Average person is just dealing with life, hearing all these crazy things called AI agents that are gonna take their jobs. They're scared of it. There are companies that are trying to adopt AI, but their employees are resisting, not realizing that there's no way to stop this adoption. To build a billion dollar company, you need a moat. And if you vibe code something that takes off today, later today, I'm gonna be a competitor and a million other people are gonna be competitors. Without knowledge work, what becomes valuable? The new skill set is not technical mastery of of code. The people who tend to have valuable existing skill set are people who have a mastery of communication. Existential threat is the collapse of reality. United States of America is seemingly on the verge of some form of a collapse caused by misinformation, disinformation. How fast this is moving, is there anything that worries you? Whether it's for tech, business, humanity, could be a problem. Yes. Ray Kurzweil has been predicting for the last twenty years that singularity, this moment where technology gets out of hand and can do like humans can't possibly even understand what it does, was gonna be twenty twenty nine. If we fast forward to twenty thirty, what has AI changed most significantly? I think the answer is. Alright. Welcome to AI two thousand thirty. We have Dmitry Shapiro here with us, founder and CEO of MindStudio. Welcome. We also have Chad Lohrey, the what are why don't you give your credentials? I don't know. Founder, chief AI officer, whatever that means. Glorified tech guy at CADRE. And I'm Keith Jensen, president of CADRE AI. And, yeah, I think we're gonna have a fun conversation today. We've got so with my studio and actually, by by the way your background for anyone that doesn't know you is pretty crazy, Google, MySpace, Music, CTO, Vio, Iconix and now My Studio. It seems like you remember things better than I do. Yeah? So My Studio, we have, it is a platform that's allowing people that are non technical to go and build agents. Question for you is, where's the biggest friction here? Like, why isn't everybody using this? Like, what if, if one thing got removed, what would allow, would just be like a floodgates are open of people coming in and using this platform? We live in an attention economy, and there are lots of amazing shiny objects that get people's attention. There's a limited amount of time, and so lots of amazing things are competing for attention. And if people, you know, are in in the right, you know, mood and and discover Mind Studio or, like, many other platforms that they might be interested in, they quickly learn to use them, and then start to realize that there's a lot of value that you can extract by automating things that today we all do manually. Right, hon? But even like, agnostic to the platform of MindStudio, feel like if everybody knew the capabilities of what agents could do for them, why wouldn't everyone be seeking out platforms to go build, right? Listen, the average person is just dealing with life, and again, hearing all these crazy things about these things called AI agents that are going to take their jobs, and, and, you know, and they're scared of it. So like we certainly see and hear a lot of stories like that, where there are companies that are trying to adopt AI, but their employees are resisting, Not realizing that one, there's no way to stop this adoption. So the more you push back, the faster it's going to come for your job. And the best thing to do is actually to lean in and automate your own job, and therefore you become much more productive, much more capable, and kind of be the last person to be laid off. I think that's one thing. And yeah, again, I think people haven't yet sort of processed and understood the value proposition of automating computer work, know, information technology work, right? No matter what office job you have, I argue you show up in the morning, you have a cup of coffee, and you log into some system and you get some data out of that system into your brain, and then you figure out what's missing and you go do some research, and you kind of enrich it and transform it, and you collaborate with others, and then you put it in some other SaaS system, and that's what knowledge work is, and it doesn't matter if you're marketing, or if you're finance, or whatever, that's what you do. And again, we live in this like crazy world where now you can actually, you know, tomorrow, show up and not do that, you know, still have you a cup of coffee. But instead of doing the work, well, maybe I can build some tools for myself that'll look like power tools that allow me to do this work faster just by pushing a button, or like some things that just run automatically every day because I know I'm going to have to do that, and for the people again that have discovered this and have realized that and started to do it, they have superpowers, and so this word gets thrown around quite a bit, superpowers, and indeed they are, they are superpowers, because there's no way to be able to compare people that can build AI agents to help themselves do things, you know, more autonomously, or augment themselves in some way, versus people that do things the old way, like the, order of magnitude. Yeah. I find that like whenever you, when everyone's, someone's new to AI and they're opening up ChatGPT or Claude or whatever, and you're like, cool, you can, you can do whatever, just, just ask it. It's almost like there's this block where they don't understand what to even ask. They're like, what were the lyrics of that song? Or like, what's a good recipe? And they're they're using it for the simplest stuff. I wonder, and and Chad and I have talked about this a lot of, like, the idea of just we have all these meeting recorders that are just, you know, listening to us on meetings. I wonder how how far we are from us just having a meeting recorder plugged into Mind Studio or something where there's, it's going, I know you have this problem. The AI recognizes the problem you have and says, I can go build an agent to do