This episode of Real Talk from AI and I Care is brought to you by Barty Software. Welcome to Real Talk on AI and I Care with your host Dr. Scott Morris. And I'm co-host Dr.
Ray Han Ahmed. In each episode we discuss current news adding a little AI education, and debate innovative AI tools and topics that will change the industry. Hey Rayan, so I got some really two interesting points in news of the day. First of all, I don't know if you've got a chance to use Gemini 2.5.
That came out a couple days ago. There's 2.5 pro. And it is, I literally, I think I have one of the previous podcasts I said I kind of use all these different large language models when I'm doing content generation and all that and I'll tell you what. I was wrong.
I it all changed when that came out I was like, I don't need any of the large language models now, at least for the time being. So you know the general. So great. What's so great about.
Yeah, what's the great. So it is it's the first one that actually takes. Transformer model, large language mode, which is everything we've been talking about so far. It's clawed.
It's you know, it's chat GPT. It's transformative models. And now what it does is adds chain of thought to it. And it's the most fascinating thing like you can actually do it.
And it's free. There's the coolest thing is it's free. You don't have to buy it. It's not a membership, at least at the time being.
It's not a membership or something like that. And you plug it in and you go watch it think. And it will literally go through how it thinks about a process. So if you struggled with how to develop prompts before and how to get to the right answer.
You can watch the thinking go, oh, well, that's where it went wrong. I need to change the way I asked this. And then you edit it. And you can watch the chain of thought.
You can watch it change. Change the thought. Now, that's part one's pretty cool. So watching it how it thinks, which used to be a big black box.
And I think later today we were talking about hype pros and cons of, you know, the hype of AI. And one of the always the cons was, well, it's a big black box. Well, now we get to see the black box. And that's pretty cool.
The other thing is that it can look, it can handle 1500 pages. Almost a couple of was it. 100 million characters before it starts to forget that is 50 times more than any other model in the past, which means you can plug it. I mean, 1500 pages.
You can plug in every book known to man, you know, and there's no book that's that long plug it in and it can do the whole thing. Right. So it changes. It's got so much more context in which to be able to do.
And the other thing you'll think it's got some really, you know, Gemini did this really well and great coding features. And, you know, I think about this and I don't, I'm not really like I'm a coder for my AI project, but I don't really think like I'm an app developer that's not who I am. I'm not a software developer. But I started to think I might become one because literally I think with this and definitely as we get into the next version of it, we're all going to be able to create our own little apps for whatever we want to create.
We're not going to need a software background. We're not going to need special coding software. We're going to tell them this is what we want an app. This is what we want to have to look like.
Go build it. And it's going to build it for us. So I have not, you know, so full disclosure, I may pay the user of chat GPT type pay the 20 bucks a month. And, you know, I've said before I like to use cloud and chat GPT and Gemini, but I've sort of just been using chat GPT for a long time.
I feel like I think I said this before I feel like it kind of knows me like a friend now. And I use it so much and I feel like it has a stickiness. So maybe, maybe I'll check out Gemini. Are you an iPhone user or are you a droid user?
So I am an iPhone user. Yes, I'm a droid user, right? So we there's that stickiness, right? I mean, my family is always like, you got to switch and I'm like, I can't switch.
I like, it knows everything about me. Like I understand how it'll work. So I do think, you know, we're all going to find our large language models that get us. And as we start moving more and more, we talked about this.
I think one of the other podcasts about agents. I think when the agents start to get us, it's going to be really hard to leave that platform. Because they get us, right? They're already know our idiosyncrasies and what we want.
Yeah, I know. I don't know. Can I is there a way to export all the conversations you've ever had with chat GPT and put it in Gemini? I don't think they're allowed.
They're probably going to be support. Yeah, I got to believe in that. I'm going to do that. Well, I'm going to my second point of the day.
I don't know if you saw this. So Zuckerberg, you know, owns Meta and we're all kind of familiar that Meta has got. You know, they're teamed up with Rayban and they have their Meta glass and all kinds of stuff. And then, you know, he's been talking a couple months ago.
He's still talking about supernova, which is kind of that new is Oakley instead of Rayban. And it's really more for like bike riders and runners. It's more like outdoor enthusiast kind of stuff. It's got cameras and speakers and all that kind of stuff.
Like I think that's going to be really cool. But I think it's going to be for skiing. I live in Colorado and I can just see myself with this behind us. Skigago going, this is awesome.
