How AI Lets Small Teams Build Trillion Dollar Ideas | Eshwar Prasad | Part 4

This episode isn't about AI hype. It's about what happens when the cost of building a breakthrough collapses - and who actually moves first.

In Part 4 of this conversation, Eshwar - a solo founder building AI products - makes the case for why the next major medical breakthrough won't come from a pharma giant. It'll come from someone with a radical idea, access to AI tools, and no institutional bias slowing them down. He calls it the DeepSeek moment in medicine.

We also get into how he decides what to pursue, what his hiring philosophy looks like in 2025 (two words: fear and doubt), and what "balance" actually means when you're spending 18 hours a day doing work you love.

This isn't a conversation about what's coming. It's a conversation about what's already structurally possible - and what kind of operator is positioned to act on it.

In this episode: 

• Why Large Institutions Aren't Aligned To Lead The Next Breakthrough
• What A Deepseek Moment In Medicine Would Actually Look Like
• AGI, Personalized Medicine, And The Timeline
• How Eshwar Evaluates Opportunities And Says Yes
• His Two-Question Hiring Filter
• The Difference Between Firing Someone Who Isn't Working And Letting Go Of Someone Who Is
• What Work-Life Balance Looks Like When The Work Is The Life

Chief Milestones is a series for operators, founders, and decision-makers navigating real constraints - in business, health, and the life built around both.

Chief Milestones is a video podcast featuring honest conversations with founders, parents, and investors about building real businesses, staying healthy, and raising families.

New episodes release Tuesdays and Fridays.

Youtube: https://www.youtube.com/@ChiefMilestones

Instagram: https://www.instagram.com/chiefmilestones/

Spotify: https://open.spotify.com/show/7HLDAe7SP3GnJSTOgAQqLy?si=6e9ac82be46f413c

Apple Podcasts: https://podcasts.apple.com/us/podcast/chief-milestones/id1861185226

Transcript

Basement Breakthroughs And AI Optimism

Eshwar 0:00

Think the next breakthrough is gonna be done by two dudes sitting in their underwears in some basement and they've got a breakthrough. But the beauty of AI right now is that somebody that has a radical thought that I'm not like a deep seek moment. Nobody thought deep seek with the with so much less intent in intensive um you know GPUs that they're able to bring a model into the market as they were, right? So you'd have a deep seek moment in medicine, you'd have a deep seek moment in every aspect of these things, and you would be surprised. And I have no doubt that next trillion dollar company is going to be a one-man guy, one person company.

Reshma Vadlamudi 0:42

So anything else that's concerning about AI?

Eshwar 0:45

I know we just touched upon the security part of it, but um yeah, I mean, you know, you can have like you know, videos, a fake videos, a deep fake videos, or or whatever, but I think you can you can have different AI agents to be able to track those and then flag those. So I think most of it is overblown. I think you would have uh AI um kind of uh secure AI itself, right? So all you need is some other somebody else to write an agent that would actually say, like, okay, this is a fake version, it's an you know automated version, it's a real version, right? Yeah, so so I mean, uh definitely there are such certain issues there. Um, but I think AI, I think is definitely going to help. Um the net positive, the net effect is gonna be positive than negative. I think uh and I think it's gonna be way more positive than way negative. And I think everything in life, when the internet came, you know, people that uh the security was a question back then too.

Reshma Vadlamudi 1:40

Okay.

AGI Explained Without The Hype

Eshwar 1:41

Uh I think that is not gonna go away. You know, whether it's AI or whether it's GI or AGI, which is another aspect of it, which we've not discussed. Um, it's not gonna go away. And I think those concerns are gonna remain. And uh, you know, my data is my data, and I don't care if somebody looks at my data. You know, like for me it doesn't matter. But some people they don't want privacy and security.

Reshma Vadlamudi 2:05

Um, what is GI and AGI?

Eshwar 2:09

AGI is uh artificial uh general intelligence. Uh that's that's what um you know that is like super intelligence that is able to take decisions, that is able to in like all this work is being done to achieve AGI. So that is like a God level power, right? So that's you're creating a God essentially.

Reshma Vadlamudi 2:27

Okay.

Eshwar 2:28

So you take all of this and then you can become an age, you can make somebody that can be like a human. But you have a characteristic of a human, then that's AGI. So that's artificial general intelligence that you can actually, you know. That's if it if it has a consciousness as well, if it has, you know, a brain like you know, it can think like that, then that's it like does it have a shape and form?