that, and maybe it's done by the time they're off the call, like how far are we from that? We're already there, companies are doing that, yeah, we're already there, you nailed it, like we have all this data that's available about like what's happening in the business. Calls are being recorded, we have email, we have Slack, everything's digital now, so like everything, everything's digitized, and you know, especially like, you know, more mature companies, they've got data warehouses, and and they've got, you know, kinds of abstraction layers, and they are plugging in, you know, the building agents with Mime Studio, and and the role of those agents is to watch what's happening in an organization to find insights, and some to go off and make things better already, go refactor the business processes, create new tools, and do all that. Yeah. We've been there for a year. For the less mature companies that don't have a data warehouse, or let's say they still have data all over the place, how do they get started? Right? And ideally, you know, the more mature companies already have this and they can just plug into a platform like Mind Studio, but if you don't have your data all in one place, what do do? You still plug into a system like Mind Studio and you have it help you get the data, and so your data might be in twenty different sources, great, you tell the agent, here are the different places where this data lives, they're not integrated together, but that's the great thing about the agents, like agent doesn't care if it's integrated, it's the integration point. So as long as it knows about the data and has access, you know, credentials and has access to the data, then it can get the data, and then for men you can do the same thing the big companies do. So like in many ways, like those, all that infrastructure that they, you know, built out of like data lakes, data warehouse, I mean, still need those things, there's like other things they can do, it's not efficient to do it necessarily agentically, but for small companies, they never have to go that route at all, or medium sized companies. Are you saying that it's it's small companies that don't have a ton of data, it's more cost effective to go the agentic route than to invest in Snowflake or Databricks or something? I think that's clear. Yeah. Yeah. Interesting. I feel like the biggest thing that companies want, no matter the size, they want an AI interface that they can just chat with their data and ask a question, and it gives the answer that they're looking for. And from my understanding, the biggest issue there is, sure, you have MCPs or you have any kind of connector APIs that are going in and grabbing data from different places, but how does AI know what the join keys are? Like, how does it understand that, I mean, something as simple as a, you know, you're going from your Google Ads that has, you know, someone clicks on an ad, it goes in, there's a campaign ID that gets passed into HubSpot. HubSpot, they and they fill because they fill out a form, that thing gets passed into Salesforce as on the lead ID, but then the lead becomes a contact, and then the contact is an account, and the account is an opportunity, like, and you want to know, did this campaign produce revenue? Are we there yet, or do we have to actually help AI map out those paths? I think it's a really important question. And the answer to the question is another question, it's like, how does a human know, like when you bring on an employee, so like nothing's automated, like somebody architected the system that you're describing and it exists. And now there's a new employee and their job is to drive the system. When they show up, they need to learn how to drive the system. You've got some onboarding, somebody helps them. You got some standard operating procedures, you got some documentation, or you just say to them, there's nothing here, just like, here's the code base and go figure out like how this thing was built, right? And then they figure it out, and then, so if a human can figure it out, then yes, we're already there, all of these agents can figure it out, you know, we were talking about before the recording, Claude Code, is kind of this like new super popular, you know, coding agent, right, there's others. Claude Code doesn't need your help onboarding really, as long as the code base is there and the connections exist and has access, it can figure it out, no matter what kind of slop, by the way, that the developers of that system that you're describing implemented, it'll figure out like, oh yeah, I just need to do this or build new connectors, and but that but I think you made an important point there of, like, if a human if if you don't know how to do it with a human, meaning, like, I can't figure out how to connect my Google Ad data to my Salesforce, you know, opportunities, because maybe we're not storing the campaign ID, and we're not passing it through. If you can't do it with a human, we can't expect AI to do it. It's the garbage in garbage out type of methodology. So we do need to be able to ensure that we're not asking AI something that's impossible to do for a human to do. Unless your proposition is that most AI are smarter than people? And more technical. Yeah, and so again, if it's integratable, if it can work, then we not only should expect AI to do it, we're already there, where AI will do it, where again, something like Claude Code can take a big mess of unintegrated systems and figure out how to integrate them, even when the people that built those systems might not have been able to figure out how to do it. How do you feel about all these AI startups that are coming out that are honed in on one specific problem? Do you feel like those are going to be a pop and go away? Because, hey, anything that they could, you know, get hyper specific on a specific industry, any knowledge that they're feeding in is eventually gonna become knowledge that every LLM has, and they could have just used MyStudio to build whatever they built, whatever they used that one individual tool to build. I think