Ep 04 | GenAI Series | Sully.ai | Ahmed Omar | Co-Founder/CEO

Efficiency Unlocked is hosted by two tech-savvy physicians and dear friends of mine, Drs. Adam Karoo and Dale Gould, who are experts in using technology to automate routine tasks and increase efficiency in everyday and business life. Let's see what they have for us today.

Stay tuned for this series of episodes where we'll be discussing the game-changing impact of AI in the realm of patient care. Patient care and physician well-being. Get ready to unlock efficiency with NerdMDs.

Efficiency unlocked.

And at this point, we just started talking to doctors, we started talking to nurses, surgeons, and started seeing like they really have a need to this. And then we started looking into the problem. We started seeing it's beyond just transcribing, it's beyond documentation, it's beyond administrative work. It could actually help physicians and doctors diagnose better. It, you know, so it was, it was beyond just— it could actually help patients.

That was Ahmed Omar, a 3-time gold medal champion MMA fighter. He's also the co-founder and CEO of Sully.ai. At a young age of 13, Omar saw how online commerce solved a vast supply and demand problem, making him eager to learn more about it. Being an MMA fighter gave him the tools to understand where determination and discipline can take him. He found his passion when he built an e-commerce brand in New York City, taking it from zero to $55 million in about 3.5 years. Omar and I had a great conversation that we're about to share about how he and his co-founders have taken Sully.ai to the next level and where they want to take it in the future. We hope you enjoy this show. Why don't you start with, you know, kind of sharing your journey? I mean, I introduced you a bit, like a very condensed version of you, um, but share your journey and the inspiration behind, like, why you wanted to help create an AI healthcare product or solution.

Um, why don't we start there?

Yeah, I love that. So Actually, after building the e-commerce company, I decided to build an actual technology company, you know, because e-commerce is more on the growth side. It's more physical products. It's more, you know, SEO, growth engines, you know, CAC, like it's more growth than tech, actual tech technology. Sure. And at this point I wanted to build a technology company. I didn't know what I wanted to build, but I was in New York City, went back to Egypt. I'm Egyptian, but born and raised in Dubai. I went back to Egypt to visit family and one of my friends, I'm kind of attracted to solving some problems. And one of my friends was like, oh, I can't find any part like for my car, any parts, any car parts. And I'm like, why? Like Google it. He's like, there's no way. Like you can try. And I Googled it and I tried like one everywhere to look for it. Like I went basically to Google, I spent 2 hours looking for a phone number, found the phone number, found an address, went to the address and then they couldn't pick up the phone and like, you know, like went for like 2 hours and then I left home, I went down to like a market, found the parts and then delivered it to my friend and I made maybe $20 at this point. And for me, he was like, if you can like scale this, like if you could do this to everyone in Egypt, that's gonna be like, that's gonna make millions. And 2.5 years later we did. It was, and I started building this marketplace that is an automotive marketplace that could get supply to their parts and then people, car owners to buy their parts and easily find it with just a simple search. And yeah, we did that. And yeah, 2.5 years later, there's 1.8 million people in Egypt who went on the platform. There was 100,000 users we generated over $14 million in annual run rate. We scaled to 2 other countries than Egypt, Dubai and Saudi. So this experience was really a technology experience. It was more like backend, frontend, supplier dashboards, apps, even Arabic and English. Like we had to duplicate everything for Arabic and English. And it was like the first technology experience that people could actually go on a platform or on an app, download an app and buy something. And it was And for me it was, and that's basically how we met, me and Nasser. Like Nasser, I got introduced to Nasser from one of like my co-founders from one of the suppliers. Supplier told me, oh yeah, I interviewed a crazy guy like 2 years ago and he's like one of them, like he's a genius. I think he told us how to run our business, but we closed the position. And then I met Nasser and we just clicked. And together we built like this marketplace that grew beyond what we expected. At some point we, like while building the company and growing it fast, we knew this is not gonna be the last company because we knew it's not a big enough market. It's just a problem that I started to solve and it just scaled very quickly. And at this point when we like took like some US VC money, you know, YC and all these things, they made us realize that when you're spending your time, the most valuable time is like your time. And if you're gonna spend building something that you know is not gonna scale beyond $50, $60, $100 million 10 years from now, it doesn't really make sense that you need to build it. And at this point we started basically to question what do we really wanna do? What do we really wanna spend our time on? And this was like a life-changing question, I would say. Like what would we spend the next 10 years on that we don't feel like work? Masa at this point had chronic illness. I grew up, my mom like misdiagnosed with polio, so they didn't give her feed, so she never walked. So it felt like the past 10 years of building companies, technology companies had come to like this point that we can build something that could change people, that can impact people's lives. And we didn't know what we're going to build. Like, don't, I'm not going to tell you I've been working on this problem for the past 10 years or something. I'm like, we just came to the Bay Area, YC helped us get here. And we started basically building a few things. We started talking to a lot of people. We had a few encounters and also had a few encounters with the healthcare system. And then we just, and then he's like, Omar, I think you need to go to the doctor and like see how the healthcare system is designed. And we went to Kaiser, you know, we went to biggest organizations, we went to smaller practices. And it was really weird. Like, it was like, wait, was this the only industry without any innovation in the past, like, 30 years? It was like the day for healthcare or something. Like, every founder just say, oh, you know what? I'm not going to do healthcare. And then we were a bit naive because we were new in the Bay Area 2 years ago. We didn't know a lot of things. And for me, the only way to know is to, like, choose the lowest level. Position you could choose. And it was like front desk assistant. I worked as a front desk assistant behind like one desk in one of the clinics in Senegal.