I guess I know where I should be turning instead of waiting to get the bottom and go. Oops wrong, wrong avenue. But then he started talking this week because you know, he's been saying for a long time that he thinks cell phones are. Smart phones are almost over and I kind of agree.
And you know, I've had that conversation. Rayon is that. I think that my kids who are in their 20s now will kind of remember it. Like I remember a record player and eight track paper.
They track tape player. Your kids are much younger. I don't even know if they're going to remember it. Like I don't know if they're going to remember it.
Like I don't know if they're going to remember smartphones. And you know, a lot of people are starting to say, hey, me, you that's it. And so now he's talking about his next generation. Which he started talking about this week called hypernova.
And that is. I almost a little bit of Google. I asked like it's going to have a little bit of augmented reality in them. I think it's the camera.
It's the right lens of the left lens. A little bit of augmented reality. But you know, you can get app updates. You can see notifications.
Now instead of having it on your watch or on your phone. Now it's head. It's a big thing. I don't know.
I don't know. I kind of wonder if the future isn't all about my care. Yeah. I'll plug my.
The pinnip piece I did for AI and I care last month and it was on this topic. And. And you know, I'm double minded about this whole topic on one hand. I totally get it.
I see it as a phone replacement. But on the other hand, to have yet a piece of technology that further. Disassociates you from the person ahead of you. And it's a frame on your head.
That's in between you and the other person you're talking to. So like when my brother wears his Raven meta glasses. I'm most kind of looking at the side is like, is your light turned on like what? Are you taking a picture like why are you recording?
I'm kind of like as much as I love the technology. I don't know if I want to be recorded all the time. And if that's if that's really, you know, and also there's a fact. And look, let's not underestimate the fashion element.
So. I think this is an art generation that feels that way. No, I don't know. I'm definitely not betting against it.
But I just I think there are some some big issues here that need to be carefully thought. Through. And so maybe I'm double minded about it. But what I'm not double minded about is that I care providers need to be on this conversation.
Like we need to be front and set or either helping develop this answer to have that. But also informing the public on it. I I completely agree. I think I want to do one of these.
I think we need to do one of our, you know, our real talk, long segments that that we're going to get towards the end of today. I think we need to generate one of them towards just this subject. Like this is. This is such a cool subject about what is the future of heads up display in the world of I care.
How does it impact what we do? So. Anyway, all right. Well, I'm interesting.
You know, I think it's probably that's probably got for the news of the day. Right. And what were you going to say? Well, I always try it.
Yeah. Try to give them a little bit of something to think about in terms of improving your education about the world of AI. What do you got for us this week? So tonight, I want to zoom in on one of the biggest success stories in our field, which is AI and diabetic retinopathy screen.
So as we know, diabetic or not, the is a huge cause of preventable blindness in adults in America and throughout the world. And I didn't know. I didn't know this many, but 50% in some studies, they say 60% of people with diabetes don't get their annual retinal eye exam. That means millions of people are at risk for silent and reversible vision loss.
And we see these patients all day. You know, they come in and they have a picture of hemorrhage or they have pdr and they need lasers or injections. So this is kind of an obvious point where AI can actually save sight kind of right now. So I just want to review what's available in the marketplace for the history.
And then I'm really looking forward to doing this as a lead in the pro concept we're going to talk about. So just sure. Just by way of that. There are three FDA cleared AI systems that that autonomously meaning on their own screen for diabetic, I diabetic retinopathy.
And this means that there's no doctor required other than making the initial order. And so the way it works, the patient comes into a clinic, a primary care doctor or even a retail clinic or whoever. Maybe it's a Walmart, Walgreens, CVS, maybe it's a lot for. And they get a screening and there's three devices that are well, there's three companies that have that are approved.
There's a digital diagnostics. There's I knock and the newest players AI health, a, a, EY, help. And so let me break them down so I digital diagnostics was called IDXDR. That was the first one.
That was back in 2018. So seven, seven years ago. And this was the first FDA cleared autonomous device, not just an eye care, but in all health care. So it takes two pictures using a top on and W400 and it gives an answer either refer or not refer meaning it's more than mild diabetic or an apathy.
And next one was I art by eye knock in 2020. This one was a little different. It just needed one image per eye and it did both diabetic or apathy and also macro diabetic macro. The last one AI health, they are the most recent one they got cleared in 2023.
It's one image per eye. They are sort of thought to be the fastest in terms of like less than 60 seconds, but they're all sort of very, very similar. This also uses the top on this, the AI health also just kind of approved for a portable sort of a handheld retinal camera, which is pretty neat. I think that's pretty cool.
