317 Audio.mp3: Audio automatically transcribed by Sonix

317 Audio.mp3: this mp3 audio file was automatically transcribed by Sonix with the best speech-to-text algorithms. This transcript may contain errors.

Steve Brown:

What I'd been doing in AI was a concept of using a mixture of agents and models to try to create different pathways and through the knowledge that's compressed into these large language models. But when all of this happened to me, I repurposed what I was doing, took some of the soft I'd been working on related to working with multiple agents and multiple points of view and multiple models, and I started having them all look at my medical record because I had been feeling like there was something going on and feeling like I was ill for some time. I've been an entrepreneur and much of that time in medical technology, working on things like chronic care management and patient monitoring, and kind of left healthcare for a while to become a documentary filmmaker. but after the pandemic, I got back into tech and especially with AI because it was so um such an exciting time to be in tech. Um I didn't think I was gonna go back into healthcare, but then I actually had a health issue myself personally. Um, about a year ago, I was diagnosed with a rare blood cancer, and that dragged me back into healthcare. So I started taking what I was doing in AI and applying it to my own medical record, to my own situation. And what I'd been doing in AI was a concept of um using a mixture of agents and models to try to create different pathways and through the knowledge that's compressed into these large language models. And I was doing a lot of that in the kind of educational field, a lot of things with Peter Diamondas and Abundance 360, um, and for that community. Um and uh thought I was going to be not doing healthcare, but when all of this happened to me, I repurposed what I was doing, um, took some of the software I'd had been working on or the concepts I'd been working on related to working with multiple agents and multiple points of view and multiple models. Um, and I started having them all look at my medical record because I had been feeling like there was something going on and feeling like I was uh ill for some time and but I wasn't really being diagnosed with anything. My doctors were thought it was a fishing expedition. Um, but then things got a little worse after our well, this is kind of a crazy story, I know, but it's like our house burned down in the big fires in LA that displaced me from the healthcare system where I was and the doctors where I was, and I ended up in near Palm Springs staying with friends, and that's where I ended up having a uh severe abdominal pain and all kinds of things that's led me to the emergency room, um, thinking that I was obstructed from uh steak dinner. So I went into the emergency room and I said, I need a CT scan to rule out obstruction. Um after I got the CT scan, they found a bunch of things on there, lymph nodes and things that that uh they guess they hadn't noticed when I had my CT scan just two weeks before, including a, you know, I had a colonoscopy, an endoscopy, a whole bunch of tests just in the weeks prior, but um hadn't really been diagnosed with anything serious. So it was kind of a surprise um that uh to get a really serious life-threatening diagnosis where um the prognosis if you catch it too late is not very good. But luckily, because of the fire, I actually caught it just in time. Say just in time, it's you know, like not too much damage had been done. Um, and I got on kind of the standard of care therapy for it. It's a it's I basically have a version of multiple myeloma, which is a cancer in your plasma cells and your bone marrow, but the version I have, they create a misfolded toxic protein that can accumulate on your heart and your kidneys and in your gut and uh cause all those organs to fail if you if you don't catch it soon enough. So I knew it was very urgent and I needed to get this under control very fast because I needed to prevent getting my heart damage and kidney damage.

Craig:

Yeah. Let me ask, just to establish your bona fides uh for listeners, I mean, uh to give a little bit of your tech background. You were working with Peter Diem on this when this happened. Is that right?

Steve Brown:

Yeah, so I was his chief AI officer. and you know, what did that mean? That means I mean he's got a really interesting organization that has an annual conference and some other things. And I was building um apps for the membership of Abundance 360 um for the conference. I built a kind of full conference application, and then I was building in a lot of AI features into all of these app applications. So um, and then I was doing workshops and I was doing a bunch of things. Like I spoke on stage the year before. I gave a 35-minute um main stage presentation talking about what I was doing in education, which was I had brought back to life or back to AI life, um, all these great thinkers from history, you know, Socrates and Plato and Aristotle and you know, a hundred great thinkers from history. And I created them in a way where they could talk to each other. So and I could have them debate things and talk about uh topics and all kinds of things. I even had it so I even made this thing where you know you could ask Socrates to write an app for you, and then they would he would take over your computer, write an app, and like to show it on screen. So I was kind of demonstrating all of this sort of mixture of things, um, uh doing a lot with AI, just a fantastic playground, learning a lot. And um, but I was very hands-on. I was actually, you know, I'd been a CEO of tech startups a couple times before, and I had always had a chief technology officer and an engineering team, and I was dealing more with the business side, but my background was studying physics and computer science at Stanford, and you know, I started off my career as you know, writing code, but with the pandemic, I went back to you know, like hands-on writing code, kind of kind of um uh inspired a little bit. I heard uh Sergey Brin from Google talking about going back to the office and actually writing code and how this is such an amazing time to be working hands-on in technology. And I'm like, I'm gonna do that too. I mean, I started writing code. Um, and the leverage today with AI, I mean, I could do things in a weekend that it used to take you know, like 10 people a year to do. So, what does that do if you have like ideas? It it's it means you take on more ambitious projects. So I was taking on more and more ambitious projects and doing a lot of different things. I just wasn't doing healthcare because it's like I'm done with healthcare. Healthcare is too slow, you know, too painful, the life's too short to do another healthcare deal. But then, you know, so I had a a lot of technology that I'd been working on. And my I mentioned that I became a filmmaker for a while when I was maybe a little bit sort of tired of you know the solution to every problem as a you know, some version of a software. Um, I kind of went a little bit different direction for a while, but my startups were all tech startups building you know pretty serious technology platforms in in healthcare. Um, you know, we were the pioneers of remote patient monitoring and chronic care, the technology side of that. Um and you know, we were the ones that got the national contract with the VA, the Department of Veterans Affairs, which was the largest health system. And then based on that data, we got our model and our ideas into an act of Congress, and that led to getting into Medicare. And we were very early and kind of paving the way with new technologies and healthcare, but everything just took forever.