Yeah.

And that was like, this is gonna make, help us decide either we want to build in this problem or not. And the initial idea was basically to build a doctor basically, or an AGI doctor, a version of the doctor that could prescribe something for you until you see a doctor. And looking at this, it didn't make sense that we need to get this to life when the doctors, half the doctors' times are spent in talking documentation. Like most of the doctors' time are just gone because of documentation. So did it make sense to build another version while we can double or triple the doctors in the world by removing these stupid administrative tasks? And yeah, This was like when we were, I worked behind the desk, we started with a billing like, you know, product and then we thought that billing is never going to be solved if there's a misinformation from all over, you know, from the time the patient comes in until he leaves. So we started building like AI agents to help, you know, intake, check-in and scribe and then also communicate with patients. And because they all have and run on the same language model, they have 100% context about this patient and about the doctor and about what the doctor said to the patient and about what the patient answered in the new questions, you know, all these things. And yeah, at this point we said like, we wanna, we can disrupt healthcare. We disrupted the automotive industry knowing nothing about automotive. You can't do that. And we had half the team that we have, like the team in UAE, like the leadership team, wasn't even half as good as the team we have today. So yeah, we decided to take this opportunity. It all resonated. Chaitanya, our CTO and co-founder, had 2 years of chronic illness dealing with the US healthcare system and has a really interesting story how we met. And yeah, Sully Chen built an AI medical assistant. He couldn't ship it to doctors. He couldn't deliver it to real users. And when he saw us, then we gave it to real doctors and users. He's like, oh shit, you guys were able to give it and get doctors to use it. Yeah, he's like, yeah, I could do everything without that. And yeah, Sunny is one of the GPT-4 contributors. So I think the team just came together, aligned on the vision, and we just basically started working on Sunny.ai. And at this point, everything just made sense. Everything just came in place. The clinic that I was working behind the desk, we acquired it as a customer. And then, uh, and then, yeah, and then very cool, etc. I would say, yeah, yeah.

So when— so what is like, you know, I was gonna ask you your why, but like, clearly you just answered the why, so we don't need to go into that. But I guess just for context, so when did, when did this like first iteration of Sully.ai kind of start? What was the timeline roughly, like when you were, you know, when you like tried to explore that billing product and what's the timeline from that until like today?

