PTW3 055: AI's Power Medicine, Diagnostics and Healthcare with Bert terHart and Donna Mitchell

Thanks for checking. In the Pivoting the Web 3 podcast. Go to pivotingtheweb3podcast.com to download and listen or web3game plan to check out the videos. Thank you. Good morning, good afternoon, good evening. Welcome, welcome, welcome, welcome to Pivoting to Web three. And this is Donna Mitchell. And welcome, welcome, welcome.

Donna Mitchell

00:00:18 - 00:01:26

We have an exciting guest today, and our guest is from Canada. It's Bert Tearhot. And he's self described as a soldier, a saint, sailor, adventurer, serial entrepreneur and author. Just seems to have a knack for snocking off the impossible and talking to him, I believe he's done just about some of the craziest and nicest things, all at the same time. A Fellow of the Royal Canadian Geographic Society, Explorer in Residence for the British Columbia Historical Society, founder of the Canadian Interactive Waterways Initiative, CTO of Med Wake, Canada, CEO of Leadbrain AI and author of, among others, the children's book Sur Salty Goes to Sea. Bert has sailed solo nonstop around the world, into the Bering Sea and out to the Aleutian Islands, all in an effort to follow in the wake of some of the world's greatest explorers and cartographers. In the same vein, he paddled solo across Canada from the Pacific to Atlantic Oceans, covering more than 7,800 kms. What is KMSS?

Bert terHart

00:01:27 - 00:01:27

Kilometers.

Donna Mitchell

00:01:27 - 00:01:28

Kilometers, yes.

Bert terHart

00:01:29 - 00:01:30

It's like 5,000 miles.

Donna Mitchell

00:01:30 - 00:02:00

I should have known that, you know, I could tell I'm losing it here, but my kilometers by foot and canoe, what's coming next is even more extreme or crazy. And I don't think it's really too crazy. I've heard a little bit of his background and his story. You guys are going to love him. And at the end of the day, I think he's had an exciting life. Some I would have liked to have, especially with all the water. So tell us, how did this become part of you and who you are in your story? Give us a little insight before we really delve into AI and everything else that's going on.

Bert terHart

00:02:00 - 00:02:23

Well, firstly, I want to say thank you for having me on. I think it's a real thrill and I'm really excited to share some of this stuff. And it's nice to talk about technology because typically people want to talk about some of the crazy stuff. And my dad, whenever I told my dad that I was going to sail nonstop solo around the world without any electronic navigation, the first thing my dad said to me was, I thought you were smarter than that. So. So he was. He was. He's right, of course.

Bert terHart

00:02:23 - 00:02:52

But, But I, I think I, I simply have salt water running through my veins. My, Both my mother and my father's family and their, you know, their, and their families are so. Their extended families have all been in the Dutch merchant marine. I mean, literally every single one. Uncles, my grandfather's on both side, my uncles on both sides. So. And when my parents came to Canada, they just stuck me in a little boat when I was a kid. And, you know, as a kid, you, you, you paddle this here and there and put a sale on it and try to get it to go from A to B.

Bert terHart

00:02:52 - 00:03:24

So it's just been in my blood and I've, I've pursued that, that same sort of passion in different contexts as I've, you know, as I've. As I grew and, you know, fell into different things. So like you said, I was a soldier, I was a scientist, a sailor, serial entrepreneur, all these crazy things. But as a scientist, I was a physical oceanographer. That meant I studied large, large period waves in the ocean. You know, as a sailor, of course I'm sailing. And as a, as an entrepreneur, I've always been involved with something that has to do with the water. In this case, the Canadian Interactive Waterways Initiative.

Bert terHart

00:03:24 - 00:03:46

So they're all tied together in this rather, you know, loose bundle. And that rather loose bundle has allowed me to, you know, pursue my passions and do things I really like. And one of the great things about being an entrepreneur is that you get to work with people you really like. You get to create the culture and the companies that you're trying to run and trying to grow. That's. It's up to you. You get to decide how, you know, how everyone plays together. Nice.

Bert terHart

00:03:46 - 00:03:56

And you get to decide who you really want to work for. So it's a very, It's. It's been a real privilege and a real thrill, and I hope I never get too old to be thrilled by all those things.

Donna Mitchell

00:03:57 - 00:04:02

Well, with all the things that you've been exposed to and the direction the world is going, how did you end up in it?

Bert terHart

00:04:02 - 00:04:26

Well, my. When I was in, When I was in college, I had to study programming, and I hated it. But, but I had to do it because I was a. I was in. I was a physicist basically, you know, solving applied math problems. And of course, you do that with a computer. And then when I got to graduate school, I did, I did the same thing. So when you solve these, you know, if you do anything, if you have, if you study the ocean in any way, shape or form, as a physicist, you have to model that.

