if you don't learn that, you know, it's, it's not the same time. But I definitely believe that learning from the very smart people at Google and Amazon has served me really well. The opportunities that I got are, nearly, you know, to the top 0.1 percent of at all. And I got them. And that goes a long way for me to go to your. Yeah, I appreciate you highlighting that. And in Hazel AI, obviously, now, just

41:51for purposes of understanding, the normal everyday, today consumer, myself or others, we wouldn't be utilizing Hazel AI. Maybe it's embedded in something else we're using. But there's no like consumer value for Hazel AI, correct? Okay. Yeah. So I just wanted to make sure everyone's not like, Oh, I got to go use Hazel. It's not, you know, clod. My understanding is very enterprise level. It's embedded within these very large companies. And that's where you'll, you know, be able to use it. Any last any, you know, listen, I could talk about AI. It's here to stay. Obviously, do you have any, you know, if you could be no stradontists and kind of predict where we're going to be going? Right? I think everyone has their opinion. But when someone like yourself, that's literally been in the trenches of it for two decades, I tend to believe you more than the influencer says, whatever is going to happen, right? Where is this going? I mean, medically, I, you know, I was listening to a friend of mine, Ed, my let and he says in the next two years,

42:53the medical advancements are going to be unreal, right? I know that that's not your wheelhouse per say. But give us a thought of where this whole cohesive AI movement is going to go. Yeah, I've seen a lot of these generational technology advancements. And I do believe that we definitely make one giant leap for all of us. We definitely reinvent ourselves all over again, be it in the medical field, be it in in transportation, be it in how consumers now interact with some of these applications that are pros and cons, even with Kihau, even with, you know, mobile, even with apps, all of these transformations happen. Some are good, you know, some were forgotten. And very similar that we happen with AI. But that said, what I do believe in is if we don't reinvent ourselves now, we will be forgotten ourselves.

43:57And that's the biggest takeaway. Snaha, you are amazing. I love your thought on this. I mean, you're very AI forward, but you're human. And I think that goes a lot to why I think your AI has been so successful. Hazel AI, I can't wait to hear the news. Thank you so much for joining us here on Entrepreneur DNA. Absolutely. I really appreciate you having me. And I hope that, you know, your listeners enjoyed this. Yes, I'm sure they did. All right. Well, if this was pretty cool, or you think some people who are interested in AI and technology need to hear this, please share this with at least to your friends and drop us a five star review. Snaha, thank you so much for joining us. Thank you, Justin. Very nice to meet you.