Reshma Vadlamudi 2:54

Like does it look like a robot or something?

Eshwar 2:57

You can actually give it a shape if you want to, but essentially what it is, okay. Like if you ask a question, it's um you're amazed how it's able to give you an answer, right? So that's that's one part of it. Artificial general intelligence is uh is uh is is uh is probably ten times more than that. Like if somebody if it's like a human-level intelligence, right? Um so where uh you know you can give it a robot shape and then that can become like a person that can think and uh and and do things, you know, like that's that that's the thing. Like, you know, we've seen a lot of these robot movies, right? Yeah, so yeah, so it's just essentially the same, right? People that you know, you know, just you know, just thinking themselves and killing people, you know, those are the things, right? That you would they would come up with, right? So uh, you know, like if you does it have feelings, yeah, can it have ways for it to actually be insecure, you know, like you know, all those aspects of it might come through, right? So the general intelligence, yeah, um the super intelligent models, right? So then things, how things will be then, but I think it's very long way away, right? Even there is anything, I don't think you should my the one thing that I say is your life is gonna be so much better, right?

Personalized Medicine And Longer Lives

Reshma Vadlamudi 4:16

With it than without it. Yes, yeah.

Eshwar 4:18

Because if it's able to be super intelligent, you know, imagine you have a drug for any disease that is out there. Um again, you know, all these drugs are generic drugs, right? So they get tested and across a various amount of people that that drug might not be effective for black people versus white people, Asian people, whatever, right?

Reshma Vadlamudi 4:37

Yes.

Eshwar 4:37

So now if you have an ability for you to have like an AGI that is able to just decide like what composition works for you based on your blood type, skin type, you know, whatever, gene type, a protein type, and then it is able to automatically create a medicine for you, which is called personalized medicine, right? So that is food, that is dosed and effective for you. And you take an advil that is different, somebody else takes an advil that's different, that's tailored made to them, right? So imagine having um, you know, AI that is able to be so hyper-focused in in being able to prescribe something that is so uh you know inherent to you and so personalized for you, right? If that's a future, then you know, and that's just I'll be I'll be happy to be living in that future than not. Yes. So like I've been reading somewhere, like, you know, please, please, please do what you want to do, just live for the next 10 years. Right? Because because the amount of technology that is being innovated, after 10 years, you don't have to worry about just dying of a cancer or uh or a heart attack, right? I'm sure somebody might might shoot you and you might probably get in an accident, but a physical ailment killing you is gonna be less, right? So that's I've I've been seeing a lot of these things that online that you know, just come what may, just don't kill yourself for the next 10 years, you know. You know, do your exercise or whatever, do whatever, but just be stay safe, stay healthy. And and I think I think GLP ones, like the the ozumpic and one that those are like medical drugs that that seem to really do well um in getting the the blood blood sugar levels in control, and uh, you know, and that that's a whole aspect of it as like a medical drug um that uh that's gonna make the level playing field. Uh, you know, everybody can be healthy now, right? Yeah. So it has certain and certain inherent aspects of it. I think you're a doctor.

Why Big Companies Miss Breakthroughs

Reshma Vadlamudi 6:34

Yes. Yes. So with with that, I think in the next 10 years, just because how these guidelines in medicine in general works, it it might take more time than not. Like other aspects I totally understand, it might be easy to shift, and then the way the medical system works, I think it's just going to take a little bit longer than that.

Eshwar 6:55

And even though most of the I would say I think the next breakthrough is gonna be uh is gonna be done by two dudes sitting in their underwear in some basement and they've got a breakthrough. So so these big behumoth companies are not gonna innovate the way they want to innovate, right? Um because they have these these big motes and they have to protect their modes, they have to protect their IP, they don't want to you know innovate faster, sooner, you know, they have their their their their agendas. Um but but I think I think the next big innovation, the cancer drug or whatever it's gonna be, is gonna be done by two dudes in a yes, two dudes uh in a basement. Yes. So so so to your point, when you say like, okay, that you don't think that's gonna happen, yes, the big players do not have the incentives aligned for them to be able to do it. But the beauty of AI right now is that somebody that has a radical thought that are not like a deep seek moment. Nobody thought deep seek with the with so much less in intensive um you know GPU inference that they're able to bring a model into the market as they would, right? So you'd have a deep seek moment in medicine, you'd have a deep seek moment in every aspect of these things, and you would be surprised. And I have no doubt that next trillion dollar company is going to be a one one-man guy, one person company.

