Inspired Execution

00;00;04;22 - 00;00;09;11
Chet
Dmitri, welcome to the inspired Execution mini series.

00;00;09;14 - 00;00;12;21
Dmitry
Thanks so much for having me. I'm excited to chat.

00;00;12;23 - 00;00;24;10
Chet
You worked at MySpace, Google and have started three companies. Looking back on the journey, what advice would you give to a younger version of yourself?

00;00;24;12 - 00;00;28;25
Dmitry
Don't do it.

00;00;28;27 - 00;00;34;00
Dmitry
Go, go, go. Travel the world. Have fun.

00;00;34;03 - 00;00;37;21
Chet
You're looking for better ways to make money.

00;00;37;23 - 00;00;41;16
Dmitry
Certainly there are. Certainly there are.

00;00;41;18 - 00;01;25;09
Dmitry
But beyond that, look. Iron. I was born in Russia. I moved to the U.S. when I was nine years old in 1979. Didn't know the language. I wasn't a very good student in high school and college. I got a degree in electrical engineering but didn't really like it. And so I've never done a day of electrical engineering. I discovered sort of software after watching the movie War Games and the Theaters and late 1983, 84 and and just started nerding out in front of a computer, you know, in front of a monochrome screen.

00;01;25;11 - 00;01;47;18
Dmitry
And I've always found, you know, code to be fascinating and the ability to be able to create something out of nothing without having any, like raw material. Right. Like you want to build a little box, you need to get some wood and glue and things like that. But you're on build a house. You got to spend a lot of money and get land and get raw materials.

00;01;47;20 - 00;02;08;09
Dmitry
With software, you don't need that. You just need to invest the time to do it. And I've always found that to be just extraordinarily sort of empowering and and have been doing software since 1984 and have had this like long career of doing it. But right after I got out of college, I wanted nothing to do a tech.

00;02;08;09 - 00;02;31;12
Dmitry
And so I got a job as a sales guy selling phone systems. And again, looking back, that was one of the smartest things I ever did, was because it taught me to be able to explain things to people. And I spent basically four years doing nothing with technology, selling it, but not really building it. And then sort of my love of it brought me back and doing it.

00;02;31;14 - 00;02;51;09
Dmitry
So I guess the advice I would give my younger self is just go explore, take chances. Don't worry too much about where you're going to end up. You know, follow your nose like a dog follows its nose when it's looking around for a morsel on the floor. You might look at that dog and say, Wow, a dog's an idiot.

00;02;51;11 - 00;03;12;05
Dmitry
The food's over there, just like lift up its head and see that it's there. But instead it's sort of sniffing around. I tend to find that to be the the right approach to living. And I have five kids now. The oldest is about to turn ten. So five kids, kids under ten. So I think a lot about like what advice I would give them.

00;03;12;08 - 00;03;21;29
Dmitry
Yeah, And I think I find myself in this thing of like, don't give them too much advice. Let them figure things out and sure, explore the world on their own.

00;03;22;02 - 00;03;44;13
Chet
Yeah, I agree. I one of the things I lined up telling folks where the professionally or personally is is that especially what people that I'm close to is my goal is to make not to make sure you make less mistakes. You probably will make the same number of mistakes I made. I just hope that you make different ones.

00;03;44;16 - 00;04;00;04
Chet
I just hope you make different ones, that's all. Because there should be some learnings, but you know, you're still going to do the same mistakes that I made of some percentage of them. I hope the large percentage of them are new mistakes that people have not seen before. So that's that's my general take.

00;04;00;07 - 00;04;37;09
Dmitry
I think that totally makes sense, although I'll point out that. All right, learning experientially is radically different than sort of learning by. Yeah, for reading and observing other people academically for sure. And you might know something and but unless you've experienced that, you don't actually know it. You don't feel it, you don't operate in that way. And so and it might be that we all need to make all of those mistakes, even if we know that there are mistakes in order to be able to really solidified within ourselves.

00;04;37;11 - 00;05;00;18
Chet
And what's really interesting me is that, you know, I find the the for the the Valley folklore is that you look for people who failed because because you learn from failures. And that's absolutely true. But I've also, you know, going through COVID and through different experiences, you also want to look for people who have patterns of success. Right.