your question is pointing at a, like a really big, big elephant in the room, which is in the world of agentic AI, with more and more power, and the ability for AI to write code, which were there, and to write code to do anything, and all information technology is code and data, right? And so, if code is available to everyone and anyone can write code, then only, you know, sort of like asset becomes proprietary data, and so if these startups are accessing or compiling somehow proprietary data, that's not easily available in public so that, you know, anybody else could just get it, then I think that has some value. Although again, others could come and compete with that, but, but like, I've built three venture backed startups, you know, companies, raised over a hundred and forty million dollars, and pretty much in every VC pitch that I've ever done, there have been the questions of like, well, what happens if, you know, it used to be Microsoft does it, then it was like, what happens if Google does it, you know, then after Google, I'm like, let's not worry about Google doing it. I've been there. Good luck. Okay. But like the things they were pointing to is like, you know, like some bigger company could come and rebuild what you're doing. Well, now it's like, what happens if anybody else does it? Well, like I see people saw Gary Tan recently make a post, but there's lots of people who've made postage, Jason Galcanas is another one, maybe pick on him, that made a post about the like one person billion dollar startup because of vibe coding, somebody's gonna be able to vibe code a billion dollar app, and I was like, to build a billion dollar company, you need a moat, and if you vibe code something that takes off today, later today, I'm gonna be a competitor, and a million other people are gonna be competitors, And so now you got a million competitors, and no matter what, you can't build to a billion, you know, and and so I think like the the this democratization of everyone's got the infinite number of the best developers in the world, basically, at their whim, completely changes what business is. With that said, like, where's the edge? I guess, where's the the arbitrageable opportunities in this case? You said data is one. Right? But, like, if you're a new company, you want to start a company, you're not going to have data, right, as your moat, so Why do you, why do you want to start a company? I don't know. That's the fundamental question, like, it used to be you want to start a company because you want to build some product, and you want to have a moat, you want have some exit, maybe that's why, you want to be your own boss because you're insufferable, and nobody else wants to work with you, and you got to be the boss, okay? You know, Andy Warhol talked about everybody's fifteen minutes of fame. I think we have a similar situation here, where, again, in the past, you needed to start a company because building products was expensive and took a long time. You needed to hire people. You needed to have a long road map, etcetera. But when building products is a matter of Claude Bot building it for you, you went to sleep and you say, surprise me. I'm gonna get up in the morning, surprise me with a few products. And you wake up and you've got a few products. In a world like that, even if you have your one day of life cycle of your product, you built some funny little photo generator that turns people into cats, and all of a sudden it went viral, and you made ten thousand bucks, maybe you made a million bucks or ten million bucks in that day, because the whole world decided to turn themselves into accounts, and then tomorrow your product's completely worthless. But so what? Claude Bot built it for you and it costs you five dollars in tokens, And so you start getting basically, I think like this glut of products, which on one hand is amazing. Another like, what is product hunt in a world like that? Well, it's a million products a day. Okay, but then what? You know, was talking to someone recently about like, okay, without knowledge work, what what becomes valuable? Right? And you've heard Jensen, Elon and so forth saying, you know, plumbers maybe, or people who can do HVAC or people that can build things, right? But we've got this whole generation, Gen Zers who grow up on their phones, doing digital things and they probably can't hammer a nail in a wall, right, and so you're going to have this gap in folks that can actually build things, you do, do you believe in that or what are some of the other jobs that Those jobs are gone too. I mean, look at the humanoids that are coming. How far how far are we from that? Again, I think we're there. We're seeing prototypes of these in warehouses, and there's nothing that points to this thing stopping, know, do you have a bunch, do you have Rosie in your home walking around making you coffee? Maybe, I think, whatever, that's, but I don't think that's like the use case. And the use case is like all kinds of manual labor, certainly plumbing. I will get there. I don't know when we've got like those plumbers. Maybe it's five years, maybe it's ten years, maybe it's twenty years, but no matter what, it's not But a much longer than life horizon than knowledge work. Right? Oh, that's right. Knowledge work. Or at least for those that are like, what comes next? Right? You know, I I think of, like, self driving, you know, Iman said we'd have it ten years ago, and it's good, but it's not perfect, and so robotics is probably gonna have that same curve unless we're truly in a massive exponential and it gets shorter and shorter. But, well, we are. I mean, yeah, we're in the massive exponential. Like, we're seeing it with information technology, right? Like robotics is a function of research in software and materials and all of that, and research is accelerating because of AI, because AI is being used to build the next generation of AI, right? So like this version of Clogged Code was built by the last version of Clogged Code, each one gets faster. Now that the machines are tied into code and they can just talk twenty