Yeah, so I think the first encounter we had with healthcare was July, June, June 2022. And then we started building something for the doctors, a transcribe tool that can help them transcribe. And then we started basically say turning it into medical notes. And then we started showing it to them, like before even building anything, we just like, you know, did a simple, you know, like a 6M Beauty product, like just kind of, yeah. And actually we had hardware used into that. So we were using smart glasses hardware. So it was like, it was like a lot, you know, it was a lot of changes and a lot of components to it. And at this point we just started talking to doctors, we started talking to like to nurses, surgeons, and started seeing like they really have a need to this. And then we started looking into the problem. We started seeing it's beyond just transcribing, it's beyond implementation, it's beyond administrative work. It could actually help physicians and doctors diagnose better. It, you know, so it was beyond just, it could actually help patients down the line basically be proactive about their healthcare, you know? And at this point we just decided that would this problem, like one human, one doctor, the vision of the company is to bring a doctor to every human on the planet. One human, one doctor. Would that be something we not feel like we're doing it for the next 10, 20 years? And I was like, yes, it would. And then September, basically I took a job behind the desk in one of the clinics. And then we actually started building the product in February because we were leaving a lot of technology. I think that's a mistake a lot of people do. You know, they bring the technology up front. And then they try to sell it. But we wanted to lead with the problem first, understand deeply the problem. Where does every dollar go from like the insurance, like from your personal, patient personal insurance, you pay for insurance and then you go to the doctor, you pay a copay and then you have to pay a bill if you pass the deductible. So we tracked every single thing, the revenue cycle, the the dollar amount. And yeah, we just wanted to understand it deeply before building something. So when we build, we have a lot of context about what is needed, what is desired, and what is really not there, you know, what is really used or nice to have. And yeah, we did that in Feb. We had Chaitanya join us as CTO. He's just telling you like PhD in robotics, built Google Assistant, like he built like massive technology at Google, 25 patents. So he just resonated with what the problem and we all aligned on the vision, which is really hard sometimes. A lot of the time, like the team can align on what you do today, but they're not aligning on the vision. But the first thing we make sure is everyone is aligned on the vision. We know what this product will look like 10 years from now. So, so yeah, I think the timeline I would say we started talking to doctors was in June 2022.

Got it. No, that's good. That's good context. And yeah, no, I mean, everything you said there just resonates. I mean, the thing I was thinking, a couple of thoughts, you know, you were talking about like understanding the problem first, you know, and like that is like so foundational to, I mean, everything in healthcare, you know, but everything in, on frankly in the world. Could operate in that same way. And, you know, the thing I stand by is let's figure out like a problem to go after. Then, like what you mentioned, you don't jump right to the technological solution. I mean, even though I love technology, I'm always thinking, okay, like what could be the solutions to this? But it's like then you have to understand the people that are doing this problem and the process they're following, because all of that part can be kind of optimized or potentially changed to before you land on the actual kind of technological solution that augments it all. So no, that's, that's spot on. I'm sure we could talk here for hours on this topic alone. Um, let's, you know, I think you, you talked about this already a bit in what you just said, but kind of leading into the impact of, you know, where this technology is, and, and I want you to share a bit about what Sully.ai already is doing in this space. I mean, you gave me obviously a little bit early kind of preview demo of the product and I'll just say I was particularly impressed with some of the front-end solutions that you guys have, like the kind of pre-gathering of information from a patient. So I definitely want you to share that, but I just want you to share like kind of why you think this space is so important and that kind of halo impact that it's going to have on the healthcare industry?

I think it resonates. I think so when I look at it. So I'll tell you, I'll say, I'll answer the first part of your question of what Sully is.

Sure.