Bert terHart

00:04:26 - 00:04:55

It's like a meteorologist knows math. So. And they solve those. And we all know they solve those problems with computers. But way back when I first started, computers actually weren't a thing. We had to stand in line literally to get time on the Cray supercomputers that were running in Colorado Springs. But I had programming skills. And a very good friend of mine, I've known him since I was in grade three, he became a medical researcher, and he came to me and he said, hey, Bert, I want to start this really big clinic in California.

Bert terHart

00:04:55 - 00:05:21

And I have these ideas about how I want that clinic to run, and I want to run it on software. But I don't know. You're the only guy I know that knows anything about programming. Can you do it? And I said yes. So I fell into it. I fell into the IT space, and I've been there for 25 years, at least, probably longer, just writing code, and I started writing out code for healthcare professionals. And that just grew and grew and grew and grew. And you know, when the Internet really came, well, when Google really came online, it was.

Bert terHart

00:05:22 - 00:06:04

It was started by a couple of scientists, a couple of scientists in California. And I immediately recognized, well, the whole scientific community immediately recognized what they were doing when it came to Google search algorithm. So I said, okay, well, I can understand that a little bit. If, and if this is going to be important, then why don't we just give Google what it wants? So that's, that's what I've been doing for, for a while now is just literally giving Google what it wants. And it's way better to, like I used to say, it's, it's much easier to, to, to join the high costs of healthcare as opposed to fight against it. And if you want to be found anywhere online now, I mean, it's much, much easier to play along with Google than, Than it is to decide to sort of swim upstream. So I literally fell into this, into this space because I had a. I had a.

Bert terHart

00:06:04 - 00:06:08

A pretty rare scale at the time, and here I am all these years later.

Donna Mitchell

00:06:09 - 00:06:58

So now in the healthcare space, you've been in here for a while, you've seen it change from what it was and what it is today. And now we have the technology of AI. How has that supported what you were doing and have it move even further into healthcare and actually helping that space, or has it hindered it in some ways? Let's talk about the good things that you've seen that you've made happen with AI. In it, in healthcare. I know there's a lot of work being done on the diagnostic side. There's a lot of work being done on the patient care side in regards to records and transferring of records. Can you share with us a little bit about that and what we need to be aware of and understand?

Bert terHart

00:06:58 - 00:07:39

Well, I think the first point that I would make is that we've, if you'll forgive the pun, we've been literally swimming in a sea of AI for a long time. Like the first program I wrote for this doctor I mentioned a long time ago was differential diagnostic software. So that we were able to intake a large number of patients and do the triage very quickly. And we did it very accurately because we were using a computer to do that. And that was a long time ago. So this AI has been creeping into the medical field for a very long time. And when you go to, like today, for example, just Forget today, take 10 years ago ago and you have an X ray done, then of course a doctor will look at that, you know, will look at that X ray and then, and then make some sort of diagnosis. A radiologist will do that.

Bert terHart

00:07:39 - 00:08:46

But a computer can do that better than perhaps a doctor can because the computer has a much finer eye to look at details that, that the doctor might miss or the, or the computer can compare images, can compare thousands and thousands and thousands of images, and it will never forget one. And AI's been, you know, I shouldn't say AI, but computer tools have been doing that for a very long time. Imagine you have a tool that will just lay two images over top of each other and you can sort of see the differences. So this, this, this AI has been creeping into the medical field for, for decades, but it's, it's just recently exploded. So here in Canada, for example, in terms of, in terms of things that, that, that are helpful, when you sit down in front of a doctor and he asks you how you feel and you start, start, start listing off symptoms. The doctor just, all he has to do is Google those symptoms and he'll get a diagnostic, he'll get a, a diagnosis. In most instances, the diagnosis will be more thorough than that which one of his colleagues could have done maybe 15 years ago, because his colleague, 15 years ago or 10 years ago or even five would have had to remember everything. And of course, to become a doctor is extremely challenging because there's just so much to remember.

Bert terHart

00:08:47 - 00:09:05

It's just shockingly, there's a shocking amount of information to remember. And over and above that, as soon as you Graduate. I mean, this is, this is something else. As soon as you graduate, I mean, you have to. You have to go back to school. There's continuing education credits. There's all these things. But I mean, they.

Bert terHart

00:09:05 - 00:09:31

Their training, their formal training ceases basically at the time that they graduate. And like I said, there's continuing education credits. And some doctors may not agree with this, but of course they should, because I'm not saying that they stop educating themselves, but the bulk of their education stops at some certain time. The bulk of my scientific education stopped when I, you know, when I stopped being a scientist. So. And AI just continually learns and learns and learns. It never, it never. Its education never stops.