fourseven, just a function of like, can you pay for the inference costs? Like innovation across any industry is going to accelerate dramatically, so I think like all those things get compressed. I think it's a more fundamental question, just kind of again, I don't have an answer to societal question of like in a world where, you know, why do we need to work? We need to work because dynamics that have been set up by, you know, our society that says well in order to be able to eat you need to work, right? But you could theoretically have a society where, you know, you don't need to work to eat, right? But then humans obviously need something to pay attention to, and no matter what they then become, they start to work. And that work might be because they flip burgers because they love flipping burgers. I, confession, I love flipping burgers. My happy place, if I could just go and run, I don't want to run a restaurant, I don't want to run the restaurant, I just want to flip the burgers. I like making burgers, I love cooking in general, I just want to cook. Let me cook, I'd go cook, would I work in the kitchen? Yes I would, if I didn't have to go and earn you know, salaries of IT people, I would go and I'd flip burgers, and so I think you get people in the right positions, theoretically the world becomes better if we can figure that out, but no matter what, we'll have to figure it out, because here we are, and and there, like there's, I don't see how you stop it. You must see some really interesting use cases at at MyStudio because you've got, anybody could be using it. Do you have, are you seeing any like commonalities of it's people in product that are like the the, I I don't know that you would call somebody an early adopter at this stage, you've been at it long enough, you probably feel like people jumping on now are laggards on this. They're not, they're still actually early adopters, I'd consider, yeah. Do you have any, is it industries, is it people, personalities, or anything you could say that like the people that are really leaning into this right now are? And to add, do you serve mostly B2C or businesses too? So we are a self-service platform, so anyone can show up, individuals or businesses, and sign up and learn to use it and start using it, and we've got a mix of all of them, over four hundred thousand of these AI agents that have been built and deployed by individuals, by SMBs, mid market enterprise, government, etcetera. The people that tend to find us are because all word-of-mouth, we have no salespeople, zero. So like everything is product led, everything is bottom up, everything is word-of-mouth. Even like all the enterprises, some random person in the enterprise discovers it, brings it in, starts using it, know, trains some other people, and all of a sudden you get an enterprise deal. So that's quite random of like who finds and starts using it. And so like every job function is already represented from the receptionist to the CEO and everything in between. But the people who tend to have a more, I think, existing skill set are people who have a mastery of communication. So the new skill set is not the mastery of, you know, technical mastery of code. It's a mastery of being able to articulate clearly what it is that you want done, how you want these things to function. And so indeed product managers do very well here, and there's lots of product managers using it because one, they have a desire to build things, and this is amazing to build things. There's always a bottleneck that they've experienced, which they've got an idea, but now they've got to convince a team, an operator team of developers and designers and copywriters and all that to do that. And now they don't have to do that. They're just gonna have an idea, and congratulations, you're now the boss, that's amazing. They also have obviously, because their job is to communicate to these different teams and to write product specs and all that, they have a more practiced, I think, skill set or an aptitude that certainly benefits in the world of AI. I remember living in New York City and seeing billboards, was a long time ago, I said billboards for when iPhone was getting better with photos, right, they would have a picture on a billboard and it would be, you know, taken with an iPhone by, you know, Samantha or something, just showing that like, damn, you could take an insane photo and then, you know, you know they didn't take that small resolution thing and blow it up, they did a lot of extra work on it to make it enormous, but but I wonder, like, from the the MyStudio side, you must it must be really interesting to see all the stuff that's created using, because, you're you're an agnostic platform in terms of, you're not building a specific they can use it to build anything, and I would imagine your marketing campaign, if you ever did one and you needed to go out and it wasn't all product led, probably be just showing off the insane stuff that's been built, are there any things that come to mind that just kind of blew you away that somebody was able to create? I've been asked this question before, and it's always a tough one to answer, because the things that **** ** away mostly, is just how common the things are that have huge ROI, my favorite way that people describe Mind Studio is, it's like AI duct tape. Duct tape is not a very sophisticated tool, but duct tape is an extraordinarily valuable tool. Propose duct tape, you must have duct tape around all the time because it's got so many uses, right? And duct tape allows you to take things that aren't quite connected together and kind of like slap something on there, and all of sudden they're connected, and is that by the way the right way to connect them? Like if this table all of sudden sort of like breaks, this little piece comes off of the veneer, is the right thing to do is like, so like duct tape it at the bottom, no, probably not, you should find like another piece and like glue it and whatever, but that would take a long time and cost a lot of money, so instead you just have a piece of duct