So Sully.AI is an all-in-one copilot that is built to remove all the administrative work off of doctors. Day to day. So what this means, what does an all-in-one mean? So we particularly designed Sully to an agent-like, to an AI agent-like experience. So, and this is the reason we did this is because I work behind the desk. So I saw that each and every person, administrative worker is like an agent that does that conflict. So they conflict each other, but they also do their work. Each other's work, but they all have context about a certain patient. So we built it. So what we did is an all-in-one. So an all-in-one AI agent for healthcare is simply we built an AI check-in, you know, AI nurse, triage nurse, where it can take all the interview questions, take all the medical information from you, and then basically regenerate or dynamically personalize the intake for you or the check-in for you. So if you said, I came in because I have flu, and let's say you, Adam, like I said I have flu and you said you have flu, it's a totally different set of questions. It's not the same because the medical, my medical history is different than yours. And that's how it should be, right? Because healthcare is really specific. And since transformers came to play, there wasn't really a way that you could personalize at scale. You could do it with templates, you know, all these things. And I think that's why the biggest industry that uses templates is the healthcare industry, is because you're trying to personalize, but there's actually an infinite amount of templates that you need to use. So the reason we thought, we think this resonates a lot. So yeah, like an AI triage nurse, I'm sorry, I'm gonna get back to this, the AI triage nurse, And then once the patient walks into the room, the information that it's being discussed with the doctor is really important because it's not only that it can help doctors document, which is key, saves them 2 hours a day, 1 to 2 hours a day. They love it. Doctors, like yesterday we asked one of the doctors there at Health Partners, one of the big organizations in Minnesota, and we asked her, what would your life be without Sari? She said, it would be miserable. So because we removed a lot of documentation time from her end, but we also, it's not about that because the AI, like the human scribe inside the clinic doesn't only document, they also send pharmacy orders, they lab orders, they can call the customers at some point. So we want the AI scribe agent to do all this as well. Add the orders in your chart. To remind the patients as well, you know, 2 weeks because you and the doctor discussed it. And all these tasks was, didn't really make sense that it shouldn't exist. Like an AI automated, AI agent, an AI support agent. Why does a doctor need to reply to the insurance company? You know, there's data, there's CPT codes, there's ICD-10 codes. You can just generate responses and just make it ready for doctors to send, or you can just automatically send it if it has nothing to do with the medical or any health-related questions regarding the patient. So we started looking at these things and one of the things that we saw most companies do is that in healthcare specifically in the US is that they say, oh no, we cannot prescribe or we cannot recommend the diagnosis because we need FDA. And the way we started building is why do we really need FDA? If we could just generate it, show it to the doctor, if they want to use it, they can use it. If they don't, they can just edit it and just close the note. As long as we're not closing notes, we're not doing the final approvals, we don't really need FDA. At this point, we can start training the models and improving it based on the closed notes that the doctors are closing and the generated output by the model, and then run a dice for it to see the similarities. So we started looking at all these things that could train models, make them better. And also at the same time help boxers get like a lot of things done and build it in an agent-like model. Like doesn't mean that you want a Scribe AI agent that it's just gonna scribe for you. No, it could do other things. It could send an email for you. You could chat with it. You could ask it about the whole database of the questions, removing all the clicks from Epic, I would say. So it doesn't mean that it's a scribe, it will just scribe. And this way, this approach didn't really resonate with us. We didn't like that. We like, the doctor can still ask a human scribe what happened with this patient on Thursday, right? You can't do that now with any scribe on the market, but you can do that with Sonny now. So this is like the way we built it is like we built it with this approach, you know, if you know what I mean, like an agent-like approach.

Yeah. No, this is fascinating actually. This is even more, I'm getting more out of this than our initial like conversation in the demo, which I figured we would. And, you know, I just think it's so worth to, like, reflect on a couple of things you said. I mean, one, I think this, like, agent approach, you know, and being that kind of true kind of copilot to be able to do all those helping hand things that are a part of delivering care is fascinating. And I think it's going to resonate with a lot of listeners. I think the other thing you brought up, which, you know, I think is a hot topic right now, is this, this whole FDA approach to approving different algorithms that are in healthcare. But I think to your point, and you're right, the FDA really just wants to regulate digital devices that kind of make a determination that continues down the path. But like you're saying, it's really all about surfacing this information for the human to actually make that ultimate decision. You know, that's the thing that we're— even this technology creating a note for a physician or a nurse is so transformative. But I think that you guys are seeing that future potential of where this is going. And, you know, I like to call it the, like, the tsunami wave that is just starting to build. And that's like where we are right now. And it's like that wave is going to get huge and hopefully, you know, all of us will be able to ride it in a positive way.

Yeah.

But yeah, I mean, what you just got into was like, honestly, like the differentiator, which I do want to give, you know, everyone a voice to explain that. And do you have like, I mean, do you want to like fine-tune what that question and like specifically for the listeners, like where do you think Sully is like fundamentally different than, you know, any of the current kind of competitors or potentially even where they're going? That you know of?