Craig:

Yeah, what was that? What was that company that you were doing the healthcare work with?

Steve Brown:

Yeah, it was called Health Hero Network. And it was, it was really, I mean, we were not we were not focused, we were focused on the patient in developing applications where the patient patients could be monitored from home in a way that kept them out of the hospital by because you would identify problems early. But really, the kind of the heroes in the picture were the nurses and case managers who are kind of the front line of chronic care, and we had a system where, you know, on one hand we're monitoring for patients and a lot of kind of education and feedback and behavioral stuff for patients, like you know, every day asking people how they're feeling and um uh collecting information about symptoms. Um and uh and then on the other side is a management system related to chronic disease and identifying problems early.

Craig:

Yeah. And then so when you went when you became ill, you went back to that software and instead of uh pointing it at the patient for the doctors, you pointed it at the patient yourself in this case for your own uh management of care. Is that right?

Steve Brown:

Well, I didn't go back to that software because that software was old news by the time I got um uh sick. But I had I had been since then, since the pandemic, and over the last three years, I had been developing all kinds of new applications, AI applications, applications that were using the APIs of the foundation models and doing pretty extensive you know ideas in in a lot of different fields in education and entertainment and um you know things related to movies. And I was doing I was doing a lot of different things with AI., I just wasn't doing healthcare. But the so I had a lot of experience, I'd accumulated a lot of experience of developing uh agentic AI applications at the point that I uh got sick. Um so I you know when I'm in the hospital, I'm when I started feeling it, you know, like you know, I'm on pain medicines, I'm on oxy, but I'm like writing code. Um so I started writing code to take and organize my medical record and create kind of a context and kind of perspective of different medical perspectives. So, you know, I had I had a you know an oncologist and a hematologist and a gastroneurologist and a cardiologist and an emergency room doctor, I had all these doctors. So what I did is I made agents kind of modeled on all my doctors. And I had them all like analyze my medical record and then debate each other about my medical record. And the interesting thing is, like everyone had a different opinion about what it might be. But they all agreed that on in order to settle this argument, you should go get this test and that test and that turned out to be exactly what I had needed. Now, I was doing this after I was already diagnosed. I basically went back in time with my medical record and asked the question, why didn't they catch this sooner? So what so I was looking at my prior data and saying, like my doctors at the time were saying it's just stress, it's just gas, you know, it's like it's something that's it's no big deal. so I well, you know, I was, I was, you know, it might have been a fishing expedition, but we were fishing in the wrong pond. Um, but I uh I went back and I said, well, why didn't they catch this earlier? So I took all of my other my old labs and imaging and everything that I had, and that's what I had the agents looking at. And they and everybody agreed, everybody meaning all of these agents, they all agreed that I needed certain tests. I needed a free light chains test, and I needed, and maybe uh that I would need a bone marrow biopsy. So that was very enlightening to me because if I had known about that test, I would have asked my doctors, I would say, hey, what about this test? And they would have given me that test if I would have asked for it, because it would have made sense. They just didn't think of it at the time. Um, but they would have given me that test. I probably would have been diagnosed, you know, maybe even a year earlier if I had known what to ask. So, you know, they like in healthcare, you know, you if you just walk in and you don't advocate for yourself or you don't ask any questions, you just say, do what you're gonna do, you're gonna get kind of whatever's in the checklist most of the time. I mean, you know, like the like the doctors, my doctors are great and they care, but you know, they're also like stressed and they've got a lot of patience and they don't have a lot of time. So the default for most people is you're gonna get whatever's in the checklist of what you're supposed to do given what we know. Um, and you know, there's not gonna be a huge amount of digging deeper until things get you know much worse. Um, but if you, you know, like you know, your doctor has like a lot of patients and has maybe 10 minutes or you know, five minutes to deal with you, but you only have you. So if you if you um have something going on and you know you want to make sure that you're getting the right care, you need to you need to become a bit of an expert in what's going on so you can ask the right questions. So I saw that for myself, and I realized, you know, like I need to become an expert in my disease if I'm gonna get the best possible care.

Craig:

Yeah. And uh the agents, this initial um pilot or whatever uh that you developed that was looking at your medical records from before the diagnosis. Uh, you said you trained one on to act as each of the specialists, I imagine. How do you do that? Is it are you fine-tuning uh uh a base LLM? Is it an open source LLM and you're fine-tuning it on you know hematology or on oncology? Or how do you develop these agents just at the not in the product that you have now, but when you were uh working this out, uh what were those Asian agents based on?