00;05;00;18 - 00;05;20;18
Chet
And they actually have some sense of what great looks like, right? Not don't that they need to copy and paste it, but they must have some sense of what that looks like. Have you found that to be the case? Because it's very easy to say, hey, I learn from my failures, but you know, you also learn from your successes.

00;05;20;21 - 00;05;41;21
Dmitry
And indeed, I think it's critical to to, you know, follow the middle way, as the Buddhists would say, like you need both. You need balance. Yeah. If all you're doing is failing and that, you know, you might just be a, you know, not very good at something you, you're learning but you're still not good at it. You know.

00;05;41;23 - 00;06;01;04
Chet
I feel that, you know, this particular subject we can spend another hour talking on. Right. Because I think we both have a lot of experience and a lot of thoughts. But but moving on, you've been in the mobile application business for a long time. What has changed over the years and specifically with the rise of Jenny, I.

00;06;01;07 - 00;06;48;12
Dmitry
Well, look at the way apps are built, continues to change, has been changing, you know, since I started building applications in 1984. Right. And so a lot of things change the languages we used to build backups, the architectures we tend to use, the systems available to us, the cloud, you know, so new operating systems, like many things that transformed the way we build apps and generative AI is a part of that spectrum of things changing of how we think about building applications over the last couple of decades.

00;06;48;12 - 00;07;26;05
Dmitry
Certainly there have been a bunch of attempts at creating generally what are called low-code no code platforms to sort of abstract the code and allow more business focused users to build things rather than having to have sort of hard core developers doing the work where the developers do the work to create these platforms so that and less technical people who might be the right people to build the applications because they understand what the problems are that they're trying to solve to show up and build things there.

00;07;26;08 - 00;08;13;18
Dmitry
On the problems with those types of tools have been that they still require a person to do a tremendous amount of work. You start to understand generally how to build these things in these new type of of ways, and then you still have to do all of the work required for any other developer to do it. Like, for example, you have to envision all of the use cases and all of the edge cases and, you know, any conditions that you might need to to account for as they are, and then to spend time to create the, the no code modules for them or write some code where generative AI is radically transforming application development is

00;08;13;18 - 00;08;44;03
Dmitry
in a couple of areas. One area is for developers, for people who are writing code generative AI basically copilots. Yeah, that grok again, what you're attempting to do and you know an easy way for people to understand this like is think about like an incredible auto complete that doesn't just complete your sentence but completes. If you're writing a book parts of chapters of your book, it does that for code.

00;08;44;03 - 00;09;11;25
Dmitry
And so developers who are somewhat junior or average developers can now be radically more productive and and more powerful. And then developers who are ready, great can just get radically more work done. So that's sort of one place we are is changing that. Like for a decade now, people have been using sort of cut and paste from StackOverflow as a method for building things.

00;09;12;02 - 00;09;48;02
Dmitry
Well, you know, StackOverflow traffic is great, but it doesn't make any sense. This is radically better and it is amazingly powerful. You know, I don't write any production code anymore, but but I'm still, you know, competent and and do play with code. And so I've, I've played with GitHub copilot and other copilots and they're amazing. Yeah they are on the other side on this low code side I generally I has made it radically easier for business minded people to build applications ones.

00;09;48;06 - 00;10;14;02
Dmitry
Yeah. That do incredible things. Yeah. Without having to spend a tremendous amount of time learning no code platforms and really the skill set that's required to sort of wield the power of generative AI to build applications is more of a sort of product management skill, correct ability to articulate and human language what you want the system to do, how you want it to behave.

00;10;14;05 - 00;10;41;13
Dmitry
And if there are again, certain use, case use cases, edge cases, etc. that need to be explicitly addressed and aligned on to be able to articulate those in clear, disambiguate odd human language. And that's amazing. And so you can build things in a tiny fraction of the time and things that a year ago would have been impossible to build with a code platform because it would have taken you forever to do.

00;10;41;16 - 00;10;49;10
Dmitry
Sure. On those dimensions, I think software development is just radically transformed already in the last year.

00;10;49;12 - 00;11;08;26
Chet
So to be three years on the on the second part, co-pilot's. Yes, right. And we're going to have copilots everywhere you know, these are and they will have copilots for many personal things and things for personal productivity. And it starts with development, which is awesome. Right. But it's going to happen all across the board, which is, you know, I want to use Workday.