Yeah, so one of the things that we, like, I saw when I worked behind the desk in Santa Clara at this clinic was that I witnessed how nurses, how, you know, MDs, doctors work on the software and use tools. That was one of the things that I really wanted to see, and I really wanted to observe and Doing that, I also saw people selling software. So I saw salespeople coming and selling software. One of the biggest, like one of the biggest problems that I had was that doctors needed to use an average of 4 to 6 tools to solve basically the same problem. So they would use IntakeQ for intake and then link it with their DrChrono, which has intake as well. And then link it with Spruce to send the message, which has a patient portal, but they still use Spruce. So the way we looked at it is that there's not really a company that's solving the whole problem. And it's the same problem at the end. Like what they're doing is the same problem. It's administrative problem. And the way we saw it is that everyone is just trying to sell something to the doctors. But the way we looked at it is what is the real problem and how can we solve it for good? And the way we did it is that think about Sully as a brain that is connected to every healthcare provider, healthcare system, like Waystar for insurance information, Availity for insurance and verification. It's linked with Twilio for SMSs. It's linked with Powbox for HIPAA-compliant email. It's linked with all these tools that are independent of your EMR. So for example, if you wanted to do a certain or automate a certain task, it won't really need your EMR's API. You can integrate it, but it will just say, okay, we can, like, you could just send it, it will just send it for you. So you can say today, you can say, hey Sonny, can you send an email? I'm not sure if Masa showed you the new demo. You can say, hey Sonny, can you send an email? And it will automatically send it by just the voice prompt.

So yeah, very cool.

So the way we solve the healthcare is like, All of them are trying to sell all these tools, but the way we really wanted to solve this, how can we really solve this problem for good? And the way to do it was to build a universal, is to take a universal approach to it. Is say, what is happening in all these baby steps and all these points and how can one model automate all this? And that's how we built it. And we saw like massive, like people are laughing on the calls. People are like, oh my God. People are like, oh yeah. You're like, holy shit, that was gonna take me 50 clicks on Epic, you know, all these things. So I think that's what you wanna see. That's what we wanted to see. We spent a year to try to get to this point, but once you get to this point, you know, like it's about a matter of time that you scale this. And I think that's what makes us different. The approach that we took about healthcare, the approach that we took to solve this problem was a little bit different than most companies. It wasn't about selling a Scribe selling you a scheduling tool. It was about solving the whole problem because I can sell you a scheduling tool and you could still have administrative problem and patients will still say, oh, you didn't follow up with me. You know, or patients could forget to book the appointment. So all these things just aggravated us. It made us more, like it made us more driven to solve this problem in healthcare.

Yeah. Now, and so, yeah, sorry to cut you off. No, that, No, I appreciate that.

And I think for the, like, you always had the subconscious fear when you go to a doctor that they don't really know the latest information about certain things. How are they treating you better? And this was one of the ultimate questions that we wanted to answer too. Like, there's no way doctors can digest 10,000 pages, pages every single hour of research, right? And that's impossible. But there is a way that models and transformers can do it. And if they can do it and doctors are using it, they can alert doctors. And this was the second goal that we want to build, basically the diagnosis part of it. And that's the point where we started saying, yeah, for us to do this, we have to have a maximum, like the maximum amount of context about the patient, patient's history, even from other EMRs, not just this or the doctor's EMR.

So, yeah, so just, just to clarify, you're talking about like, you're talking about the health information exchange aspect of this, like to really make this the best it can be, it has to be tapping in to all those other sources of health information, which frankly is also a huge problem to fix that has not been solved in the US or the world about.

Yes, that is true. Yeah, that, that is what I believe and I have believed about a certain vision for other companies and I was able, thankfully was able to bring that to life. We were able to bring it to life because it was a team. It was always a team. I look for the team before anything else. So we see the same vision and I believe we'll make it a reality. We have the capability, we've done it before. No reason for us not to do it again.

Amazing. Jack, I mean, we've already been talking for almost a half hour. Time flies, right?

Yeah.

These kind of conversations are so important because I think this is like, this is such a hot topic right now. And I think, you know, I think what a lot of health systems are struggling with right now is like, who do we like listen to? You know, who do we get a pitch from? Because, you know, a lot of times it just comes out as like a, you know, as a sales pitch. But like, I mean, if you had an opportunity, like you had 2 minutes to sit with you know, the CEO of like, say, one of the largest health systems in the United States, but you only had 2 minutes, like, what would be your pitch to that person to like get the next meeting? What would you say to get them convinced to take a look at Selly.ai?