Reshma Vadlamudi 8:19

Yes. And that's where I was going. Like the guidelines might not be passed, and all this might not be happening, but there are certain doctors that I personally follow. Um, so what do they do is they they practice very differently. So they don't they don't follow these traditional guidelines. So they do their own retreats. Uh they just can't do it within US. Uh, if they were to do it, they might lose their license.

Eshwar 8:47

Uh but I mean all these are hegemonies that people want to protect their IP and they don't want you to just uh go wayward and and do something, and you know, I think so. So but I think with AI, I think all that everything is gonna get because you needed so much you know money, and uh, you know, for you to be able to do something like you know, creating new drugs used to take a lot of money, but not anymore, right? Uh so especially with quantum computing, that's another aspect of the secular tailwind that is that that we're gonna come come to. So quantum computing is gonna be also um something that's gonna be you know, you could just run calculations of, you know, you could just take um you know various combinations, right? Of all the everything you just you just dump it into um into this this model and then you just let it let it cook, right? And then it might come up with like two, three, uh, two, three scenarios with um, you know, with all the all the aspect of it fully figured out and then give you an answer, right? Um because there's so much data and then the computational power that that it has, um, you know, you know, that's that's something that I would think of, right? Um so now you could actually um rent a quantum computer and run your simulations for for less than two, three hundred dollars per hour, right? As of today, right? I mean, I think I don't know about that, but um, but I think you can do it, right? So so there's companies like RGTI and um QBTS, it's D Wave Quantum. So they they have uh a quantum computer on the cloud that you can rent and you can just you know spin up a specific load and you create a complex, unsolvable problem and just give it and then see what happens, right? So now things are so available for everybody to do it. I can do it, you can do it if you want to have to solve it. So so I think you have to take a completely different approach. People will think like, why did we not think about it? So easy. Yeah, sometimes things like that are so apparent and it's so easy to do, but people don't do it because they have a bias of being it being extremely difficult. So often things that that work are this are the simplest things, we'll have the simplest idea, and then you go back to the basics, and uh, people that do not have any experience or any knowledge or whatever just comes about. Everything that we come come come to know about is always gonna be some experiment or somebody just thought about something. Without a thought, then nothing would have come come about, right? Yes, so it's as it's as simple as that. So I have a lot of um optimism about AI and how it's gonna help us all, you know, have live better lives, be help, you know, live uh uh live healthier lives, right? I'm I'm sure that's gonna contribute to it. So I'm hopeful um of a good AI than having negative negative um effects of AI.

Using AI Agents To Ship Fast

Reshma Vadlamudi 11:48

Yeah. Yeah. So how are you leveraging AI across your uh let's say other businesses? Like are you?

Eshwar 11:55

So I'm in the process of completely enveloping myself to build something within AI right now. So I'm I'm I'm not looking forward to, I'm not looking up to anything else at this point. So uh, you know, I think I think uh, you know, I just uh want to ship certain things that I call that I built myself because I always wanted to build stuff, but I didn't know how to, and I had to depend on developers to be able to do it. And the experience that I've had is is horrible, you know, especially the freelancer market of being able to develop products, you usually end up, you know, everybody says that they know everything, they don't know anything. So um even you have an idea and you it's fresh in your mind, you want to do it, so you end up doing it then and there itself, then you have the you give it enough wings for it to fly. But if you have or conceptualize an idea and you have an app idea, for example, and you don't know how to build it, and then you think of hiring somebody from Fiverr or Upwork or somewhere else, and you give it to them, by the time they come back with specifications, it's two months. Or even the product is shoddy and it's not it's not the way it is, and then your enthusiasm about a specific idea would die off. So now you have an opportunity for you to take if take everything in your own hands, conceptualize an idea, and then be able to try it and and realize it and then ship a product and then put it out in the market and see if it works or not. You can do that in less than less than two weeks right now. Yeah, so that could probably be less than one day, right? So once your AI agents start building it for you, yeah, and they figure out everything and talk to other agents and then be able to figure out, and you can say, like, hey, I'm I'm in a mood of um you know creating my own Instagram.

Reshma Vadlamudi 13:42

Right?

Eshwar 13:42

Yeah, uh, but still connect the data to Instagram. But I want the Instagram to be not with this photo, should not be like this. I want it to be bubbles or circles or whatever, right? You could just you could just have an API to Instagram, and then you could build an Instagram that is custom made for you. The way it looks and feels is personalized for yourself. Yeah, so you could, you know, you could think of the app store becoming like that. So right now apps are stationary, everybody, everything, everybody has the same um kind of apps, you know, same apps and the same functionality and same things. Now you can customize your apps to where you want it, want them to be, right? Yeah, or you could probably say, like, I want like this app, I like something from this app, I like I want to combine all of three and create my own app. No, things like that.