00;11;08;26 - 00;11;39;02
Chet
There's a copilot to help me with Workday. It learns my patterns, things like that. On the second part, on the no code part and then we'll talk a little bit about your product is I've always had this. You're a programmer, you're a business analyst. I PM and you're a business user. It seems to me that what is happening with Lambs and Jenny AI gives us definitely makes it easy for PMS, for product managers and for business analysts to write apps.

00;11;39;09 - 00;11;59;02
Chet
But we've not gotten to a point where it is simple enough for a business user who does not have any strike, has does not have a structured background to be able to go and build apps. Would you agree? Would you do what? Would you do a cut off point there? Or would you say that the technologies evolving fast enough, even though it's not there, it will get to a business user level pretty soon.

00;11;59;05 - 00;12;29;08
Dmitry
It's already there. It's already at the. So we have over 13,000 APIs that have been deployed now using our Tool MindStudio, which lets people build AI applications, AI powered applications, as we call them, utilizing models from open AI, anthropic, Google, Madame Mistral, open source, etc. The vast majority of the people who have built these AIS or business users are not developers are not product managers.

00;12;29;10 - 00;13;01;18
Dmitry
Wow. Having the skills of again, these skills to be able to understand systems and architecturally how to abstract things, to be able to get them to work better together. Yeah, these are skills that typically are found in people who do things like product management or Yes. Or like other things like that. But but one does not have to be a professional product manager to do it.

00;13;01;20 - 00;13;36;20
Dmitry
Yeah. I'll give you an example. There are two consultants, Jessica and Kimberly, that neither have written any code ever in their life. Neither have worked as product managers. They are consultants to academia. They're both PhDs, and they consult with universities, you know, faculty, supervisors, staff. Yeah, they discovered MindStudio and in less than a month built over 55 zero A.I. applications for university faculty, staff, Ph.D. students.

00;13;36;22 - 00;13;52;16
Dmitry
They understand the subject matter and the needs of their of their intended audience. That was the thing that was needed. And the rest they were simply able to use generative AI to do the work of structuring it for them.

00;13;52;18 - 00;14;18;20
Chet
Would you see those? I'm going to get a little deeper on this particular thing because I think our our audience would like this. Would you see that that is and and please I might be getting this wrong is more a lightweight app than an enterprise app. And the distinction between the two is the enterprise app to me is, you know, I'm buying a phone on Verizon, right.

00;14;18;20 - 00;14;39;08
Chet
And it cannot go down. It has billing information has got PII going on. It uses an Al-Alam from a from a chat bot perspective of but it also uses chat captures private private data. Right. That's how I'm thinking about it in a very loose way of an enterprise app. What is a non enterprise app or what I call the lightweight app?

00;14;39;11 - 00;15;19;06
Dmitry
That makes sense? Yeah, I think that's a great distinction. I believe what we're seeing again as as more and more people are able to build applications and more and more of these AI powered applications are deployed. Yeah. Is that this old way of thinking about enterprise apps as being like these rigid and and robust? I put it in air quotes, whether they're robust or not is questionable, but like these at least percent perception wise, they feel like robust applications on that.

00;15;19;07 - 00;15;43;26
Dmitry
This is a new look if you think about it, like what are applications for? Like, why do we even use applications? We use them to solve problems, right? There are tools we use to help us do our work to solve problems, correct? And as long as the application solves our problem, correct, arguably better application are the lightweight applications that fit into all the nooks and crannies of our unique workflows.

00;15;43;26 - 00;16;14;15
Dmitry
Yes. And just solve our problems. Individual in know, robust enterprise application. Correct. Actually can do that because all robust enterprise applications must be built for a large, you know, groups of people. That's the only thing that makes sense economically. Nobody's going to build it from my unique workflow. And then I as a business user enforce to sort of create my workflow around the abstractions of those applications and how they're all put together, correct.

00;16;14;18 - 00;16;21;17
Dmitry
In this new world of generative AI, each one of us in a sense, can be our own app developer.

00;16;21;19 - 00;16;22;07
Chet
Yes.