I think I would say we're a team of 4 founders, a GPT-4 contributor, a 15-year Google lead, lead engineer who built Google Assistant and built most of the technology that we use today. And me and my co-founder, we built a previous company, disrupted a massive industry before. We're building an all-in-one AI agent to automate administrative work for the healthcare system and your doctors. Our goal is to remove all the admin burden on doctors, make sure they're not burned out, make sure we give them back time. So you as a healthcare organization can be more efficient and see more patients, which aligns with our vision. We are a small team today, which is good for a healthcare organization because we can customize and tailor each and every agent that we built specifically for your healthcare system. And we can prove that in not just weeks, but in days. We can prove that in numbers because we have many, many doctors working on the tool. And we have testimonials of them saying how many hours it saves them daily. And we integrate with every EMR, with every device, so we can make it super easy for your doctors to use. Whatever they're using today, they can still use, and we can automate all this on the backend. We can remove all the clicks your doctors go through with a fraction of a cost that you will pay to any other AI tool. Our margins are transparent. Our tech is transparent and something that we have that we know a lot of companies lack is that our product really, really works and your physician could test it. Yeah. It works consistently. It won't work one at a time. It works every single time. So I would say that would be the pitch.

All right. Well, if I was a CEO, I would definitely be enticed by that pitch. So good job, Omar. And hopefully some of the listeners are also itching to hear more. To that note, you know, what's the best way to get in touch with you? You know, we'll obviously include this in the post-production notes as well, but what's the best way to get in touch with you?

I don't know, do you prefer people like saying my email out loud? I can say it's Omar@solid.ai, so it's Omar@solid.ai, or people can reach me out on LinkedIn too. I'm very active on LinkedIn. LinkedIn. Yeah, I think like anywhere, Twitter, like LinkedIn, whatever people prefer to use. I'm all about user experience. So it's whatever platform they prefer to use, they can reach me on. And my email is my last name, Omar, and @sully.ai. So yeah. Great.

No, that's, I mean, well, you better start creating an AI to answer all your emails pretty quick too, because I do. I did the entire, uh, bullet list like last—

email.

Oh yeah. What were you gonna say?

Yeah, I, I, I'd love to read each and every email that I, I'm gonna receive. I, yeah.

No, amazing. Now you've been super responsive to, to any of like my inquiries, which I, I really appreciate. Um, so let's hit this like what we call the Nerd Alert Lightning Round. Yeah, it's gonna be a series of questions that are kind of familiar across the board for these, but, um, you know, this whole podcast series, but, you know, it's just a little bit more to get to know about you. Obviously you've shared a lot already, I think, kind of even personally about your why, but, you know, first question, if you could vacation or live anywhere in the world, where would that be and why?

I think vacation, I would say, I feel home in the Bay Area and I haven't got that feeling for a long time. I was born and raised in Dubai, but I would say if I would vacation, I would go to Egypt. See my family and my nieces. Yeah.

Awesome.

Yeah.

Yeah. Being away from family is tough. I, uh, yeah, I don't live in a different country, but I definitely don't live really close to the majority of my family. Not in Sonoma. Yeah. Yeah. Not in Sonoma County anymore. No. Um, awesome. What's your, what's your, this is another kind of, I guess, softball kind of question, but like, what is your favorite, um, vehicle that's out there, or potentially even like a dream car that, you know, could even be a historic kind of car, classic. Do you have, do you have something like that?

I think in your mind? I think when I was 7 years old or 8 years old, I drew the design of the Cybertruck, you know. I think every one of them.

That's so cool.

Yeah, I love the sharp edges of the Cybertruck, so I think I really wanna buy this car because I just, you know, like the design of it. But I would say it was always the G63, the Mercedes. I think it's just like I resonated with when I was a kid. It's a very famous car in Dubai too. They customize it, they gold plate it, they do crazy things with it.

Oh yeah, I've seen pictures of those.

Yeah, the plate numbers are like $9 million and the car is like $400,000. And it's like crazy. Yeah, they spend like 20 times more on the G63 than the actual cost of the car. So yeah, 33 resonates with me.

Awesome. Well, both of those resonate with me. I definitely have a pre-order interest. I would prep as well. Yeah, for kind of similar reasons. I just think it's so different, but like fundamentally cool at the same time. And I think it's going to be a badass vehicle too. Yeah.

So I actually had a nice car. I rode a Nissan Sunny car everywhere because I felt like I was paying for more things when I went out with a nice car. You know, I was just like, I don't really mind what kind of car I drive, but if I'm gonna buy a car in my garage, it's gonna be the GT3. But I don't think I'll drive it a lot too.