Steve Brown:

So we if you hear, you know, you're you have an audience of people who know a little bit about AI, I assume, but the that you hear these, you know, this new model came out and it has a million token context window. What the heck does that mean? That means every time you interact with the model, there's a potentially like, you know, like 750,000 words, you know, like hundreds of pages of documents that can go in with that prompt. You might be prompting something that says, hey, what's going on? You know, hey, what's going on? Four words. But if in the background of that prompt is you know 100 pages of your medical record, you know, it's setting a context. So you're not actually training the model on your um medical record. What you're doing is you're um organizing the context for your question. Um, and that is part of what the what is prompting the model. Now, there is such a thing as fine-tuning the model, adding your layer of intelligence on top of the model. We're not doing that yet. Um, but the first test was um what kind of training already is there in the foundational models? And can we get to it in a way, can we get to the medical knowledge in a way that that we can trust? Because you know, one of your challenges, you go to ChatGPT and you know, you can cut and paste stuff from your medical record and put it in ChatGPT and say, what's going on? Like there's nothing that stops you from doing that except when you've got cancer, you might have a really long, complicated medical record, so it's really not feasible uh to do that. And their terms of service say, don't do it, but because there's a lot to organize and there's a lot of kind of quality assurance that's needed. So my methodology of getting to a more um kind of reliable result was to rather than saying, hey, every time you ask the question, I get a little bit different answer rather than a deterministic response. Well, let me intentionally try to get a diversity of answers, you know, like I would with a board of directors or a tumor board or a medical advisory board. Let me get a bunch of different opinions and have them kind of cross-validate each other. let's have them sort it sort it out. And it turns out then when you do this with a mixture of agents, with a mixture of point of views, and you get a diversity of ideas, when you find convergence around ideas, that's pretty important. Now it's one thing, it's like, you know, like if you've got like a you know, whatever chance of getting some sort of hallucination. I mean, this is a problem that the that the foundation model companies have been working on like crazy over the last couple of years. And so it's gotten really, really good. But let's say there was, you know, like you know, some finite chance that you know uh something is a hallucination. Well, the chance of getting two agents or three agents or four agents or five agents to have the same hallucination is almost nothing, is almost zero. So you know, this is a methodology of getting to a more reliable result. And I think some of the models are doing some of this kind of mixture of agents in the in the background. I mean, they're doing that, and you don't even know they're doing that, but we're doing it kind of intentionally and kind of transparently because in the real world, you have a bunch of doctors. In the real world, they have a difference of opinion. Like, who are you gonna listen to? Yeah, well, you kind of want to listen to them all, get all the ideas, learn what you can about um you know what all of this means for you. And you know, like if you have cancer, there's a whole bunch of things where they're saying, like, hey, we could do this or we could do that. Steve, you decide.

Steve Brown:

Right. You know, like how do you make a decision like that? You need to get educated. So, what we're building now, it's really in this category of like, let's help educate you. The doctors are diagnosing and treating you, but it really helps to get educated.

Craig:

Yeah. Now I have a couple of questions. So you're architecting this mixture of agents. The agents, you're feeding them context in the context window, not using rag or necessarily uh fine-tuning.

Steve Brown:

There's rag means, you know, what am I taking something out of a database somewhere? You know, whether it's a vector database, you know, or you know, this sort of a uh semantic search with all these AI tools, or it's just like pulling data out of a database. Rag means retrieval augmented generation. What and what how am I going to augment what goes in there? So when I take your medical record and I kind of format the kind of most recent relevant things, you know, I don't need data, you know, I need your metabolic panel from 10 years ago. I need your most recent stuff, but I need your, you know, I like when you kind of take that um out of this sort of messy large medical record and compress it down into a really organized way, that is a form of rag that is going in. So that plus a pretty extensive uh, you know, rag puts things in the context window. That's what Rags is.

Craig:

Right, right, right. But through with search, I mean you're not taking all of your curious. medical documents and dropping them physically into the prompt, you're having uh rag search for the relevant.

Steve Brown:

I mean no well it's a I mean so rag you know there's a lot of different versions of RAG I mean if you just went you know whenever the whenever you go to um you know chat GPT and it says hey searching the web you know that's RAG it's searching for some stuff it's going to add that into the context it's not in stuff that's not in the foundation model. Anything from your medical record obviously that's not known to the to the model that's new data that came out of a database that that is you data that's specific to you know that is a that's a form of rag we retrieved it from your medical record we actually did a lot of kind of organizing it to make it processable by the AI because you can just dump it in um you need to kind of organize it. So I mean you can call that rag um you know we we're but everything that we're doing whether it's coming out of your medical record or the kind of the kind of the guardrails around how they're supposed the agents are supposed to behave or the perspective of the agents themselves that's all organized into the context that goes into the AI.

Craig:

Yeah I understand and then agents uh in what way are these agents simply because they can look things up or yeah what can you I mean as opposed to just an LLM that you're having another LLM talk to right so I mean agent is a term well here's how to think about agents that comes from agency.