00;16;22;10 - 00;16;36;08
Dmitry
And can get exactly something that fits us like a glove. Yes. Like water wraps around things, correct. From the receptionist to the CEO, because every receptionist has a different workflow.

00;16;36;10 - 00;16;37;07
Chet
Correct.

00;16;37;09 - 00;16;56;03
Dmitry
In digital, receptionists can see in their own workflow where inefficiencies lie. Yeah. And once they understand that, they can simply build an application to fill all of that in, Yeah, that's a radically new approach to software building and software using.

00;16;56;06 - 00;17;04;07
Chet
And the reason why I was smiling as I was what I was going through my head is everybody has different childhood issues and hence they have different working.

00;17;04;07 - 00;17;05;23
Dmitry
Styles, right?

00;17;05;25 - 00;17;41;13
Chet
It doesn't matter whether they do the same job or not. And so you have to make it work for you, even though the output may be exactly the same. Right. Some people can do it in an hour, some people need 2 hours and the output might be the same. Now, when you're thinking about this and when you're thinking about this in the enterprise context, right, whatever it might be, I'm assuming there's some level of templates that you kind of give people so that you say, you know, generally here are the guardrails, right, where you actually can and cannot do.

00;17;41;13 - 00;17;57;01
Chet
But once you want your resort, here's some templates that you can go off and build for yourself. Here's an example, one that we built for all receptionists as an example. Is that the way you envision it, or you just want to say, Here's a blank sheet of paper, just go to town? Or a great.

00;17;57;01 - 00;18;25;21
Dmitry
Question. I like. The answer is both are clearly human beings struggle when just given a blank sheet of paper to do something with it, you get a blank sheet of paper to somebody and say, Draw something. They get stuck. Like, What should I draw or write something? Well, what should I write? But if they see some patterns of other people doing it correct, they very quickly grok it and say, Oh, I see what they want.

00;18;25;21 - 00;19;03;09
Dmitry
General, and they can start doing so. I think templates are extraordinarily powerful and the ability to sort of make copies of other people's work, if they allow you to do that, do that and use those as scaffolds are extremely powerful. And so this for example, in my own studio, we support that. We support, yeah, the ability to take something, make a copy of it, just like with get you can clone it and start there it but at the same time generative I dramatically reduces that sort of writer's block of a blank sheet of paper.

00;19;03;11 - 00;19;11;11
Dmitry
So for example in MindStudio, the AI itself can generate a lot of the scaffolding for you.

00;19;11;14 - 00;19;12;07
Chet
Yes.

00;19;12;09 - 00;19;32;06
Dmitry
Rather than you having to find a template. So we we have this thing called prompt generator is one of the pieces of of building an AI. So you built like these workflows, correct, multi-step workflows you can to one or many alarms, various data sources and all of that. And part of that is also sort of giving the system some instructions.

00;19;32;06 - 00;19;59;28
Dmitry
I'm like, what is it? How it behaves? And you can start with a template. You can start from somebody else's work and make a copy of it. You can start from blank or you can start from this prompt generator. And the prompt generator says, What are what do you want to build? You can say, I want to build, you know, a system that looks at this data source here in extracts insights, correctly read reads between the lines great.

00;19;59;28 - 00;20;15;08
Dmitry
And it will simply push the button and it will write all of the scaffolding for you and you can edit it there. That's how this was impossible before. Generative API. Yeah, you need a template, but now the generative. I can create custom templates for you on the fly.

00;20;15;11 - 00;20;39;00
Chet
You know given give it your awesome experience. Right? You would agree one of the hardest problems that Jenny is going to do is to is is going to solve is changing people's behavior. Right. I mean, just to give you an example, we were looking at, you know, going to Australia and what would we do in Australia and Google, you would just say, give me an itinerary.

00;20;39;02 - 00;20;57;12
Chet
But in Egypt you say, give me an itinerary and then you can go deeper and deeper and deeper into it and say, Give me this for someone who likes doing this, who does not like doing this, this is the kind of food and that I think the prompt engineering part of this is, I believe what happens in in 2024.

00;20;57;14 - 00;21;16;04
Chet
Right. I don't think people I think, you know, the haiku era was the haiku. The haiku year was like 20, 23. I think we're beyond that. And now people are getting I'm talking about not the early adopters. I think the rest of the population is getting more and more comfortable with prompt engineering.