Choosing Ventures And Saying Yes

Reshma Vadlamudi 14:29

Yeah, yeah. And then how do you choose what ventures to pursue? And so how do you say no to most of them? Like when you land on idea, or you maybe people might approach you, or everything might be like exciting at first, but then how do you know, okay, this is what I want to take up, and now I don't want to take this one up?

Eshwar 14:52

So inherently I say yes to everything. Um, because I don't go out up as much, and you know, you don't get I don't get a lot of opportunities like I want to do this, let's do this. But if I do, if somebody might find me and come to me and say, like, okay, I have an idea, I want to do this, chances of me trying to do something with it are always yes, because I'm not getting too many things that I can take, and I get only one or two a year, uh, or maybe like one or two in three years in this in this case. Um, and I just do it, you know. You know, as long as it's not too much money that I'm risking, if somebody asked me for 20, 10,000, 20,000 to start a startup, I would just say why not? You know, yeah. So in the process, I uh I learn about that person, understand what it is, and you don't you never know what's what's what's where. So I don't really um be too calculative about it if I feel like I want to do it, and then it's just some something wipes, you know. So I sometimes have to like the person, I don't care what the idea was, right? And I probably might say yes, just to just to see what's out there. Um so I don't get too many um people asking me to do stuff, so that's why um, you know, maybe it's a different story if I'm getting like 10, 15 people wanting to do some different things, I have to say no, but that never happened to me. Um, but yeah, I mean, um, but if if you don't like, if you don't see it, and if it doesn't excite you when you're on your first hearing, like if it doesn't, then you say, sorry, I can't do it. Yeah, sometimes you're constrained by money, like sometimes your market is you're completely invested, fully invested, and you don't have any money, you obviously have to say no. Um sometimes uh you go yes, you say yes, even if though you don't have money, you start trying to figure out to make it work. Um so for me, I just say yes to most of most of everything that comes my way because I feel like I don't go out. If somebody all of a sudden finds me and says like something, then I would say yes, and I just put some money and see what what it works, right? So if I put if I have to put 50,000 to just find out the idea, I'll probably do that.

Hiring For Mindset Over Experience

Reshma Vadlamudi 17:03

Yeah. Do you follow like any principles when it comes to hiring?

Eshwar 17:11

Um yeah, I was really bad at hiring. Um, I feel like again, I I have the same feeling, like you know, like um so that's one thing that I could stuff totally improve on if I look back uh uh on in terms of hiring. But I think um my perspective about hiring changed uh with all the spirituality uh dosage that I've been getting for the past uh 18 months is is um and again that kind of also factors into the AI and how things will be in the future, that even though you you probably lack some some experience specific thing, that's not a barrier anymore because as long as the person has the willingness uh and and hunger to be able to learn, you can learn. Anyone can learn anything now. So the way I look at it in 2025 in terms of hiring is uh the person should not have no fear and should not be should not have any doubt, doubts, right? So structuring your questions in a way to find these two two fundamental things, fear and doubt, to gauge that person whether he's fearful of trying new things and fear of failure, or he's he's doubtful, he's too he questions everything and he's not too doubtful to be able to do anything, right? Um or uh that's why you see like most of these big companies that are big companies when one were small are are are uh headed by a couple of people, right? Two people. Um uh and then you know most often they don't have experience, and even if they have experience, it's just how getting to make things work, even though the odds are stacked against you. So the moment if somebody comes and says, like, oh, I can't do that because these these experiences, then it's it's it, then then they they are basically a uh a prisoner of the past. And they're willing, they're they're willing, they're willing not to try anything new. So you can't succeed in that in with that with that kind of mindset. Maybe in a big environment where they're they're they're um some big companies where their aspect of it, their work is not so critical in nature, the decisions are not so critical. Maybe they can survive there, but at least in a startup, you need somebody that just wants to be able to pick up something, run with it, not worried about failure, not scared of anything, and they're not doubtful. They often take, they just pick an idea and then they take a decision and they stick to it. And then um, and then if they fail, they'll go back and look at it and say, like, okay, I didn't work out, so I'm I learned this, I'm gonna move on.