Bert terHart

00:09:31 - 00:10:23

That's, that's, to me, that just keeps the building blocks of what it's relying on to form a diagnosis or to. Or to create some sort of prognosis just gets better and better and better because the knowledge base keeps getting bigger and bigger and bigger. Like, think of the number of scientific papers that get written in a year and how is it possible for any human being to keep up with the scientific literature if even if you are incredibly specialized in your field, it's an impossible task. It literally is an impossible task because there's more and more scientists who are studying more and more specialized things that can happen in your field, things terrible for things. And AI has no problem keeping up with all that. And doctors, of course, are trained to ask really smart questions. And if you ask AI a really good question, you'll get a really good answer. So I think that what I've seen in general, in terms of its benefit, is that it's, It's.

Bert terHart

00:10:23 - 00:10:44

It has this amazing ability to diagnose, and it has this amazing ability to create a prognosis, and it's unbelievably accurate. And it's. If it's done properly, there's. There's no bias in it. Like, there's, There's, There's. It's only just the information. So it's, it has this. It is.

Bert terHart

00:10:45 - 00:10:52

There's an incredibly bright future for it. Like, the trajectory of it so far in healthcare has been. Has been fantastic.

Donna Mitchell

00:10:52 - 00:10:53

There.

Bert terHart

00:10:53 - 00:11:04

There is other sides to it that. There are other sides that are. That are troubling, but the way that it's used today, I think, has been. Has been a huge plus. So.

Donna Mitchell

00:11:04 - 00:11:08

Bert, I'm chuckling because, you know, I want to get to the other side of it, but before we go there.

Bert terHart

00:11:08 - 00:11:09

Yeah.

Donna Mitchell

00:11:09 - 00:11:11

Let's stay on the positive side.

Bert terHart

00:11:11 - 00:11:11

Okay?

Donna Mitchell

00:11:12 - 00:11:14

Positive side. Let's make sure everybody Understands.

Bert terHart

00:11:14 - 00:11:14

Yeah.

Donna Mitchell

00:11:14 - 00:11:43

So you're in the healthcare space and physicians, diagnostics and all of the analytics take place and it's really great. So it's probably, is that a combination of the machine learning, predictive, generative. Is all of that together? Can you give some insight? People are starting to learn about that, that there are differences in AI, so they understand their businesses, what they're using on the applications, which ones are really doing the diagnostic work. How is that on the back end? Maybe can you share a little bit of insight without getting too geeky?

Bert terHart

00:11:43 - 00:12:35

Well, I think there's, there, there's all kinds of companies that are, that are developing applications that are for doctors in AI, but they're, they're all basically using one or two engines. So basically everyone is using ChatGPT and they're just building applications on, on top of ChatGPT. And what they're doing are the companies that are doing that are creating very sophisticated front ends, you know, that are, that are, that are founded on, on ChatGPT and, you know, Gemini Copilot. You know, there's, there's a couple of different flavors, but it's really the UI that's, that's important and embedded in the UI are very sophisticated prompts. So I had alluded to asking really good questions. So if you ask a really good question, which is, which is the analog to creating, you know, a very sophisticated prompt. So the, and the more sophisticated the prompt, the better the answer. So some of the prompts that are written for healthcare are 3,000 or 4,000 words long.

Bert terHart

00:12:35 - 00:13:29

So it's a question that's basically, you know, the better part of a chapter of a book. And if you ask these really good questions, you'll get really good answers. So for example, we've tried using TALLY in, in the, in the clinic that I'm involved with, and we see about a thousand patients a day, all remotely, by the way. All, all driven using, using code that's, that's, that's reliant on AI in a very, in a pretty limited way. But, but what the doctors use TALLY for is that the interactions they're having with, with, with patients are recorded. So a doctor used to do that by writing notes, right? He would scribble notes on a pad or scribble something down. It would go to a medical assistant whose job then was to translate that chicken scratch into something that made sense. And then the doctor would get something, file, that thing would be charted and the patient, if they wanted, would get something like, you know, hey, you know, maybe you should do this.

Bert terHart

00:13:29 - 00:14:28

But all that was manual and tedious and fraught with error and peril. So, so we, you can get Tali to, to listen to the, both sides of the conversation and produce notes that are, that are impeccable and immaculate and make sense because you, you can say to the, you can tell Tally that for the chart, you want it to be written, you know, very, I guess simplistically, you want that to be written for someone who's a board certified plastic surgeon with 25 years experience, who's the, who's the director of a very large medical clinic, you know, with all these qualifications. And that's how you want his notes to be written. That's what you want to appear in the chart. And then you'll say, you'll have, you'll have it, have it something differently to say, this is what I want, this is the precis of that conversation, that I want to be in the, in the front office. But I want that to be written for someone who has administrative duties, you know, with blah, blah, blah, blah, blah. And then you say, I want, I want, I want a letter to go out to the patient. And I want that to be understandable based on that, you know, based on what, what you know about that patient in terms of their, in terms of their own life history.