00;21;16;06 - 00;21;33;07
Dmitry
So I think there are problems there. I think there are a couple of things that I'd like to push on there. Computing didn't really take off until we went from DOS to Windows.

00;21;33;07 - 00;21;35;18
Chet
So yeah.

00;21;35;21 - 00;21;40;22
Dmitry
Most people have not learned to use Google search operators.

00;21;40;24 - 00;21;41;22
Chet
Yes.

00;21;41;24 - 00;22;26;05
Dmitry
Even though it's a really powerful thing. Yeah. And while we nerds like to do things in command line interfaces, the regular and user should not be thought of as that type of a nerd. Yeah. And so I believe that prompt engineering is not a skill that end users will or should pick up. But in fact, I believe that all models, whether they are language models or image diffusion or code video, whatever they are, are best used as back end services, cloud infrastructure.

00;22;26;08 - 00;22;55;11
Dmitry
Yeah, let's call that the intelligence layer. Correct. And that can users should not even though they can because chat exists and cloud exists, even though they have that capability to learn how to think like a prompt engineer or whatever that means doesn't mean anything but correct saying the same thing. Right. The nuances of how these models might react and how you manipulate them, even though that's possible, that that's not practical and that's not the right approach to doing it.

00;22;55;16 - 00;23;21;15
Dmitry
And in fact, that the intelligence layer should be abstracted by an application layer where nerds show up in package things for end users to use and just push a button and use. Yeah. And so I don't believe that 2024 is the year that everybody becomes a prompt engineer and starts treating these things like Google, even though they might attempt to do that.

00;23;21;20 - 00;23;49;24
Dmitry
Correct. But, but I submit that the reason chatter beats usage has done one of these jobs went up and then crashed down. Generally. Yeah. Is because people realize while it's amazing that you can do that if you look at the effort required, it doesn't make you more productive in many ways. It makes you less productive, makes you not do some thinking necessarily because it can enumerate things.

00;23;50;01 - 00;24;14;16
Dmitry
Give me all the variance of what I can do in Australia. Okay, that's cool. Okay, so for that, that's not another side. I agree with you deeply that the ability to interact with information by being able to have a conversation with it and to take it in different places, each one of us thinks very different. We figured out this is why I was a poor student.

00;24;14;22 - 00;24;17;20
Dmitry
Yeah, is I don't follow most people's reasoning.

00;24;17;27 - 00;24;18;11
Chet
Correct.

00;24;18;11 - 00;24;23;07
Dmitry
But if you let me ask you a bunch of questions, I can very quickly learn things.

00;24;23;10 - 00;24;23;25
Chet
Yes.

00;24;23;27 - 00;24;39;16
Dmitry
If I can be the one in the lead asking questions rather than somebody lecturing to me. And so I think that radically transforms education and certainly usage of information technology is just that we're trying to educate ourselves when we're searching Google or doing any of that.

00;24;39;18 - 00;25;00;26
Chet
That was I was getting ready for a rebuttal on your first point. And then you said your second point. I'm like, okay, I agree now by we're in we're in complete agreement. I do think apps will have its spot. But I think I really do think that there's a use there's some user behavior changes on how they interact with certain technologies that will show up in the next few years.

00;25;00;29 - 00;25;03;16
Chet
It may not be 2024, but I think what the.

00;25;03;16 - 00;25;38;02
Dmitry
Ability to go seek out knowledge, whether it's the travel itinerary or grounding or subject or anything like on Google is antiquated Google search. Yes, typing in keywords, even though behind the scenes it's using AI anyway to expand all of that and do fuzzy and mean like all that on conversational is better. Yeah. There's some ranking things that need to come into play, meaning large language models have deficiencies as compared to Google Search as well.

00;25;38;04 - 00;25;46;14
Dmitry
Those things need to be solved. Yeah, but that is absolutely, in my opinion, the better approach to engaging with information technology.

00;25;46;16 - 00;25;54;10
Chet
So a blank sheet of paper. What's your prediction for Jenny this year.

00;25;54;12 - 00;26;28;04
Dmitry
On more models, meaning more language models and many efficiencies and language models? Open source plays a big role. You know, small models become clear that they shine in many ways above large language models where they're made just for a specific person purpose and that they operate faster, cheaper and give better results than than simply using large language models.