I saw that the other day. Now you get a wonder, I mean, how does that affect how we do remote care right now if you don't have to have a big huge camera in your office. That's pretty interesting. Yeah, I've used the app so it was approved on the optimist.
Maybe it's with me and maybe my texts are way better than me at taking pictures. I said a hard time using the optimum head and handle camera. That was just me. I mean, no, no financial interest here, but I said I had a hard time using it.
But I'm excited about the handheld cases. I think it's really good. I'm also excited by just having a very simple device, even if it's a desktop device. I think we need both.
The nice thing about all of these devices that there's a there's a reimbursable code for this, which makes it, which makes it nice. And that code is like the nine triple two nine code and reimburses. You know, it's gone down since it started. I think it started on like $65, $60.
Now it's I think sub $40 or $40. Not terrible, but, but, you know, we don't get reimburs much more than that for a rental photo. I know. I know.
It's, is it really that much different? Probably not. And probably not. And the key thing is, and is the heaviest sort of the quality metrics that you can be helped with this.
So a lot of health plans like Medicare, damage plans. They sort of live and die by these metrics. So if you're on these plans, if you have a higher star rating, you will get bonus payments. $10,000 per patient per year.
So that's another big win for these types of systems. This episode is brought to you by Barty, the AI powered all in one EHR built for I care with AI scribe voice over internet phones, websites, payment processing in R.C.M. all bundled into one system. Barty is redefining modern I care spend less time clicking and more time with patients at Barty.com.
Great. So like, okay. So what's the catch? So the catch really is, I think these are great, all three companies, great companies and the companies that are coming out.
It's a kind of proud space. They've attracted a ton of capital investment. Really cool technologies. Really great team.
But adoption has been slower than probably what I would have guessed back seven years ago. And we know that because that CPT code really hasn't induced that much. And we know that we're talking about these technologies out there that we're talking like thousands of times, few thousands of times. So it's kind of tiny in comparison to the diabetic eye disease population that we know.
And it's mostly being used as academic medical center. So a lot of, you know, maybe the sort of the Tesla impact things happen. And we're not going to be able to get that out into the community, but we're still sort of waiting for that. And, you know, why is that happening?
I think there's multiple reasons. It's probably workflow. Maybe it's reimbursement. Maybe they can also be part of you.
Yeah. I think it might be capital expenditure, right? I mean, this is an expensive device. And, you know, do you buy a new device every time a new device comes out?
No, you're kind of like, okay, let me use it long enough to pay for it. And it starts to be maybe not the newest thing before I go buy the new one. And that's kind of, you know, you and I probably need to do one of these two about what is the, what is, you know, we kind of covered this a little bit last time was what is the practice the future kind of look like and, you know, maybe it's not every one of us having a device in our office. And, you know, we can build it.
We can all be excited. And I know this kind of a lead in maybe into our next topic about the hype of AI is that, you know, we don't get excited about it and people can invest a lot of money and then they can build it. But if you build it, will they come right to feel the dreams comment. And it ties in really well into what we're talking about next, which is that we're going to be doing a little pro con segment on whether AI and I care is is over hyped in its current form or not.
And Scott, why don't you want to tell me why AI is an I care. Or maybe we can talk about more generally, but we're sort of focused on I care, but why is it not over having. You know, I think we can probably both fight either side of this and go, I think that you just said it right we already know that some of that technology exists. I mean, so diabetic right now to the exist, we look at what it's doing radiology, what is doing in pathology, what it's doing already in drug discovery in terms of, I mean, you know, the reality, regardless of what side you're on.
In terms of vaccinations and that kind of stuff, and we look what we did with the COVID vaccine, I mean, what should have taken years took months and now they're saying, you know, we can develop vaccines in six months that you think for five years, that's because of AI can look at and just pound out models and do experimentation. I mean, I think the more appropriate stuff is look at what it's doing in terms of task automation for marketing for contact management systems for finance. I mean, oh my gosh, I dropped my entire practice spreadsheet into Gemini this morning, said I want to look at this and then 30 seconds, it gave me a report that I know that I couldn't build if I worked on it for two weeks and it did it 30 seconds. I mean, that's AI right maybe it's not always clinical AI, but all we do is not clinical, we run businesses, we we deal with patient care, we deal with education.