Yeah, I think it's a trophy car, right? Yeah.

Yeah.

No, makes sense. You know, I mean, I can see you're like a big thinker. I think you've already alluded to that multiple times on this episode, but you know, outside of like, you know, the, your current thing that you're fighting for, I mean, what other, what other big problem do you see in the world, in the United States, in other industry that you think would be worth to potentially solve? Not that you don't have plenty on your plate right now. And why do you think that problem is important to you, to society, or the world?

I think people need a new user experience to use a computing device. Me being— me saying that would be the smart glasses. That looks extremely normal, extremely thin. Components are not there yet. The market is not there yet. But that's a problem that I know. I hate phones, by the way. And if you know that about me, the phone could be anywhere in the office that I don't know where it is. I'm just always on my laptop or just, you know, on Zoom. I don't like the phone. I don't like the phone experience. I don't like you have to look down every time you look at the phone. I don't like the way people look when they hold the phone. I feel like there should be a new computing device and a new user experience to interact with digital things. And I think that's— I don't see it in any other form other than the smart glasses. Some people believe different than that, like Humane and all these companies.

Yeah.

I believe it's gonna, it has to be really close to the eye so you can see bubbles behind the screen, you know, just flowing and, you know, something just swimming there. And I think with Gen Z's attention span, the only way to get their attention is that.

No, that's helping them mentally as well. Yeah. Well, yeah, I mean, those types of devices will obviously be heavily AI driven, right? I mean, it's going to be. Surfacing that contextual information and analyzing and, you know, answering their questions. Yeah, no, I, I actually just got the new Meta Ray bands. Like, they just arrived yesterday, the new version. And, you know, first version I passed on because I was like, these are going to just be a kind of a clunky toy.

But I'll tell you, the—

I mean, yeah, yeah, I'll say the second version is— I'm fairly impressed. I mean, it's not at what you're talking about, but it's like I think it could get, you know, incrementally to that direction.

Yes.

So yeah, I'm looking forward to playing around with those too. I'll take a picture here at the end of our interview so we can share a smart glass photo with the sharing of the episode.

We all find ways to surface the market because we think the best way doctors can experience healthcare would be using those glasses. If we run multimodal language models behind the scene and we could assist on what they see, what they do, every single part of it. And at some point, computer is going to be much better, much faster. These models have some latency to it. So I think within time it's going to be like a life-changing experience. So you could build life-changing experiences with certain hardware. And yeah, I can't wait to even build a Suddy app on one of those.

Yeah. Very cool. What, you know, outside of all this stuff, what like today do you feel like, what today are you using? You know, outside of your own product or anything, but like what is something you're using today that you just think is like, man, this is like, this helps me be so much more efficient than I could be without it. Is there something that you would not want to like give up that you currently use? Can you share what that is? Do you have an idea?

I do. I do have a lot of tools that I use now that I can't live without, but I'm trying to think which one should I mention first. Something that I use today. So many things that we use today nowadays. I don't know. I don't want to say a software tool because it's going to be obvious. I'm trying to think of something that is— are you saying something personal or something?

Yeah, it could be something totally personal. Yeah. I mean, something to just kind of like think of like how it helps you to kind of automate or augment your life.

Yeah.

And it can be something totally like generalizable.

Maybe something that we're doing. It's not something that I'm using, but there's something that we do in the company every Saturday and it's really, really not— I'm not trying to talk about Sadi, but it's something that I feel it's really pushing us forward. So every Saturday we meet at 3 and we discuss research papers, whatever research paper we want to read. It could be something completely like, we were discussing the other day in the standup about like, you know, URT2, you know, the nerves between the two brains. If you cut this line, like line of nerves and have two separate brains in one brain, it will, like your right hand's gonna stop your left hand, you know what I mean? Or it's gonna see the language, like the language side wouldn't really understand what this is, but it can draw it. So it's fascinating to learn about a lot of things. So we do that every Saturday. We come up, read the research paper on the call and discuss it.

Yeah.

Also sometimes discuss what we could do in the company. If there's any use case.

Oh, super cool. That's actually very cool. I love that. I mean, that's like stretching— Ahmed Omar.

Yeah.

Obviously it's stretching your mind and your team's kind of like education and yeah, 'cause when you start reading some of those papers, it's probably, I mean, it gets very esoteric pretty quickly. That's super cool.