tape on it. Now, AI duct tape is much more valuable than regular duct tape, because it comes with intelligence. And so AI duct tape can bridge the gaps of poorly integrated systems. And if you think about it, every organization can be described in I think in two ways. One way is as a collection of people, with a common goal, you know, with different job functions working together, leveraging really powerful tools, we call it the stack, the tech stack, and tech stack consists of on average now or median, I think like one hundred and thirty SaaS products for a mid market company, that's the tech stack, and humans are there showing up in the morning, have their cup of coffee and like running this machine of all of these stacks, right, and then creating some positive yield, so that would make sense to describe it, you know, crudely a company like that. But another way you could describe it, be that a company is a stack, which is a collection of tools that are then sort of run by these humans. And why are the humans there? Is because the stack isn't well integrated, and the humans are the glue that glues the stack together because there are so many fuzzy things. And when there are fuzzy things, we must need these humans to come and glue the stack, but like what if you could just use some AI duct tape to patch this up? Well, now all of a sudden you can free the human from being human connective tissue tape and to do something like more valuable, or get rid of the human I guess, there's also the other option, but I think most organizations now are seeing like you actually don't need to get rid of humans, you need the humans who have learned to apply duct tape, you need more of those, you need more humans that can apply duct tape, because an organization is constantly morphing too, and so it's just fill in, another way of maybe thinking is like intelligence filler to fill all the poorly integrated nooks and crannies of things, you know. To add to that, so, you know, obviously you can ask the question, once you have all the humans start applying duct tape, then what do the humans do? Right? But I think what you're saying is a company will consistently evolve, so you are constantly reconfiguring that duct tape versus saying, hey, we're gonna use your humans, put in the duct tape and then get out. One hundred percent. Competitive advantage I think becomes, for companies is like what company is able to continue to evolve, because one, the capabilities of the duct tape, if we stick with the duct tape analogy, continue to evolve and evolve exponentially, at least for some period of time, until I guess the singularity that seems to be near. And And two is as you start to, again, once you've patched everything up and it's running and now you see your organization running autonomously and you're stepping back and you're like, well that's dumb, like I know I did it that way because that's where we were coming from, but now that I can see it working, well if I was gonna do it again, I would completely redo it, like I completely get rid of the system right now I'm bridging, because we don't need that anymore, we can just go from here to there, and so you're constantly refactoring, and again the amazing thing about, you know, this world we live in and certainly with Mind Studio, is that the time to refactor, the time to build these new things, the time to grab a piece of AI duct tape and apply it, it's like five minutes to thirty minutes, like that's the median. We're not talking months or weeks, and it's for one person, not teams, and so if you need to have a meeting to plan something, you've already squandered your time, you should have just built that piece of duct tape, don't have any meetings. Say that a little louder. What's stopping someone from using Cloud Code to rebuild Mine Studio? Nothing. Yeah, nothing. Yeah, we're in the world where, again, if you're gonna use Cloud Code to rebuild Mine Studio, I actually think that's probably not the best use of Cloud Code. Just use Cloud Code to help you build your business thing, and you kind of don't need Mind Studio. In that world, I think this becomes fascinating, you know, because I obviously also get asked this question as like, well, what what does that mean? Like, you're running a company and like, do people just need cloud code? Is that it? Right? And the thing I point out to people is like, forget like what the stack is and what's being used. The the biggest problems is the human computer interface problem. I've had friends that a few years ago with like, you know, especially after a couple of cocktails would like argue this thing of like where we're going is that you tell AI come up with an idea for a trillion dollar company and run it for me, go. And like, and that's where we're going, I swear, and we're kind of going there, right? But even if you could get there, that doesn't make any sense, that's not, first of doesn't make sense because if you can do it, everybody else can do it, since there's no point building any companies, there are no trillion dollar companies anymore, everything gets diluted, but if you ignore that, even if you could have something that like was so autonomous, unless again, you just kind of don't want to be involved at all in business, that's not what you want, what you want is the ability to be in control, and to understand what's happening, and to be able to impact it, and then not do the heavy lifting. And so you just want a better interface to be able to do more with things. And the problem is the chat interface, or conversational interface, is not the right interface for the vast majority of things. It's the right interface for certain things, to ask a question and get an answer. But to be able to articulate, anything, and we were talking about as we were walking in here, the people are starting to build with Mind Studio tools to produce, you know, studio grade images and videos, to produce motion pictures, to produce explainer videos, to produce TV commercials, and a whole new slew of tools to be able to do that. Like, how do you articulate to an AI model