Bert terHart

00:14:28 - 00:14:59

So say, for example, if I was another doctor getting diagnosed by another doctor, then I don't want to be talked to like I don't have any medical knowledge at all. Right. I would want something I'm getting back from that doctor in terms of a letter or what I need to do or what the diagnosis is or what my prognosis is. I would want that to be at, you know, at my level, a medical professional. But if I'm, if I'm a school teacher, then obviously I have a different background, a different training. I would want that to come in a way that I can understand it. And AI does that like that you just tell it to do it. Yeah, you can say, I mean that AI can read the patient's record.

Bert terHart

00:14:59 - 00:15:37

We know what they, we know what their, you know, what their life story is, their life history, their other medical challenges, whatever they might be. And then you can, we can literally tailor the letter that's going out to them after every doctor's visit, whether it's remote or whether it's, you know, in, in, in the clinic to that very particular individual. When they're very particular. Yeah. So it's, it's, those things are amazing. Like, what's, I mean, think of it think of how frustrating it was to take your prescription to the. To the. To the drugstore that no one could read, right? And you hand it over to the pharmacist and, you know, they're scratching their head and they're on the phone back, they're trying to decipher it, and they're on the phone to the clinic.

Bert terHart

00:15:37 - 00:16:06

You know, what. What is this? And so there's. That's a very small example of just one patient. Physician interaction gets amplified. Amplified in the sense that it's way more useful to everyone, to the doctor, to the staff in the clinic and to the patient, and eventually to other professionals who are associated with that particular visit. It might be a pharmacist, it might be a physiotherapist. It could be anybody. Right? So, yeah, there's lots.

Bert terHart

00:16:06 - 00:16:09

It could be you. Yeah, so there's a drug rep, remember?

Donna Mitchell

00:16:09 - 00:16:36

So when I go into offices and ask questions about diagnosis or why do I need to take that medication, can I do a reversal? I don't want to get diabetes, but I don't want to take that drug. And, you know, the meantime, I'm very clinical. They have a raised eyebrow and get a little put off. That would really help. Really, that help that interaction, is what you're saying. And I get my notes. It makes more sense. And I won't be looking at it like, well, this doesn't really mean anything based on what took place or it means something, but it should have had more detail.

Bert terHart

00:16:36 - 00:16:50

Yeah, exactly. It can be. It can be entirely tailored to your own. To your own life experiences and the way you want it, because you can go in and say, hey, you know, the next time you go in, they could say, well, I needed more. I. I need. Because some people might. Might be very concerned about the, you know, the, The.

Bert terHart

00:16:50 - 00:17:08

The contraindications of a particular drug. And when they make that, when they. And there you go. And when you go to the staff and say, hey, you know what, what about this? They just make a note on your chart. Contraindications are very important. You know, please elaborate. And then it will just do that, and it won't do it based. It won't do it on what the.

Bert terHart

00:17:08 - 00:17:42

On what the doctor thinks they might be. It'll do it on what the latest research is because you told it to do that. So the doctor's idea of what the latest, you know, contraindications are for that drug might be 10 years, might be 5 years old the last time the drug rep was in there. Right. And of Course, there's five years of research and five years of other clinical trials or five years of other data that the doctor in all good, you know, in all good faith and consciousness just might not be, you know, he's human. He can, he or she is human. They just can't be everywhere, you know, and all knowing all at once. So there's, there's, there's real, there's, there's a really, really, really bright side to AI in that context.

Bert terHart

00:17:42 - 00:18:30

And it's literally exploding because there's people, there's just vast sums of money pouring into that sort of, that sort of AI where you're looking, you know, you go into a business like healthcare, for example, and you look at all the friction points. So the ones I describe would be a friction point between the doctor and the medical assistant who has to transcribe notes. Okay, that, that's one. And then you go to the friction point that, that has to do with like for, for your example, you walk in and, and you know, you, you look at the letter, you make you, you, you call the office and say, hey, you know what? I'm not sure I understand this. That's another one. So you can always. And you know, they're wildly separate, those two and everything in between. And when, you know, people just keep, businesses just keep stepping back, looking at the tools they have available to them, which is ChatGPT and Gemini and Copilot and soon to be Musk's Aix, so, or X AI.