Yeah.

What, so just a little pivot. This will be the last question for me, but like what, you know, there's a lot of talk about, you know, any form of new technology being kind of like uncontainable, like just being difficult to kind of contain from a societal perspective. What, I mean, do you have, in this space, do you have a concern about that or, you know, maybe the possible kind of solution to that?

I think that's a very good question, by the way. We ask, like when we were early, really early in the model and like looking at diagnosis, we'd see like the model recommending alcohol intake for some patients. You know, it was really funny to see. I was like, the doctor told me to drink, I can drive. So it was really funny to watch, but I think the models really react to what we feed it, like what data we give it and what constraints we can give it. I think in healthcare, if we can reduce the number of people that dies from misdiagnosis every year or misprescription or misleading prescriptions, it's about 500,000 people. My goal is to cut this into half or even more than half. And then whatever constraint we can add to the model and make sure it doesn't hallucinate, we could do later. But Technically, as we speak, this is achievable, but no one's doing it. And this kind of annoys me. You have 46 people die every hour from misleading diagnosis, misleading prescription, or misdiagnosis. So I think we're missing a point here saying what this could harm as much as basically saying how this could help. And I think in healthcare specifically, it's really easy to constrain those models. They shouldn't be going beyond certain point. They just operate in healthcare. I think the general part of the aspect of it that you have a big language model that can do everything, this is a little bit hard to contain because it's not vertically focused. But I think in healthcare, we found ways to actually contain it without even playing with the language model. We can just prompt it and you can contain it with the prompting. So there's a lot of ways, like graph of thoughts, chain of thoughts, all these things was like amazing ways that you could basically, 'cause it's like a brain basically, you know, it learns and a brain growing up in an environment that is harmful might harm others. A brain growing up in an environment that is not harmful and there's a feedback loop of what is right and what is wrong. I think can help contain those models.

I love that analogy of like, I mean, that's kind of the nature versus nurture type of thing, but yeah, that's super cool. I mean, lastly, Omar, do you have any questions for me? I've like spent obviously this whole time asking you questions. Do you have a nerd alert question for me?

Yeah, I want to know why Denver.

Why Denver? Yeah. Yeah. So yeah, full disclosure, we did not prep for this question. Okay. So these are always like softballs for me too. But the— now, so Denver, my wife and I met in medical school and we, we, you know, she's from Minnesota. I was from Northern California. And we, we did what's called a sub-internship in our fourth year of medical school. Out here in Denver at the hospital that we ended up both doing our residency training at. But honestly, we did it, you know, we came out to Denver because, you know, we did those sub-Is over the winter and we wanted to like do a little skiing and snowboarding. And this was obviously well before having kids. And so we came out and we honestly just like fell in love with the place. You know, I think, so we, before we knew where we were going to go for residency, we actually, made a decision to get married in Colorado only because as a destination, it was kind of halfway between my wife's family and halfway between where the majority of my family live. And then we ended up matching here and we did what's called the couples match where you can kind of link your two applications together electronically in the system. And yeah, so that's how we actually landed here. And I'll be honest, you know, initially I missed the ocean because living on the coast, you know, it was something I enjoyed having. But I'll tell you, man, Colorado is a cool place.

It's—

it has everything, I would say, except the ocean. And as long as you can vacation at the ocean or, you know, have time for that, that's honestly why we decided to stay here and we enjoy it. So good question.

Yeah, actually, it's really interesting. And I really, really liked when you, when we reached out and basically talked about NerdsMD and what you're trying to do. And yeah, the fact that you have this tech aspect of healthcare is really fascinating to see. And the way you're trying to deliver it with nerds is really even really, really cool. I think we should talk beyond this health spot. Yeah, thank you.

Happy to continue the conversation. Conversation.

Yeah.

Well, thank you again, Omar, for joining today. And like I said, we'll definitely continue this conversation offline, but appreciate your time, taking the time on a Saturday. And, you know, we look forward to bigger and better things from Tully.AI.

Amazing, Adam. Thank you so much for having me. And it was— I had a fascinating time. Thank you so much. Great. Thank you.

Have a good day.

You too.

Thank you for listening to Efficiency Unlocked, Nerd Tested, MD Approved, a NerdMDs podcast by Drs. Adam Karoo and Dale Gold.