camera movement and framing and various other aspects? If you have to do that in language, that takes too many words. What you want are nice draggable three d interfaces with visualizations that say, let me move cameras around, let me move lighting around, and let me see all the variants, like that has nothing to do with being able to chat with ChatGPT. And, again, the cool thing with it is now, you can build those kinds of interfaces that become front ends to all of these models that today, their front ends are chat. If I understand this correctly, you're saying that, we're not necessarily building interfaces for the sake of building a a pretty UI, but we're building interfaces to help us input more high fidelity information to a model. So we don't have to keep refining our prompts and getting better and better at prompting in order to make something happen. You could never refine your prompt to do these things. You that it would literally be impossible to refine your prompt as a human to do that, because you would need to say too many words and think in too many dimensions, you'd never do it because it would be cost prohibitive, right, even with like a prompt helper. And so yeah, the ability for humans to be able to tell AI what we want done, that's the bottleneck, that's already the bottleneck. Like if all innovation stopped in AI model, you know, development, if we had no more innovation, we are the gap between our ability to tap into that, and what we've been doing so far, is giant. Like there's so much more potential out there with the technology that already exists. I think we're barely scratching one percent, I don't even know how to quantify it, but we can't, yeah, exactly. Let me ask you, with how fast this is moving, is there anything that, you know, actually worries you? Whether it's for tech, business, humanity, like, there anything about where we're going that you're just like, oh ****, that could be a problem? Yes, I think there are like countless things, let me pick a few if we want to scare folks. I think our nearest term existential threat is the collapse of reality, and we're seeing this, for example, in our politics today. United States of America is seemingly on the verge of some form of a collapse, and that collapse has been caused by misinformation, disinformation, our adversaries manipulating us with information technology to fight with one another and questioning reality, along with not just our nation state advertise adversaries, but, like, just anybody trying to make a buck because chaos, you can make fractions of pennies if you create chaos. That's becoming one, can scale that now infinitely, and so you can run a thousand YouTube channels in from your basement, you actually don't need a basement, it's in the cloud, right, like so, and that's happening, and it's happening in a massive scale and you know, like, if you're paying attention to it, you'll see that YouTube has changed dramatically, that so much stuff on YouTube or TikTok or you pick your platform is now AI generating and it's compelling. It's like Elon Musk selling you Bitcoin or something, right, on, like, all these live feeds, I mean, I see them all the time. The things that have taken, for for a moment, again, everybody gets fifteen minutes of fame. Right now, fifteen minutes of fame are for these channels that have, like, old celebrity meets their younger self, and they embrace, and the camera pans to the next celebrity meeting their younger self, or Lassie, saw the other day. So you have, like famous pets that are coming back in, you know, these things are awesome and nostalgic and compelling, and they're bite sized, and so these things are getting like many millions of views, these channels have like two videos without millions of followers, right, it's like crazy stuff like that, and they're automated, right, in fact actually, tomorrow is, I don't think it's come out, comes out before that, but, tomorrow I'm teaching a four hour workshop, specifically on that, on how to build with Mind Studio, these types of pipelines where you can just build infinite number of channels about anything, and they don't have to be video, they can also be blogs, you can build media properties, people have done that, where there are now home media properties that are completely do their own research, create their own content, distribute their own content, and you just set and forget it and scale. Give it an angle, or don't, you can come up with its own angle if you want. Yeah, this like again, this vulnerability of humans to believe what we see with our own eyes, but now for that to be manipulated for nefarious purposes, I think is the existential threat we see now, and there's like all the job collapse and like So I haven't been on social in like two months, my wife and I just said, hey, we're gonna take like a break, and it's been ******* liberating, I love it, but I'm jealous, you know, at Cadre we post some stuff and I see on Slack, oh go, you know, check what we just posted, so I went on Instagram to just go view that one thing, and it opened up instead of opening up that directly, it opened up the home feed, and I saw it and I was like, I just scrolled one to two, and it was and it was clearly an AI ad and then another AI like content creator, and I'm like, I can pick that out now. Like I, I, we're still at a point where I were like, I can pick it out, but I almost feel like to your point of, like, to getting get us getting drawn apart by all this content going out there and and contrarian views and conspiracy theories and all this stuff that, makes us kind of question the reality. We're at this point now where we can still identify it if it's fake, but we're pretty close to not being able to. We're extraordinarily close, and again, I would propose, while you saw some and you realized that it was AI, people are being fooled already. Oh, yeah. If you're not looking out for it, if you're not as savvy, my mom, you know, she's eighty six years old, and some of the stuff that she points out, I'm like, mom, that's not real, she's like, do mean that's