Steve Brown:

It depends on how much are you trusting them to do things where you it wasn't exactly what you told them to do. The minute I take two of these agents and have them talk to each other you know that that's you know like they're prompting each other you know that's an agent. Right. Now you know a lot of agents means that means there's just some set of actions that they have some level of autonomy that they can go have agency, make decisions and do things. But certainly when they're talking to each other that's an agent. Now you know that but there's a spectrum from you know just a you know chat bot to you know something that's going and autonomously going and doing a bunch of things. We're somewhere in the middle right now where and but they you know the agency part is talking to each other cross-validating each other um and you know that that's where agent comes from but you know as you know over time this evolves and there are more tasks that we allow the AI to do to help you do um you know it becomes more and more agentic. So agentic is a spectrum of agency. And you know some of those some of those features I mean like you know there's one aspect of like I need to get educated what does this mean? I had this genomic test what does this mean for me? Like what is what do these terms even mean? It looks like Greek. And you know a lot of it is Greek symbols in these reports like how do you even know what that means like there's a lot of it is just you know like it that's chat. But there are a lot of other management tools you know if you get a serious disease like cancer and this is true of you know like pretty much anything serious it's not just you know that you have a lot to manage over time. And you know there's a lot of lifestyle factors and you know like you know if you're putting off diet and exercise and sleep and stress and all these other things, well, you know if you're in chemotherapy you got to worry about those things. I mean they interact with your treatment. So you know and if clinical trials it's not just hey is there a clinical trial that I match for but you know can I can I monitor those clinical trials and can I get notified whenever there's something new that I that I match for there's and there's AI that's determining the match and uh you know whether or not you qualify or might qualify. And so there's a lot of there's a lot of management and work to be done. So there's a lot more pieces of that where the kind of the grunt work of it can be offloaded to AI. So those are all part of the features and roadmap of what we're doing. But the on-ramp is I've got questions and uh I I need to figure out what's going on.

Craig:

Right. And the medical data that that you uh organize um it's when you say organize uh it are you curating that data and are you know putting it into like a graph knowledge graph or into a vector database or what do you mean by organizing and is that something that a lay person will have to do once your product is in general availability think of it this way when you go to a doctor and if a doctor refers you to another doctor they do a write up or workup on you and say hey this is Steve here's what's going on with Steve and then here's a few pages of stuff about Steve.

Steve Brown:

You know and it it's you know they're not gonna put in you know the fact that I had a flu shot 20 years ago they're not gonna put in stuff that's not relevant to the issue at hand they're gonna try to be complete you know we're basically reproducing uh that well um uh if you just pull the electronic medical record I mean there's a bunch of stuff in there that doesn't have anything to do with cancer um there might be things that do have something to do with cancer but you know we only care about the more recent values not the 10 year old values um so you know there's a there's a and then we have a chief medical officer and we've got a lot of medical input on this it's like the real question is what is the what is what is the kind of out of your medical record what is the kind of essential um uh information for asking you know for ed for patient education and for asking you know what does this mean and what's going on i mean it it's not everything it's a lot um and we don't want to miss anything that this relevant but if you look at the you go to my chart and you go like you know look at it it's there's a lot of data and it's scattered in different places there's a lab section and there's the visits and then there's a click on something through the note from the visit and you know there's a lot of stuff and it's kind of all over the place. And you know like how do you know like we need to categorize all that stuff. We need to put all the labs together we need to just care about the kind of the recent stuff that's relevant um there's another category of genomics and molecular profiling in cancer like that's a special category. So it's more like categorizing and organizing your personal health record so that it makes a little more sense. It's not just a data dump. Yeah you have medical record you pull your medical record you get this like massive JSON file it's kind of a data dump that's not particularly useful we need to organize it. Yeah and are you automating that organization so that so that it you can put in the URL and you know credentials for your my chart yeah it's and it goes and pulls everything it's not putting in my chart credentials but you know after the Cures Act um you you're legally entitled to get your medical record yeah it's your it's your record so and these things are you know there are health information exchanges and these things exist and like you know a lot of this work I remember when you mean years ago when I was doing uh healthcare information technology and remote patient monitoring and chronic care all the conferences every i all everyone ever talked about was interoperability interoperability and standards and all this stuff a lot of that stuff has kind of been you know it's it I think a lot of people in the industry will say oh it's still a mess but a lot of that stuff has been worked out um and there's pretty good electronic medical records the data is there um you still need to organize it and the fact is if you don't have it and you all you get is a you know PDF that was a fax from a lab to whatever you know but then we can OCR that and analyze those documents as well

Craig:

Yeah and then uh how many agents do you have is does that expand and contract depending on the complexity of the case or and what uh foundation models are you using um is it like you're you've got one based on Claude one based on chat GPT one based on grok one based you know like that or do you yeah I mean how does how does that work?

Steve Brown:

And the model is independent from the actual context management. So you can the models are interchangeable uh in that sense and when I was doing this for myself and you know the first version of this which is a in a private beta with just people that we're are able to work with closely because it's still in in development all the models are in there I mean all the foundation models are in there and too many models are in there. But as we go now to market with it um we're narrowing that field to the to the major foundation models that we can do in a HIPAA compliant way. So you know that's Anthropic and OpenAI and Gemini are the main ones but all the other ones work um you know whether or not they add you know it's kind of like if you go to perplexity for example it's like you can go to ChatGPT and you're just dealing with open AI. If youxity it's kind of the same thing but you can decide hey do I want to use open AI anthropic you have a choice of models. So right now there is a choice of models um but we're not expecting a uh you know the average person to have any idea well which one should I choose so we're doing some work there to kind of to sort that out to make it easier to kind of have a default mode where you can you know there's an easy way to just get going um and we you know we made like 36 agents but you know we based on your um kind of the summary of your condition we recommend the you know here's five that are would be good for you to use like you know like you know if you've got something going on with your liver you know then oh you should add a hepatologist in there if it's you know some cancer where it's like surgery or radiation or whatever you know add a radiologist or add a like a radiation um person add a surgeon so what we do is we kind of say hey here you can talk to anyone um and we have some that are just modeled after um you know like different branches of medicine and we have some that are sort of more sort of scientific like you know like immunology or you know immunotherapy and um genetics things like that. So you know there's a we come up with a you know like a uh a kind of recommended list of here's some that you can talk to part of it is kind of point of view we want a diversity of points of view um and we want things that are more kind of relevant to you because a lot the way people use this a lot is like hey I have a I have an appointment with my doctor coming up I have an appointment with my cardiologist coming up let me go talk to the cardiologist agent for a while to kind of get to to practice I'm only gonna get 10 minutes with this guy if I'm lucky you know like can I make sure every minute counts can I kind of rehearse that's how I use it today.