00;26;28;04 - 00;27;03;24
Dmitry
So things, you know, image diffusion models are already amazing in a bunch of ways, but have issues with consistency. And so we start to see that there video comes into its own certainly this year and I think generally gets up to par with image generation. And so you can generate extraordinarily compelling, realistic, long, longer form, you know, video code continues to level up.

00;27;03;24 - 00;27;42;19
Dmitry
That's happening really quickly of what's required to build an application. So yeah, all of those things and the rest is hard to predict because I very much respect and like Sam Altman's point of view, I think about I yeah, and I think by the way, that sort of creates some problems at Openai because on one hand everybody was brought in to go in and sort of search for ragi in the mission.

00;27;42;19 - 00;28;16;06
Dmitry
It's like, let's go to Mars. Yeah. And then all of a sudden they said, well, they said, well, let's launch this chatbot thing. Maybe people like and people said, Oh my God, that's it, That's your business now. And there's just like a you know, Paul, there's AGI or like, let's build cloud cloud infrastructure, right? Correct. And depending on how successful some researchers are in and then along with AI itself in sort of finding those shortcuts to take us to places where, again, the human human already is the bottleneck, correct to leveraging A.I..

00;28;16;10 - 00;28;36;08
Dmitry
This is why we should leave that prompt. Engineering is the right approach or the chatting that, you know, command line interfaces are the right approach. Correct? A picture says a thousand words. It's hard to type a thousand words with your thumbs. Correct. But even the picture is the wrong input into these models. The right input, of course, are large datasets.

00;28;36;10 - 00;28;43;18
Dmitry
And so the key is how much can we figure out how to get datasets out of the physical world and out of the human mind?

00;28;43;18 - 00;28;44;07
Chet
BRAIN Yeah.

00;28;44;13 - 00;28;59;11
Dmitry
To be able to extract that. And once we do that, we can granted to AI. And so we might see incredible advancements and you know, certainly in medicine and in energy, yeah, countless other things. Now that's exciting.

00;28;59;11 - 00;29;01;06
Chet
Times. Oh, very much so.

00;29;01;06 - 00;29;04;10
Dmitry
Very good time to be alive like it is.

00;29;04;13 - 00;29;22;29
Chet
It is phenomenal, right? It's mind blowing. All right. Time for Rapid Fire. I'm going to ask you a bunch of questions very quickly, your responses, and then we'll move to the next one. What's your favorite way to use air in your day to day life?

00;29;23;01 - 00;29;53;16
Dmitry
Favorite? I, I actually I don't very much use air in my day to day life. Like I build a platform that allows people to use air. And I've played certainly with large language models and they've got lots of use cases, but I still write my own things and and intend to do it in that way. So I, you know, other than playing around with some image generation to create some thumbnails, I actually personally don't, don't at this moment rely much on that.

00;29;53;18 - 00;30;06;21
Chet
All right. We'll we'll will double click on that if you if you could have any assistant for anything readymade customize see you fits like a glove like we talked about what would you choose?

00;30;06;24 - 00;30;36;13
Dmitry
I would have a a proxy to the digital world. Yeah. Something that looks at all of the things going on in the world digitally. Yeah. Is tracking at wire speed. I understands me and my needs and sees the world and says oh that's valuable to me. Tree. Yeah. And takes that and packages it in ways that I can understand it and brings it to me rather than I having to do with all of that.

00;30;36;13 - 00;30;47;03
Dmitry
So it would read all my email, it would be my proxy to the web, to content, etc. And by the way, I'm certain that's where we're going. Yeah, for sure that I.

00;30;47;03 - 00;30;48;03
Chet
Would by the way, I would use it.

00;30;48;03 - 00;30;58;27
Dmitry
As well. Stir ups and don't use it. That's going through feeds, looking for good nuggets. Yeah, it's it's horribly inefficient and I agree.

00;30;59;00 - 00;31;05;11
Chet
I agree. One world's problem you hope I will solve.

00;31;05;13 - 00;31;28;21
Dmitry
I think the most important problem to for us to solve in the world you countless really important problems right inequality global warming like we could rattle them off but I think the number one problem is the ability for human beings to align with one another. Yeah, we live in a world where more and more we are getting fractured.