not real, she doesn't understand the concept of AI, it's like, or like even with me, and like, I've been posting some AI stuff, and she's like, I'm like, that's not Nimon, that's AI version of me, she's like, I don't understand you, when did you go there? I'm like, I didn't go there, you see, and so like, yeah, so like, we're already there, and we're certainly going to be way, way over that line of like, that you can tell, certainly within a year. There's this, there's this thing called like the dark internet theory, right, which is, you know, as more and more AI generated content gets fluttered on the internet, the internet becomes useless and it's just bots on bots on bots, and I think what that leads to is, you you, where do you get information from? You know, trusted peers, communities, your your your work friends, if work is still a thing. That's what I mean by this collapse of reality, is is that, one you need sources of truth that you trust, because we want to trust reality. Reality isn't something we believe, reality is something we count on, that's our fundamental thing, but if you don't have reality and all you've got is like, think I generally believe this, well that's a hard place to be, and very manipulatable obviously. And two is even if you have sources of truth, but they are then biased because they are humans also in some way. But again, people are using agents to help with that, and so you were asking me like, what are some of the use cases that have kind of blown me away, I guess. Humans are obviously, we're creative, and so we find we can build our own prosthetics, and in a world of information overload, people are building things that help them process information, so lots of things people are building are things that like monitor the web, or monitor social media, databases, monitor sources of information, and so the human doesn't have to pay attention to that anymore, they can just go monitor it. Like one of our customers is a giant newspaper holding company, they operate a bunch of newspapers. They've got over four hundred Mind Studio agents now that they've built, by the way, one non technical person in the world built that, and they do all kinds of things that used to be done by humans, like news gathering. It used to be journalists, they cover court cases, journalists used to spend an hour a day going and sort of scouring the web and various sources for like what's been happening with the courts, well they don't do that anymore, there's an agent that does that. How does that agent determine if the content is real? That agent, when it's set up, is is taught, is configured with here are the sources of truth that the humans would be using if they were doing it manually. An agent just does what the humans did. It's very kind of boring. It's not an agent that goes off and tries to figure out how to get new sources of truth. It just says, oh, here's this like finite enumeration of sources of truth. My job is every morning to go and log into all these places, do the same thing human would have done, which get all of this in, understand it, maybe do some more enrichment, and it can spend a lot more cycles, a lot more time doing things than human would have. And and so so that becomes much better than human doing. It doesn't just replace humans, you know, give human an hour back a day per journalist, but makes it better, does a deeper, better job. The the key hitter is that these sources have to be trusted. Like, you're inputting trusted sources, and you're not you're not saying, you know, go diverge from these sources, like, stick to these. Yeah. And so the second that those become untrusted or biased by AI, right? You get issues. Yeah. Or other things people are building are things that do fact checking for them. You can you can live in a world where you don't trust the sources as long as you trust some sources to then go fact check. The problem is, again, the cost of that and the effort is too great, so we don't do it. But if the agent just does it, and so the agent reads it before you ever read it, all the spam disappears, all the, you know, false goods disappear, or the majority, you you reduce the noise, and now you're only consuming the things that are generally past the filter. So I guess the question becomes, is it the source of truth? Is it like, oh, this is a reputable source like CNN or something like that? Because if you if you have, you know, somebody using Mind Studio to create enough channels, YouTube channels, sources, media sources, it can overwhelm, like if someone wanted to paint a conspiracy theory, they could in theory, go blast it everywhere across all these different sources, and an AI may look at it and say, wow, there's a hundred and eighty seven thousand citations of this being true. Yes, CNN says it isn't, but, like, which win? You're spot on. By the that's how it's done. That that's how we have that's how we have the political environment we have today. This was done to us in twenty sixteen. It was kind of the first year that social media was manipulated to influence our election. I was at Google then. My job at Google was to pay attention to this during that election. It was clear that's what was happening. We didn't know the names of these people, Internet Research Agency in St. Peter's program, Macedonians, like all that stuff, but we saw what was happening that was there, and since then, we've experienced it at greater, greater scale. So that's not a very good signal of truth that it's everywhere, quantity, quantity is not a signal, just complete noise. Yeah. How do we ensure that humans stay in control, right? You mentioned that we want to stay at the center of this, right? We're deploying our agents, but at a certain point our agents are just telling us what to do, and we're no longer the central orchestrator. The question is, are we in control now? Have we been in control? And when you say humans, do you mean just a few humans, or like the average human? So, yeah, so I think it's a fundamental question. I mean, some people have already been measuring it. This was expected, but