Craig:

Before I have a doctor's appointment I go in and I like kind of like rehearse it and make sure I've kind of asked all my naive questions and kind of kind of educated so when I go in if I only have a few minutes you know just make sure that I'm focused yeah and uh I asked this before and I'm sorry to ask it again but I don't uh so you've got let's say you're using uh Claude and Gemini uh let's for simplicity's sake two agents one based on Claude one based on Gemini I understand that you've cured uh the system has curated my medical uh documents and has access to them organized more than curated organized okay how what differentiates a hematologist agent from a cardiologist agent uh it uh you said you are looking at fine tuning is that right but what is it that so yes we are looking at fine tuning but for a different purpose not to fine tune on like you know fine tuning the point of view of a um cardiologist versus a hematologist um now in the end if you talk to the cardiologist if you talk to the hematologist and you have your medical record organized in there you're gonna you're gonna hone in on the same result but the fact is the point of view matters um and there's a reason you have a cardiologist because there are things that you know if you have a cardiologist when like in in my condition it affects your heart um so I have a cardiologist and it's like hey you know I there's a bunch of stuff I want to measure doesn't you know whether or not did I have any heart damage is my heart getting better so you know like that's I don't go to the cardiologist and talk about you know other stuff I talk about that stuff um my cardiologist my new cardiologist knows a lot about what I have because it affects heart you know but it's a you know it's a it's a point of view now does it really you know how much does that matter where it matters is when there's when there's you're not sure what's when nobody is quite sure what's going on or what the right thing to do is to have a diversity of opinion you're surfacing more ideas so in my case you know like my doctor my initial doctors were misdiagnosing it they had different ideas and none of them were right well then you know then I got to you know diagnose okay well that's you know that's one and done it but it's you know I got my diagnosis.

Steve Brown:

But then the question is well what do we do next and you know my main doctor was like okay well this is what we do this is what the standard of care is we're gonna put you on this um well there turns out there's a bunch of different genetic mutations in the cancer which do point to the fact that the standard of care is going to be less successful with me and uh there are some other options. Well how did I find out about that? I found out about that by you know talking to the agents. It's like hey you know what there's you know may and then I wanted to monitor it really closely. It's like okay let's start with the standard of care let's see what happens but I could see after the first month I had a good response and then it flattened out the second month it was like plateau and I'm getting this data and I'm going back to the agents and I'm saying like you know what you know what what's going on here? Should I be worried about this? So I, you know what are the alternatives what are the kinds of things what else should I talk to my doctor about? And so I get to a to a point it's like hey there are other treatments that actually are you know that's a lot of interesting things going on that for this genetic mutation you should ask your doctor about this other stuff. So I did ask my doctor about that. And my doctor who is not at a center of excellence he's an oncologist that has you know lots of different kinds of cancers um you know he was not going to stray from the standard of care because you know it's not his expertise in in that but I ended up getting referred to a much more specialized hematologist at an academic center of excellence. And I then I went to I didn't just go to one I went to a couple of other ones too for a second and a third uh a fourth opinion um and you know from that standpoint from the center of excellence you know the kind of the like academic medical center I have that conversation it's like yes we can't that is that you know if I were you the doctor is saying if I were you that's what I would do too but it's not approved for your disease it's approved for something else so it's off label. So you know to get this through we're gonna have to write an appeal letter to the insurance company um you know which they did and which they got because it's the was the right thing to do which actually highlights another challenge with precision medicine. There's all kinds of new things coming out in if you watch cable news you're gonna see you know hey do you have this gene and this mutation and this ask your doctor about Ktruda and like it's based on genomics the genomics of the cancer and all these new precision medicines that if you happen to have that specific mutation this thing is really good but sometimes this stuff is not it hasn't even had fit had a full clinical trial because there's you know it's a if you have a rare disease it's too small of a group it's not worth it for anyone to do the trial. So there's a lot of off label use in precision medicine and a lot of things that are stuck in phase two and never even got a phase three clinical trial so but it it's like there's you need to ask for this stuff because this is the new medicine that's going to be that might be exactly the right thing for you.

Craig:

Yeah now I and I'm sorry I'm gonna go back to this because I don't is it is so you have Claude and you have Gemini do you simply prompt one of them let's say Gemini say you are a hematologist and assume that it will adopt that point of view I mean how in the other one you say you are a cardiologist and assume that it will adopt that point of view or is there data that you have to give the model to have it adopt that point of view we you know the kind of guardrail on that is we have like it's probably about a whole page of you know couple pages of content that are kind of defining what a top what is a top hematologist like yeah okay how do they think what do they do what what's their kind of you know and all of that is do doing is is and you know same thing for a cardiologist.