Bert terHart

00:18:30 - 00:18:59

So there's, and there's very motivated, very keen, very well funded, very smart people who are, you know, plugging in AI into all those places where there's, where there's just a little bit of friction. And again, I mean, maybe I don't want, I won't belabor a point here, but people have been, you know, autopilots have been flying commercial airliners for, for decades. It's AI. That's, that's all it is. Right. It's a very intelligent pilot. And they're way. And they're typically, they're, you know, if, if.

Bert terHart

00:19:00 - 00:19:03

Well, they're, they're very good pilots. I won't say that they're better pilots, but they're.

Donna Mitchell

00:19:03 - 00:19:11

Well, I want you to go there because that's my other life in the airline, you know, so avionics and things like that in the aviation site. So go down that road just a little bit.

Bert terHart

00:19:11 - 00:19:59

Sure. I mean, well, if you. Like I said, autopilots have been flying airplanes for a very, very long period of time. An Autopilot will fly, a human gets tired. So typically a human might be able to fly a plane better than an autopilot, the best autopilot in the world, for maybe 10 minutes, but not 20 and certainly not 20 minutes and certainly not an hour and certainly not 20 hours. So there's this idea that, you know, humans naturally get fatigued, they get tired, they, you know, they get distracted, all these kinds of things, but autopilots just don't. So when I, so when I sailed around the world, for example, it took me 265 days and I steered the boat for seven hours. And I had an autopilot, a very dumb piece of AI that steered the boat, not me.

Bert terHart

00:19:59 - 00:20:45

So 265 days, I did it for seven hours. So, and I have, I have friends in the airline industry. And if you actually fly the airplane, like if a pilot actually takes it off autopilot, he'll, he'll get a call in Canada. You get, you, you get a call, you get hauled in front of, hauled in front of the chief pilot for that particular airline. And you have to answer to the Federal Aviation Board or the equivalent here in Canada as to, you know, the faa, the equivalent, why you were flying the airplane because you're not supposed to be, you're supposed to land it and take off and that's it. And the autopilot does the rest. And in fact, you're in Canada, you have to, you have to let the plane, the autopilot land the plane, every certain number of landings just to make sure that, you know, that it's still learning, annoying and knowing what it's doing. And that's been going on for a very long time.

Bert terHart

00:20:45 - 00:21:53

And they just keep getting better and better and better because of course they keep learning about how, you know, what's the best way to fly this particular airplane, whether it's a, you know, a Boeing 767 or an Airbus or, you know, whatever the case might be, they just get smarter and smarter and smarter and smarter and better and better and better and better. So there's, it's just been around for a very long time. Like one of the reasons, as you know, speaking of autopilots, one of the reasons that Tesla and Musk has basically a decade on everybody else in AI is because he has the, he has the largest self learning database, if you will, that there is because Teslas have been gathering, as long as Teslas have been on the road, they've been gathering information about how drivers drive. Oh, he's got data, he's got data. He has, he has, he has data like nobody else. And of course it's one thing to have the data, but how the thing about AI is typically is how the data is actually talking to each other, literally. You know, it's asking itself questions and generating answers. So having the data is one thing, but then having the code on top of that, that asks all the smart questions and then gets out the smart answers.

Bert terHart

00:21:53 - 00:22:15

So he has data that no one else has, and he. It's real world. This is not, this is not something that's just made up. It's real. It's real people, right, driving real cars in real circumstances. And then, you know, this, this data layer sitting right in the midst of all that, in the same way that, you know, autopilots are sort of autopilots on. On airplanes have been gathering data, but not nearly as. As intelligently as.

Bert terHart

00:22:15 - 00:22:23

As Tesla's been doing it, of course, because they've had no way better tools. So again, it's just everywhere. It's just everywhere, you know, everywhere that you look.

Donna Mitchell

00:22:23 - 00:23:37

Yeah, it is everywhere that we look. But now that we're talking about everywhere that we look. And thank you for those explanations and the details in the different spaces and places, because even for myself, you said some things that I really didn't realize or connect or AI in some ways. I remember a lot of the aviation days and simulators and different things and we've had those conversations, but I didn't connect that some of the AI and the landings and takeoffs and just transatlantic or just different things, how that's kind of creeped into our lives throughout this time. And then on the clinical and healthcare side, I think that's so phenomenal that there's the possibility that this would really come into the health care space, which would really help the care of the patient. Oh, yeah, really help understanding or providing preventive tools or direction or more clarity where you can really understand. So do you think a lot of this takes into consideration or really minimizes, let's put it this way. Does it minimize a lot of the diagnosis or the misdiagnosis and the impact on society in regards to unnecessary or frivolous lawsuits or things that really don't get missed or shouldn't have been missed? How does that play into that space, do you think?