Reshma Vadlamudi 20:00

So you're going to look at m more about the mindset than their experience in 2025. Yes.

Eshwar 20:06

You know, in 2020, pre-2019, yeah, he needs to have right experience. He needs to know have coding, or you know, he needs to already have like different experiences based in the past, so that I'm not teaching him something new because any part of hiring is training them.

Reshma Vadlamudi 20:21

Yes.

Eshwar 20:22

Um, so with with with AI, that has become so short, right? So it used what what it should take for for somebody to learn an year. Now it's probably can be done in a month or even weeks or maybe days.

Reshma Vadlamudi 20:34

So especially if their drive is too good, they will learn it. Like they will learn in 12 to 18 hours a day. They will work towards it. And uh yeah, that's what that's what you want.

Eshwar 20:45

Like right now, I'm in a founder mode, right? I'm spending 18 hours a day to to to learn about these things, right? So I have no sleep, no thing, nothing. Right. So I can't, I like you, I love it and enjoy it so much that I don't feel like going up to P.

Reshma Vadlamudi 21:00

Yes. Sometimes, right? Yes.

Eshwar 21:02

So so so that's the frame of mindset that is in. I think you need to find people that want to be having the same mindset as me. And if you can um surround yourself people like that, uh I think I think magic can happen. Yeah, I think that's what I'm looking at uh at this time. Again, I go back to those two things, fear and doubt. If you don't have fear and doubt, you you you can do anything you want. You can you can build a rocket, yeah, you can go to Mars.

Reshma Vadlamudi 21:31

Yeah, right.

Eshwar 21:32

It's just it's possible. Anything and everything is possible, it's just that fear and doubt that makes you not want to get it, right? So you can take anything in voyage, everything is free. Um that's the only two things that I feel like I'm putting too much emphasis on those two things. Yeah, somebody is fearful and doubtful, I don't want to hide them. And you have to buy and find a way different creative ways to ask questions, uh, to gauge them. Even though you ask all these questions, you know, like you know, you would be good at hiding stuff. Yeah, you would and they would probably not be able to hire hide it once they start working for you, right? Once you feel like that's not who what you signed up for, then the choice becomes very easy for you to just say, like, you know, go to onto the next one. Yes.

Firing Well When It Is Time

Reshma Vadlamudi 22:19

Yes. Um are you good at firing?

Eshwar 22:26

Uh I think so. I think so. I think I've uh spent, you know, like in you know, it's also important for us to explain that that the firing, I mean there's some bad apples that you have to fire. That's that's no problem for me. But there's certain times where you have to let go of somebody just because it's not working out, maybe the work is not there anymore for that person and they're not adding as much value as we hoped as we thought it would, and maybe we're pivoting away for something else. The the person is really good, but you cannot repurpose him somewhere else. And uh in the per for for the uh the goodwill of that person for them to find the right job that probably they will love to work with. So you kind of see it in a way that it's probably better for me to let them go because they're not value adding any value here and they're not learning anything here. So it's probably best for that person to be elsewhere. Uh, I think that's where I have a little challenge. Uh so I, you know, try to make it work, make it work as much as possible. Um, but I've in hindsight, you you know, the moment you felt like it's not working out, it's probably better off letting that, letting, letting the people go so that they can do better better things with their life than just sticking here and being stagnant, right?

Reshma Vadlamudi 23:46

Yeah.

Eshwar 23:47

So I have difficulty in fighting that type of situations where I feel like the person is good, I like that person. I could not find anybody that is trustworthy or as good as that guy, but he's not being useful for me in this specific setup. And it's hard for me to let people go like that.

Reshma Vadlamudi 24:04

Yeah.

Eshwar 24:04

But if somebody's not working or it's not working out, then it's just it's pretty easy for me. But um, but again, you know, it's twofold.

Work Life Balance And Boundaries

Reshma Vadlamudi 24:12

Yeah, yeah. Okay. And do you believe in balance? Like when I this is regarding the family, and let's let me ask you this one. So do you think a work-life balance is something uh, or let me let me reframe again. So, do you think work-life balance is a possibility and uh for someone who wants to be successful?

Eshwar 24:39

I guess I guess we could kind of draw these boundaries ourselves, what is to be balanced. You know, that could be different for different people, right?

Reshma Vadlamudi 24:48

Um so what is the work-life balance for you?

Eshwar 24:54

You know, that's I don't think about it, think about it like that, right? I I just do whatever I feel like doing in that day and that moment and day. And I'm just happy doing that.