Steve Brown:

What that's doing is it's creating a different pathway through the knowledge. And it's you know it's surfacing different um ideas. Okay, I got it. And you could use the same agent you know that because that's all built into the whole agent uh perspective you know this all of this kind of uh content of kind of defining what a great hematologist is like um and that's independent of the model so you can use the same agent you can say hey I want to see what happens if I use this agent and I let all the models talk to each other. Yeah same agent different models or you know same model different agents so you can mix and match the agents and models yeah look that's something that somebody like me who wants to dig deep and you know to explore and try to see did I miss anything I you know I want all those options um and I want to go deep and I want to try to figure out and then I end up kind of just settling on you know my favorite one and I was okay that's the one I'm I like to use. Now, if you don't want to think about any of that, you know, we kind of do a lot of that for you and kind of set up the defaults so you don't have to think about that.

Craig:

Yeah. Now that's fascinating, and that answers my question. The uh so this is designed for the patient to manage their care, to understand what the standard of care is, to have these multiple agents debate the standard of care given your medical data, and surface, as you said, tests or hypotheses that may not have been explored by your doctor, that then you can take to your doctor. Is that right? Is there going to be a version of this that sits with the doctor uh where they can you know listen to the patient, look at the medical records, and then have the system debate for them to see if there's anything they're missing?

Steve Brown:

So I that's a really important uh idea right now. People who are using it, um most of the people are using it are using it with their doctor. Um so you know, I think that there's a lot of there's a lot of interest in the medical world of like how can AI help us? I think there's also a lot of trepidation about AI, like how is this gonna make our lives miserable? So there's a lot to work out on, like say, the how does this fit in with the doctor? Um, and in my prior businesses, that's what we focused on. We focused on that, and we were really going in from the doctor's point of view. So I'm very familiar with that and I know how important that is. Um initially, when we're doing this, we're saying we know we can't wait until we've kind of convinced everybody in the health system um uh you know, to uh about this. We're going straight to the patients, and we're you know, we're saying, look, you already are going to a chatGPT. You already have questions, you already are doing, you're going to Google, you're going to chatGPT, you're asking your questions. we can help you do that, what you're already doing. We can help you do that in a much more organized way, because we built this whole application that kind of organizes that in a way that is that is really organized for someone with cancer. so it's gonna be it's a much better way to do it. Um and it remembers that whole history. And, you know, once your medical record is in there, you don't have to keep like cutting and pasting it into a model. It kind of keeps track of all that. Plus, there's a whole bunch of other tools that we're adding on there. Like, hey, I actually do want to monitor you know my diet and exercise. Hey actually, I do want to track my symptoms, actually, I do want to um search clinical trials, hey, I actually do want to go reach out to clinical trials and gonna keep track of uh um you know whether or not there's one relevant to me. So there's a whole bunch of other management tools beyond that. So, yes, you can go to ChatGPT and ask medical questions all day long, and you can uh even upload some documents. But if you're you know, like if you have a medical record like mine or like the people that are using it, that is just very unwieldy. Sure. Um, and it's also just chat GPT, and it's like, what about Claude? And what about Gemini? And what else is out there? Like how like I want to be able to bring that together and uh do it all in one place.

Craig:

Yeah, yeah, and I can see the you know, going to a single model or even a couple of models, uh, the idea of having these models debate each other, that's uh behind the scenes, right? And then they come up with a consensus view, or I mean I can see how powerful that would be. Or can you look in and see the debate?

Steve Brown:

Yeah, you I mean, there's this so, first of all, right now in the in the beta version, there's it's a pretty extensive product with a lot of features, and one of them is you can actually ask two different agents to debate each other and you see the whole thing. Um, there's another feature where you can choose a whole panel of agents, um, make it like your tumor board and or your medical advisory board, and you can have them all respond to the same prompt at the in parallel, and then you can ask another agent to go look at all those responses and synthesize it until you kind of like give you kind of a synthesized, like you think of it as like your board of medical advisory board, and then the you know, the chairman of the board, you know, that's gonna kind of like take all those different things and kind of synthesize them. And when you kind of synthesize them, and what that does is it really does a great job of of surfacing more of the possibilities and weeding out the you know the things that that might be less relevant. And then when you take all of that, you can go into chat. So when you're going into chat, it's like, okay, now I want to talk to the chairman of the board and I want to I want to talk about this. You have the medical record plus the synthesized view of all these agents, you know, plus the you know the agent uh definition of the chair of the board. So but you know what you think of it, that this is all like let me have let me have a whole bunch of people look at this, let me surface and kind of synthesize a lot of the ideas and kind of prioritize this and make sure that that's all there in the background. And now when I go into chat, I'm going into chat that's not a blank slate, it's a very informed chat.

Craig:

Yeah. and so you developed this for your own uh care. were the you mentioned that there are some tests that may not have been uh given or ordered by your doctors that that this system suggested. And can you talk a little bit about how this helped you direct your care in a way that it might not have been otherwise?