Bert terHart

00:23:37 - 00:24:03

Well, I think that, you know, I think it plays into that space very directly. The, well, first of the AI as we know it today just doesn't make the same kind of, you know, human errors like, you know, I Think as a. If once you graduate from. From. From medical school, you head off to be an intern, which means that you work 22 hours a day, and then the other two hours, you know, you have to eat, sleep and. And, you know, do whatever you have to do. Travel or commute back and forth to the hospital. That's just the way that interns are trained.

Bert terHart

00:24:03 - 00:24:18

And they're trained for that. For. I don't honestly know what that reason is. It's. It's a lot like boot camp in the Army. But, you know, the stakes are a lot higher if you're, if you're an intern in the doc, if you're an intern, because it's not just your life, it's someone else's. But, but think of the mistakes that. That interns make just because they're tired and hungry.

Donna Mitchell

00:24:18 - 00:24:19

Yeah.

Bert terHart

00:24:19 - 00:24:28

And they're. And they. And they make them by the bushel load. And, and we're used to that. I mean, because they're humans. But all those mistakes are, Are, are. I mean, they're. I shouldn't say.

Bert terHart

00:24:28 - 00:24:43

I mean, they're not common, but they're, you know, those, Those kinds of mistakes are made. And everyone knows about, you know, everyone's heard about someone who's had the wrong leg operated on or. Or they happened to my mom. Exactly. She had a wrong hand operated on. So it's. It's just. I mean, think of the questions.

Bert terHart

00:24:43 - 00:24:56

When you go into surgery, they want to make sure they have the right person. You know, what's your name, what's your address? They ask you all these questions. You're taking chicking off the box. Yes, it is Bert. Yes, it is Bert. Yes, it is Bert. So, I mean, if you. All AI has to do is scan my face and it's Bert.

Bert terHart

00:24:56 - 00:25:13

It can't be anybody else. I don't have to answer any questions. I don't have to be conscious. I don't have to be anything other than, you know, just there. And in some cases, I don't have to be there because the surgery can happen robotically. Like, think of the number of surgeries that are done robotically. They're done, you know, endoscopically with, With. With a camera.

Bert terHart

00:25:13 - 00:26:06

Like, you don't actually have to make a huge incision in someone's leg or knee, for example, to. To do something to have some sort of arthroscopic surgery or has it has to do with, you know, fixing a ligament or something, where before you'd have this giant gash in your leg where they had to Open everything up just to see. But that's, I mean, it's, yeah, it's, it's crazy. It literally is crazy. And I think AI will go a very long way to making sure that, that those kinds of mistakes, the simple mistakes that are made because, you know, we're just go by the wayside, the simple, the simple misdiagnostics where, like, you know, some of the, some of the worst cases are, are where, where the, the symptomology is very complex and the disease or the, or whatever's ailing that person happens to be quite rare. So it's very easy to miss. Like, let's, let's say a particular diagnosis might take 10 things all lined up in a very particular order. And if you get all those right, then, you know, you, you know, you, you win the prize.

Bert terHart

00:26:06 - 00:26:43

And the prize is something that's, you know, not very pleasant. But if you get, if you get nine of them right, and not 10, then of course admit that the diagnosis is somewhere else. So that's a pretty complicated problem. You know, we have 10 variables in 10 different ways. That number is actually 10 times 9, times 8, times 7, times 6, times 4, times 3, times 3, times 1. So it's a, it's a huge number, right? So it's, it's, it's very easy to misdiagnose. And those, typically, those kinds of problems tend to be very, very severe. And tragically, quite often they occur in young people.

Bert terHart

00:26:43 - 00:27:37

So AI is way, way, way better at finding the needle in the haystack than we are, that's for sure. So in that way, it's way better. And the other thing I would say about it is that with AI and technology in general, you are able to meet the patient where they are. So the simple way to think of that is I don't have to go into the office to talk to my doctor. I can sit down and talk to him and he can sit in front of a computer and he can, you know, he can diagnose me just by asking Google some really good questions or AI some really good questions. That conversation can be recorded by, by AI. AI can actually, you know, you could actually scan me, could look at me and say, well, how does he look today then? Then how, as compared to last year, whatever the case might be, so I don't have to go to the doctor's office. Like, think of the number of elderly people who find it really challenging to get to the doctor's office and then don't go as a result.

Bert terHart

00:27:37 - 00:27:58

And my, my dad's a really. Or was. My dad's passed away. He was 96 when he passed away, but he was a really good. He's a good example of that. He was still driving at the time, but. But that was very rare. And his job, I guess, as he got older was to drive all his friends, the people, his friends who were still around because most of them couldn't drive and most of them wouldn't go to the doctor because they couldn't get there.