00;31;28;21 - 00;31;29;19
Chet
Yeah, I would agree.

00;31;29;19 - 00;31;55;04
Dmitry
Seeing it in our politics, we're seeing it in our lack of ability. I agree to our fellow coworkers at the water cooler because we're all consuming different information in different ways. And on one hand that's amazing because everybody gets what they want. On the other hand, that's catastrophic because a population that is not aligned cannot solve all the other problems that need to be solved.

00;31;55;07 - 00;32;10;00
Dmitry
And so I believe that's the number one in I believe is absolutely the thing that can do it. If they can understand each other. I agree once and connect us together and translate each other's points of view in ways that we ourselves can understand it.

00;32;10;02 - 00;32;30;29
Chet
It's really interesting. My 21 of my 2024 predictions was was I hope people had more dialog, right? Just just dialog right. Just though and as a as, as a geek, as a technologist, you don't think of dialog often, right? You just think about code and how it affects the world. But I think you're absolutely right. I think it's That's awesome.

00;32;31;01 - 00;33;00;13
Dmitry
Dialog requires patience. Yes. And time and empathy and emotional intelligence and all of that. And and unfortunately, all of those are scarce. Yeah. And so where technology I think can hope is to again, create a proxy. Correct. Where my proxy understands me in your proxy understands you and it can translate to one another how we can communicate with each other much faster, much deeper, connect much more clearly and more nuance.

00;33;00;15 - 00;33;13;02
Chet
So the year is 2034. What's the main challenge we are facing when it comes to technology ten years from now?

00;33;13;05 - 00;33;50;03
Dmitry
There are, and I think many things that I would caveat this with. One is that we make it to that year and still have something that we call technology in the sense I think in the near term we're faced with really existential problems with this technology being used for by bad actors. Yeah, in extraordinarily powerful ways that can shape our world radically.

00;33;50;05 - 00;34;12;22
Dmitry
Now, I'm an optimist and I believe we'll get through it, but we are at this moment, I think in those times, like again, the ability primarily to manipulate people's opinion about reality, we're already seeing that. Not so Bitcoin. If we can get past that again, it's hard to predict. The especially the pace of innovation as it's happening now is in this.

00;34;12;22 - 00;34;44;21
Dmitry
Ray Kurzweil points out information tech technology accelerates exponentially, especially now with generative AI. So again, I hesitate to make the prediction, but I believe the biggest problem is going to be how do we how do we create an equitable world where people who are not in control of technology don't suck up all the oxygen in a sense enslave the rest of the world?

00;34;44;25 - 00;35;13;02
Dmitry
It's very well because people with access to this technology that other people may not have or won't have a competitive advantage, they will have a ludicrous p imaginable advantage. Yeah. And in the way that our economic systems are built now and our systems of sort of understanding how we relate to one another, they encourage a winner take all approach.

00;35;13;02 - 00;35;13;17
Chet
Correct?

00;35;13;17 - 00;35;20;21
Dmitry
It's for the shareholders. Correct. And that does not work with technology like this.

00;35;20;24 - 00;35;30;04
Chet
No, I agree. All right. The last one in a phrase, what does the future hold for my studio?

00;35;30;07 - 00;36;04;05
Dmitry
We're building the application layer of A.I. models belong in. The intelligence layer is back, ends. They are abstracted from humans by the application layer. We're just getting started. We've got over 13,000. AI's deployed in big enterprises, small enterprises, consumer, etc. All of that is a little blip on the again, once the application layer gets even more mature, even more applications are there, even more, more people are able to create these things.

00;36;04;08 - 00;36;13;00
Dmitry
That's what MindStudio is in success. We are at the interface between humans and intelligence. That's awesome.

00;36;13;03 - 00;36;18;16
Chet
Very well said, Dmitri. This has been a joy to have you. I think the.

00;36;18;18 - 00;36;19;10
Dmitry
Our.

00;36;19;12 - 00;36;25;24
Chet
Audience will absolutely love this. I really, really appreciate the time. And thank you very much for joining us.

00;36;25;27 - 00;36;27;05
Dmitry
Great chatting. Really enjoyed.