Steve Brown:

Yeah, I mean, there's this concept in medicine called shared decision making, and another one called collaborative care. It's you know, you've got a high stakes, serious thing, and a lot of decisions to make. you the outcomes are better when it's not just doctor just saying, just do this. the outcomes are better when it's a shared decision making, when patients are involved in their care and they're participating in that. And it's their outcomes are better for all kinds of reasons. Part of it has to do with just um the more agency you have as a patient, the more you feel like it's possible, you more engaged you are with that, the more hope and you know, the more energy you have for it, and you don't give up and you take it takes some energy to get through something like this. Um, you know, and the fact is, you know, uh a lot of people don't get through it. It's still like you know, 600,000 people are gonna are gonna are gonna die this year of cancer in the United States. So it's still a hugely unsolved problem. But there are better treatments every day, and there's a massive amount of research published every day. Your doctors can't keep up with all that, it's just too much. So, you know, you're gonna get something that's a guideline um for most of the time. Now, what it's considered like a pretty good therapy in cancer if it works 30% of the time.

Steve Brown:

Right. You know, this if the response rate is 30-30, you know, it's like, well, that means 70% of the people are getting no benefit from it. Um, and you know what, what if what if there was something else that you might have done if you're in that 70% group? Um, well, that's where all this precision medicine stuff is coming in. It's like, hey, this thing didn't work for you, but you have this, and I have this for myself. There's a specific mutation in your cancer cell that makes that cancer cell particularly vulnerable to this other treatment. So it's not quite as sensitive to the other stuff that's in the standard of care. So you might uh not get it the response you want from that, but it's super sensitive to this other thing. So, I mean, you it's something you kind of want to know about if something like that exists. Um, you can't give yourself that drug, but you certainly can ask your doctor about it. And uh uh and in and you know doctors want to do the right thing, but you know, they're also overwhelmed with uh uh you know, there's a shortage of oncologists and cancer is on the rise.

Craig:

Yeah. Uh and so where are you in the in the product journey with this? And is it uh is it you talk about people who are using it now? Is it on the market or uh not yet?

Steve Brown:

It's in a private beta. You know, it's you can sign up on the waiting list. Um, and there are people who've signed up on the some people that were starting to let in. There are people that we only when we can kind of closely work with them right now. But uh early next year, we're gonna open that up um so that people can just do it for themselves and uh you know, have a really great cancer co-pilot that can help them uh become more of an expert in their disease. It's this what we're coming to market with is patient education. You know, every single thing we do says, you know, you might want to talk to your doctor about this, or you know, this this is patient, you know, we you know, whether or not we have all the information about you kind of depends on whether or not you gave us all the information. So it's an educational tool. And when you become educated, or on behalf of a family member, or some people are doing this for a parent or a relative, or it's like you know, they're the ones having the conversation with the doctor. But when you go in, you kind of want to have done your homework first and not be dumping a bunch of stuff on the doctor that might not even be relevant to you, but try to figure out what's relevant to you before you go into the doctor and be really smart about it so you can so you can be an active participant in in your care with your doctor. Now, I talked, we talked the examples we talked about, there was a lot about like diagnosis. But you know, that diagnosis is you know, that's you get your diagnosis and you got your diagnosis. Um the most important thing is like, you know, like things keep changing. You need to keep monitoring how you're doing, you need to understand if something's going, you know, if it's you're getting a relapse or if it's you know something's going sideways. Yeah, um, you want to know about the stuff, you want to know what to look for. You want to be educated in all these things because you're gonna go back to your doctor. If you have cancer, probably go back to your doctor every month in that first year. Um, and you're gonna get a lot more tests. And so from in my case, I identified some ideas that I talked to my doctor about, and you know, I ended up finding doctors who said, Yeah, that's what I would do if I were you. It's like, okay, well then please do it. Please prescribe it for me. And I got on to a much better treatment. Um, uh, and you know, but then the question is was, well, should I, you know, what other things should I be doing? Um, so a lot of the cancer therapies and my cancer and cancer therapy was is immune compromising. Like my plasma cells that are supposed to make antibodies were not making antibodies, they're making something else, they're making something toxic. So I need to stamp all that out, but no, I'm immune compromised. So, what can I do about my immune system? How do I keep healthy? You know, I a friend of mine, an old friend of mine, when he heard that I had um a ver a variant of it's related to multiple myeloma, but it's is you know, I had basically a variant of multiple myeloma. And he said, Oh, my brother died of multiple myeloma. And then he said, Well, he didn't actually die of multiple myeloma, he died of an infection. And I'm like, oh man, I you hear about you know, immune compromised people get COVID, and those are the ones who died of COVID. Or you hear about you know people dying from the flu. Well, those are people like you know, immune compromised people who are on chemotherapy. Um, so my question was, well, what do I do about that? How do I, you know, so I'm trying to understand what are all the possibilities. And it's well, it turns out you can get infusions of antibodies from other people. And I didn't know about this, but the minute I asked for it, it's like, oh, that's a good idea, Steve. Yeah, we'll prescribe that. So then I then I'm monitoring very closely on the key marker and it's going down, going down, but it has a little blip. It didn't go down for a couple of weeks. I'm like, what's going on? Um, so I'm talking to the agents about this. Like, you know, is there something I don't know? Is there something about uh this new drug I'm taking? You know, is it like how would I know if I'm getting resistance? I'm getting educated so I can ask my doctor, like, should I be doing something different? Like, do I need a different dose? What do I, you know, I what's going on here? Well, it turns out, you know, the pharmacist said, just take this drug with food. Well, it turns out it's something that you actually need to take it with fat.