Bert terHart

00:27:59 - 00:28:03

So if you have AI can come, come, can come meet those patients where they are. I think that's a good thing.

Donna Mitchell

00:28:04 - 00:28:06

Well, I'm very sorry for the loss of your dad.

Bert terHart

00:28:06 - 00:28:07

Yeah.

Donna Mitchell

00:28:07 - 00:28:12

Your dad had a very positive impact on your life from the conversation that we had this evening.

Bert terHart

00:28:12 - 00:28:12

Yeah.

Donna Mitchell

00:28:12 - 00:28:31

And it seems like you have really got your hands in a lot of important healthcare and just so many different places with the AI. I guess my next question is that fits a lot of positive information about AI. But we know there's good, bad and the ugly. We won't go to the ugly, but I want to go to the bad. Okay.

Bert terHart

00:28:31 - 00:28:33

All right. So, yeah, yeah.

Donna Mitchell

00:28:33 - 00:28:40

What have you seen? What raises your eyebrow? What concerns you at night? What did you see that you wish you didn't know? Or what can you share that you feel comfortable with?

Bert terHart

00:28:41 - 00:29:28

Okay, well, there's, you know, there's two sides to every coin. And of course, there's always bad actors in the world somewhere who are looking to leverage. And this is, this isn't anything new for humanity. There's always, there's always a few of us, you know, among us, who are looking to leverage, who are looking to gain some kind of advantage. And technology has been one of the ways that humans have used to leverage advantage. And AI is a pretty good example of that. The example that I would use is that there were some drug researchers in the United States, this was within the last year, who were using AI to come up with protein configurations, basically drugs, simplistically, the way that drugs work. And I don't want to get this wrong, so I'm going to be careful.

Bert terHart

00:29:28 - 00:29:47

So you have, you. Basically, it's a piece of a puzzle. You have a cell here like this, and if you want the drug to work, it has to attach itself to the cell like that because this doesn't work. There's no interaction. Right. These are receptors. You hear about these receptors all the time. So these receptors are incredibly complicated because they tend to be protein based.

Bert terHart

00:29:47 - 00:30:28

And proteins are unbelievably complicated molecules. So that the way that these Puzzle pieces fit together is unbelievably difficult, which is one of the reasons why it's typically very, very, very expensive to create drugs that actually work. Because the number of possible solutions to this problem is a very small number. It happens to be one. And there might be literally 17 billion different ways to put this thing together. It's a little bit like Rubik's Cube here, kind of my favorite toy, Rubik's Cube. So these guys had a. They were using AI with a very sophisticated computer program to ask the questions over and over and over and over again.

Bert terHart

00:30:28 - 00:30:51

You know, what's the best way to put these things together? So you have two very. They would. They'd ask the AI say, okay, here's the problem. So this is what the receptor looks like. This is the piece of the puzzle we have to fit. And then they would ask it, I need to make a protein that fits that. And it would then run through the 29 gazillion combinations and come up with the ones that worked that actually made things better. Okay, so that was fantastic.

Bert terHart

00:30:51 - 00:31:13

It was amazing. They were. They were coming up with these drug combinations and protein combinations that were unbelievably beneficial. And they could do it in a way that was. Was firstly, incredibly inexpensive compared to what it was done before and was wildly efficient. And then they decided, hmm, I wonder how many ways we can make something that's unbelievably deadly. Like, how many. How many ways? Let's just.

Bert terHart

00:31:13 - 00:31:39

Instead of. And it was just. It was a. It was a simple switch in. In computer program, in computer programming, they literally said, instead of optimizing for 1, which is the best possible benefit, let's optimize for 0, which is the least possible benefit. So they just hit a switch in the code, exactly the same code that said, run towards 1 or run towards 0, and then took it. Well, this is interesting experiment. Scientists tend to be curious people, and, well, this was wildly.

Bert terHart

00:31:39 - 00:32:05

It was fantastic. This way. So let's see what happens when we run the other way. So they turned the machine on, and then they went home and they came back the next day. And the result was that they buried what they had done, stopped the program, and went to the White House and said, we found thousands of drugs that are capable of killing an extraordinary number of people. And anyone can do this with this code. So we can't. This.

Bert terHart

00:32:05 - 00:32:50

In case no one's ever thought of it, we just did, and maybe we should talk about it. So this idea that you can basically run towards one way very quickly, which is what we've been talking about, right? Like imagine, imagine you put a bug in a. And this happened. Actually this happened with Boeing. Boeing had a, had a problem with their autopilots where they had a very simple problem in the code that, that somehow just because there was this, you know, convoluted series of events where the plane thought that it was actually in a very steep climb, it just, instead of going up, which is one they thought it was going, you know, or, you know, one going up, zero going down. So it thought that it was going towards one wildly. So then the autopilot says, well, we bet we need to go down. And that's what they did.