I think that you know, it's starting to get to the spot where is we get more data and I know you're going to pick me apart right here because I would pick myself apart on this one is, you know, I mean, it's going to be able to start looking at pattern recognition of these huge data sets. If we have access to huge data sets to be able to do that, I think that you know, we already know those successes exist, maybe not yet in I care, but when does I care, been first about adopting anything, you know, so I think it's going to take time I do, but I'm encouraged, I, you know, I. Well, you know, I definitely could probably argue both sides and it's kind of awkward for me to argue that the coincide or I saying that it is overhyped and I maybe I don't believe everything I'm saying, but if I were to make a sort of strong man argument, I would be I would just say about the dog, but we're off these stuff I started by saying this was back in 2018 it was approved and we're talking seven years later and yet yet eight years, you know, seven eight years later. The CPT code has only been used tiny fraction of times and do we have a same code for AI and glaucoma AI and macadameteration AI and.
We're now honest, our honest ROP lens counts like all these other things and make and all I hear is it's coming it's coming yet soon soon it will be there and so. You know that to me if that's not a hype cycle, I don't know what is I think the the future tense that we're talking about is. Is something as a field we have to start delivering on things so it's not kind of like it's going to be plug and play but it's kind of like plug and pop we're starting to kind of like waiting for things now but okay so let me definitely devil's advocate on that man so let's think about it so I remember I was a resident. And I remember I got the first.
Blackberry. I actually come from Miami and I got a blackberry like the week before I left and I went into my residency and people like what is that. That was a blackberry was 20 years ago right and now you look at a blackberry or flip phone and go you've got to be kidding right I mean really seriously it's evolution right and I think we're still very much in the early part of evolution and this is a time scale thing right if we look. In terms of decades and not days and weeks I still think we're all babies still crawling right we're still infants maybe even toddlers in some categories but I look at how fast it's changing if you look at just chat gpt from two years ago to what as we said earlier today what chat what Gemini 2.5 is.
You're talking the difference between record players and you know iPods right it's just such a huge difference and I think it's going to go faster and faster and faster I mean it has to I believe kind of getting a point number two is. You know I mean there's so many healthcare challenges I mean if I I did a study before I started my company and said what are all the healthcare challenges and I come up with 196 unique differential I healthcare challenges in the world of healthcare United States 196. There's a lot of things we can fix maybe we're not going to do diagnostic imaging maybe that's maybe that's the first place we dove into but that might be the last thing that we actually really fully influencing it or arms around the evolution and the path in which we take I don't think any of us know right we're going to look at retrospectively go of course it went that way right now we're like I don't know where it's going to go. Yeah I think I think you made a good point I we overestimates what's going to happen in two years two or three years and under estimate what's going to take 10 to 20 years I mean you mentioned the blackberry example I to think AI is still going to be we're still going to be talking about diagnostics and diabetic from up to 20 years.
No, and I think diagnostic you're going to look like they look now yeah yeah I totally I totally agree but I think you know as as as people who are you know listening to this podcast and thinking about this you know we have to you know we we can't automate the trust we have to earn it with our stakeholders and that stakeholders meaning other I care providers out there who may not be as you know reading everything about AI and so. It's a lot about you know patient has surgery you come back and you know maybe they're in the healing phase and you just have to kind of sit sit with them and say you know patients patients it's going to get we're going to get there we're going to get there and so I think that's that's the way we can help lead this field you know and the hype is definitely real I think you would agree with that too and we just have to manage it. Yeah, I you know I I think maybe I'll give you some fodder here right is that I mean I as I talk about this people like but I don't trust it I mean I don't I can't see how it's thinking I don't trust it well Gemini just change that but then I always pose a question of them I said well. Aren't we a big black box now we're just in the United States with 30,000 or no I'm sorry.
In North America 81,000 I care providers between ophthalmology and optometry don't we really have 81,000 black boxes. We don't I don't know how your algorithm works when you see a patient you don't know how mine works for any given disease at least with AI we're going to build a track algorithms and build a follow train of thought whereas in humans we got no idea what's going on in the gray box instead of the black box. Oh yeah, that war with AI I see like beside your on now it's like the power to the robots but that's great I think that was a great pro con debate and I look forward to do more these. This is super good hopefully everybody that listens you know we were able to give you some important things to think about definitely reach out to us you can get us at Scott as a C O T.
At AI in I care dot AI or Rayhan our E H a N at AI and I care dot AI and give us ideas of topics you want us to discuss we have a long list but we're always looking for top of the mind types of things first to chat about we love to hear from you. You've been listening to real talk an AI and I care your weekly podcast to keep you informed about AI technologies revolutionizing I care. We would like to thank birdie software for supporting this episode of real talk for may I and I care.