Steve Brown:

But it's just it's buried there in the, you know, in the you know, the detail, you know, that thing you get on the label says take with food. Somewhere down in page 10 of the fine print, it probably says, you know, take it with high fat. But um that, you know, I didn't I didn't know that. I mean, you know, who reads the whole 10 pages of all the things that come with a prescription? I mean, it um and no one mentioned it to me, but uh I I basically started taking the same medicine. I started taking it with dinner and not doing low fat, you know, and did it with fat. And instantly I see in my next results it's I'm back on track. So, you know, there's all these factors that influence how you're doing, and you gotta get educated.

Craig:

Yeah, um, and so and I'm not interested so much uh for business reasons, you know, how the startup is funded, but is this gonna be affordable? Uh and is it gonna be a subscription or sort of pay as you go? Because you were talking about uh had you had this a year earlier, um you know, you may have caught it, but maybe you weren't uh necessarily feeling that you needed it a year earlier. Is this something that people will be able to dip in and out of or have available every time they go see a doctor and they put their tests into the system and just kind of to monitor their health and see if anything's surfacing that they aren't aware of?

Steve Brown:

I'd say over time we're going to expand, but the initial focus is people who've been diagnosed with cancer or a family member of a person diagnosed with cancer. Right. Um those people, I mean first of all, because I mean it I get to go to my own experience, and I've gotten to know a lot of other people, and we have other people in our company um who've had similar experiences, and like we know what the need is there. You get diagnosed, you have a lot of stuff to manage, a lot of stuff to deal with, a lot of decisions that are you're being asked to make. Um, and you need you have a lot of questions. You need something that can help you get educated in what's going on so that you can be a more effective participant in your care. That's the initial um uh go-to-market. Um, a lot of the same principles and tools would work for autoimmune diseases, neurodegenerative diseases, other diseases. But we're our point of entry is you got a serious diagnosis, now you have questions and you've got a lot to manage, and you've got a lot of decisions you have to make. You need to get educated. That's our that's where we're starting. Um, but uh, you know, over time I think that we'll you know we're gonna keep going and we're gonna see what people need and what people uh want and this will evolve. But we got our starting point is that there is a um I mean we put it in this emission of the company is to accelerate cures for cancer. Um and that's a bold, kind of ambitious, but that's why we're doing this. And we know that the pathway, you know, and I don't we don't know how many years it's gonna take, uh, but we know that the pathway is this thing called precision medicine. The fact is, every cancer is different because it's your genes mutated in a unique way. So every cancer is just by definition, everybody has a unique cancer. Some point, everybody is gonna have a unique optimized treatment that's uniquely tailored to their cancer. That's the ultimate problem that we're gonna solve. Right now, we're doing patient education over time as we have more and more people that that are that we're working with, we will start to train our own models with this, and we will we will go deeper and deeper and deeper into oncology because our goal is to accelerate the cure.

Craig:

Yeah. Yeah, well, that's uh that's fascinating and commendable. if someone wants to track the development uh and when this might be available more generally, where do they look?

Steve Brown:

Yeah, go to curewise.com and it's gonna say you can join the waiting list. When you join the waiting list, it asks you a few questions about like, you know, are you are you a patient? Are you a family member? Are you a clinician? And you know, we ask them if uh ask about you know what your um what you're dealing with. Um from the waiting list, um, you know, we right now it's a small team and we're hiring and we're growing, but we you know we there are kind of uh a few people that we're starting to invite in and we're still gonna invite a lot more in. Um and then when we feel like we've really gotten this you know solid, we've done it with enough people and we we've kind of gone through um uh also all of the kind of security audits and those kind of things. And we make sure that it's really we want this to be something that people really trust, that um that people know has their back, that's done, you know, the founder of the company, me, I founded this because um I I it's like I did this for myself initially because I'm dealing with this. And um, you know, I wasn't sure whether or not I wanted to tell anyone that I was dealing with this. You know, it's like a lot of people, you know, it's you kind of feel like, gosh, you know, my career's over, you know, like I don't know, or my life's over. I don't know what's gonna happen to me. Um I kind of you know I've resisted at first and I said, oh, I'm just gonna tell everyone what's going on with me. And I'm gonna see if hopefully we can do something that helps a lot of people. Um and so far, the people that we have um worked with on this, uh, I think I think everyone has had a it's impacted them in some significant way. So I know that there's so much possibility here to really help people, and that's what we want to do. So we need we need to get the word out, we need you know, to for people to try it. We're working on the business side of it. You know, you the there's a reason why you know it's in the news, AI all the time is in the news and data centers and all this stuff. I mean, this stuff is does. Cost money, you know, we're gonna charge a subscription fee for it. We're gonna try to make it as accessible as possible. We're also gonna make it easy for family members or friends to chip in. So, you know, we're gonna make this as accessible as possible, but we also have to make this um something that where you know the business works so we can grow and serve more people.

Craig:

Yeah.

Sonix is the world’s most advanced automated transcription, translation, and subtitling platform. Fast, accurate, and affordable.

Automatically convert your mp3 files to text (txt file), Microsoft Word (docx file), and SubRip Subtitle (srt file) in minutes.

Sonix has many features that you'd love including enterprise-grade admin tools, collaboration tools, automatic transcription software, share transcripts, and easily transcribe your Zoom meetings. Try Sonix for free today.