Bert terHart

00:32:50 - 00:33:47

So the plane just went into this mode where it thought it was going up and suddenly the plane now is going straight down towards the ground. And you don't have much time to figure that out, right? So this idea that, like I said, you can be driving towards something good, but because it's code, it's very easy to say, instead of optimize for 1, let's optimize for 0. And the results, this is, it was public. I mean, this wasn't anything secret, but their results were public because it was part of some published research. And the flip side of that coin was unbelievably dangerous because they discovered very quickly that there was a bunch of different ways to put these proteins together that were basically deadly and anyone could do it. Not saying that anyone could actually make the proteins and do all the work associated with making whatever it is they want to make, but the actual nuts and bolts of it were just as, you know, as plain as day. So, yeah, there's, there's some things to be concerned about. And I mean, that's, that's the issue.

Bert terHart

00:33:47 - 00:34:04

Whenever you hear issues about AI and people saying, oh, well, you know, there's got to be careful because there's, you know, there's good and bad here. We can't just be running towards the good and thinking or running towards one, this one direction and thinking. No one else is going to be looking over the shoulder. That's just. Humans have never been that good natured. Let, let me say I could talk.

Donna Mitchell

00:34:04 - 00:34:20

To you for quite some time now. So I'm like, I better get ready to close and just have you back. So what is it that you haven't had a chance to share that you really wanted to talk about on pivoting the Web3 podcast? Did we miss something that's really near and dear to your heart or you.

Bert terHart

00:34:20 - 00:34:51

Want to really say, I think that. I think what I would like to say is that there's no reason to embrace these, these, these technologies because they're just gonna, you know, it's like they're coming whether you like it or not. The Internet is coming whether you like it or not. So there's no reason not. There's no reason not to embrace them, and there's no reason. And if you do, then, then you can decide for yourself, you know, what, what it is that you like and what it is that you don't like and, and how you might want to use it and how you might not want to use it. I think that's an important. I think that's an important decision for everyone to make.

Bert terHart

00:34:51 - 00:35:12

And to make a really good decision for yourself, then you need to, you know, you need to get right into it. So I would, I would encourage everybody to, to, you know, to wade into this thing. And it's not hard. I mean, it really isn't hard. It's designed to be for, for everyone. That's the whole purpose of it is to. Is to help make your life better. So I would, I would hope that, you know, people would, Would bite off as much as they.

Bert terHart

00:35:12 - 00:35:13

As they might want to chew.

Donna Mitchell

00:35:14 - 00:35:19

And if someone wanted to reach out to you, what would you like to share? What's the best way?

Bert terHart

00:35:19 - 00:35:37

Well, you can, you can just go to bert.com and find me there. So, I mean, that's the easiest way. And I have a rule. If you email me, I, I will email you back. So it's easy enough to find my email on, on that place. And if you Google my name as you can, I think you might see it on this. You're going to find way too much stuff. I think so, Yeah.

Bert terHart

00:35:37 - 00:35:59

I, I'm easy to get hold to. I, I'm easy to find. And I, I mean, if you have the, if you have the courage to reach out, then I, you know, I'm. I'm more than happy to, to speak with you because I think that's. I think that's a courageous thing to do, and I want to make sure that anyone who does that is. I want to reinforce that sort of behavior. If you, if you choose to be. If you choose to be a little bit outside the box, I think that's a good way to live your life.

Donna Mitchell

00:35:59 - 00:36:02

Everybody listening to me. You're a little bit outside the box.

Bert terHart

00:36:02 - 00:36:02

Okay.

Donna Mitchell

00:36:04 - 00:36:18

I'm a little bit outside the box. I know that. I really, I'm really glad to have you here. I've learned so much you gave additional insight. And I really am interested in what you're doing and how you're doing. We're going to have you back again if you're willing to come.

Bert terHart

00:36:18 - 00:36:19

Yeah, of course.

Donna Mitchell

00:36:19 - 00:36:33

And do a deeper dive in certain areas or maybe explore some new ones. But thank you for listening to pivoting the Web3 podcast. This is Donna Mitchell and Bert Tearhart. And thank you. Thank you. Thank you. Have a good evening. And we're here to serve.

Donna Mitchell

00:36:33 - 00:36:45

0:00 - 36:45

Thanks for checking in the Pivoting the Web 3 podcast. Go to pivotingtheweb3podcast.com to download and listen or web3game plan to check out the videos. Thank you. We're shaping tomorrow together.