all all of us are atrophying in our ability to do deep thinking and multi step work, where it used to be we had to go read a bunch of websites to go figure something out, now we ask ChatGPT, it's read the websites, and it just gives us the answer, right, and so we don't do any of that, that's convenient, and that's awesome, it's like calculators, right, Some people don't do arithmetic anymore. On one hand, that's bad, on the other hand, I guess you save mental effort to be able to do other things. I think that's like, like scroll TikTok and Yeah, Important things. Yeah, I think, I, again, I think there, there's a huge danger in very few entities, let's call them, people, groups of people, companies, governments, whatever, having extraordinary power that can't possibly be matched by individual citizens, and and and that's really scary. If we we talked about a lot of like short term, a lot of things we we said like, is it it's like, it's here. A lot of things we thought that it's like, it's here. And even like thinking about, you know, ten years ago, twenty sixteen, this where we started to see the manipulation of of content, and we're ten years later, and it's it feels like now we're at, like I said, this inflection point of like, it's now easier than ever for anybody to to do that. If we fast forward to twenty thirty, your opinion, like, what does the world look like? AI, specifically. What has AI changed most significantly? I won't pick one. Let me see if I can enumerate some. First of I think the answer has changed pretty much everything we do. Again, another four years of innovation. Again, it's exponential. Know, Ray Kurzweil. You know who he is? Ray Kurzweil? Yeah, the futurist. There's a documentary he did like twenty years ago called Transcendent Man you can find. And he talks about, in a quite brilliant scientist inventor, works at Google now, not here, but I'll tell you another story of my interaction with Ray Kurzweil once, where he was having trouble figuring out some basics about the Google campus and needed some help and fun stuff like that. But Ray Kurzweil is extraordinarily brilliant, but he talked, again, twenty years about that information technology especially accelerates exponentially. Humans have a really hard time thinking exponentially. And, you know, he's been predicting for the last twenty years that singularity, kind of this moment where technology gets out of hand and can do like humans can't possibly even understand what it does, was gonna be twenty twenty nine. So that's the year he's been predicting for two decades or or even longer. But that was the specifically the year, not twenty twenty eight, not two thousand and thirty, but twenty twenty nine. So he's pretty close. Again, we're a bunch of it were here now. So another four years, I think we have massive, like Chatuchipiti came out three years ago. Okay. We've come a long way. So so everything's gonna be changed. Some areas that I find, you know, look at the positive ones that that are interesting is, again, this ability to communicate. Many of our problems as humans is the ability for us to communicate with other humans. Communication is clunky at best. Language is most people don't have a mastery of it, and even the people who do have the mastery of it, it's limited. Waving your hands is limited, drawing things on a whiteboard, putting things in a PowerPoint, all of these things are limited. But imagine if you could communicate the way sort of James Cameron can communicate, but without the budgets, and like anything you wanted to communicate, you could say, Hey James, can you make me amazing, you know, communication package that would allow me to communicate? Yeah, you can do that. And it's even predictive. You're just like, I have an idea. And then your agent already built it by the time you got home because you had the idea where you were walking to lunch and came back and was like, here's five different ways you could communicate it with visual effects that would have cost many billions of dollars to do before in this year. So, like, this ability to generate content, to be able to communicate ideas, I think, is amazing and profound, that changes countless things. By the way, I think solves a lot of problems that we have of miscommunication between nations and peoples and all of that, or even with our loved ones. So, I think that's a huge one. Obviously automation of all things that today take up a lot of our time, that quite frankly waste our time. All this knowledge work we've been talking about, I think most people would admit that the work that they do, even though it's called knowledge work, and we call it intelligent work, actually isn't very intelligent, and it's kind of like repetitive and drabs. If you take that off, people get freed up to do things. Now, those things they do might be positive, or maybe they become hooligans and they create more chaos. Sticking on the positive side, like the things that they could do. Again, this ability for all information technology is code and data, and the ability for AI to write code without human intervention. Or people like, oh, you'll be able to prompt AI to write anything you want. Yes, you don't have to prompt AI to write anything you want, you silly human. Yes. And so for AI to just be able to build, refactor everything, and build whatever. Now again, if we stay positive, then there should be amazing new capabilities for us to be able to have fun, do things, learnings. Obviously, all of the, you know, I'm very bullish on AI's ability in in, you know, scientific research, drug discovery, I think the next four years, all of that stuff should be, like, massive steroids, humanoids, or other types of robotics. I mean, it's incredible. It's wild. I mean, I think by the way, think anybody who tries to predict, like, what happens is nuts. You know, that's why I point to Ray Kurzweil. It's like, we can't comprehend, and hopefully we don't blow ourselves up. Let's hope. Yeah. All right, well love the chat, Dmitry, Chad, thank you guys for being here, See you next time. Have a while. Thanks.