cataractcoat.com podcast series, episode number 144 with Dr. Rohan Ahmed, how artificial intelligence is changing ophthalmology and what you need to do to be involved. Welcome back to our cataract coach podcast. Today, I'm a very special guest and a very special topic.
And that topic is all about AI in ophthalmology, artificial intelligence. What do you need to know what will be coming down the road and how do we prepare for it? And for that, I have a man who has his own podcast about AI, Dr. Rohan Ahmed.
Welcome back to my podcast, my friend. Thank you for having me today. This is a great treat, a great honor. I am a long time fan, a long time listener.
I think I'm following the top, one of the first few cataract coach podcasts subscribers back in 2018. So I think you're doing great stuff and thank you so much for having me. Well, we first actually met probably 15 plus years ago when you were a resident. That's right.
That's right. You did a grand rounds at Baylor, why did my brother's at Baylor in Houston? And we had hundreds of grand rounds over the course of three or three years. I, and I could say for my class, we distinctly remember yours as being one of the best.
And it was because you brought so much energy and you talked about, your passion was obviously infectious. We were all super excited about how to have a lunch at afterwards. Like this is amazing. Because we, you gave us a lesson.
I still remember. You know what? Orthopedists do spine, neurosurgeons do spine. Other people do spine.
And there's so much, there's like a petition, there's like other things that you have to deal with that. You know who only does ophthalmology and eye surgery? Ophthalmologists, you know, it is, it is our domain. And I just loved it.
Your dedication to excellence, I love that grand rounds and I really appreciate you. You actually spent, you did the grand rounds. I remember this. And then you spent an extra hour with the residents.
Wasn't scheduled, wasn't planning just like, that's just chat about careers. Right. And it was great. Thank you so much for that.
That was great. Well, I mean, that's the goal is to have some kind of impact. And obviously, to spread this incredible enthusiasm because yeah, I'm thankful every day that I'm an optimal. I'm sorry I get this lucky in my life.
I mean, it's just amazing. The only, the only time I think, okay, wait a minute, is what I'm on an airplane and I travel a lot. And then like, is there a doctor on board? And I just think of myself, what do you mean to check the pupils?
That's right. That's right. I know you've since a Dr. Glockam-Flecken, his comedy stuff and he has a good video on that about one of the greatest group of ophthalmologists is being called on an emergency on a plane.
So yeah, I kind of slouch down in my chair too. So I'm gonna tell anyone that. That tell me more about your kind of career path. So you practiced there in the Dallas, greater Dallas area.
Yes. And what did you do after residency? Cause you did a little bit, few things that are unusual. Yeah.
I will say, so I was a big nerd in high school. Like it was 1999, it was a year wide decay and I was graduating from high school that year. But I was on a computer programming artificial neural networks. I was taking a class at a local university on the mind body problem.
And I was working on a project putting in, getting a robot eyes vision. And it was, it had a mechanic arm playing tic-tac-toe and my responsibility was programming the neural network to identify exes and ores. And that, that got me interested in artificial intelligence. Got me interested in the eye and ophthalmology and therefore I went to med school.
But I tell, I tell my mom, because you know, I'm also a basic background and she's like, you have to go to medical school. So I, you know, I'm a very good, I'm a good son. So I went, but I was like, no, what have I done AI back then? You know, it's been a totally different career cause I was interested in it.
That got me interested in ophthalmology and then did residency, got into private practice and I was just been so curious about this stuff for such a long time. Like you, that when it started coming out, I thought I'd just, this is gonna be the most impactful, way more than the internet. Most impactful new technology we're gonna face but I just wanna get into it as deeply as I could. So well, very precocious in 1999, 26 years ago.
So kind of corresponds with, you know, my co-resident, I, John Laddis, who was a co-resident. We were residents from 97 or 2000 and a 97 is when IBM's deep blue beat Casper, Robert chess. And we were just like, whoa. Yeah, yeah.
And that's kind of in that point too, maybe a year or two later, I grew up at the same time as you. Yeah, start thinking more about like, how do you incorporate this into what we're doing? Yeah. And then we get an interesting like, okay, we're in interest in the Iowa Calc.
It's like our website, Iwellcalc.com. If you look up the URL, I registered that URL July 1 of 2000. That's 26 years ago. Yeah.
Or 25 years ago. So it's just like, it's crazy. Yeah. And now it's kind of like everyone's like, of course you'd use AI for this.
But back then, it was, I mean, that wasn't even on the mat, right? Even 10 years ago, it was not obvious at all. It was not obvious, it was not obvious. And I remember, I went to Med School at Vanderbilt and I remembered doing a project in diabetic retinopathy screening saying, you know, we looked at that.
The technology was just not mature at that point to do autonomous or automated analyses of a fundist photographs. But there's been so much technology upgrades in the past 26 years that, you know, it's amazing what one person can do now. I mean, they're talking about the solo entra who's starting a billion dollar company. And this is possible.
I mean, there's so many tools that are disposed on. So, you know, my clarion, my, the thing I'm really preaching now to every out the mods that I'm using, you just got to get into this stuff. And I've seen a lot of trepidation, but I think it's just, it's fantastic. And I know you use a ton of tools.
I'm excited to talk about that. Right. And they just keep evolving. I mean, the number of tools available you have for free.
For free, for free. Well, I'll give them your data and they track everything about it. That's right. That's what I mean.
You told me, I think there's a lot of parallels with ophthalmology because we are a kind of a tech forward specialty, right? Lasers, Iowa calculations, new types of lenses. When I talk to my primary care buddies, I'm like, what do you like with your new, and it's like, oh, like, I'm kind of, again, it's amazing thing about ophthalmology. We just get so many, we're kind of spoiled.
I mean, don't you think about how many tools and gadget could we get to use? It's like, I mean, in your working, I mean, we haven't talked about horizon and the robotic stuff, but it's just fantastic. I think that gives us a like up and being early adopters of technology. We're just so curious.
I mean, I think maybe it's a selection bias. It's a huge selection bias. So it's like, we're all geeks. Every ophthalm is an absolute totally geek, right?
So it's funny how we can pretend now to be cool, but let's be straight. I was a total geek in high school. I was a nerd in college, and I was a dork in med school. I mean, come on, of course.
It's funny. When I had residents out, I would query them like, so Rehan, you and I were both high school valetron. Of course we were. But it was like 50% of my residents were high school valetron.
50. And this is back in the day when there was one out of class, like 5, 6, 700, not like now where they're like 12 of them or 15 of them. It's like it's complete. It's complete selection bias.
But yeah, we're totally geeks and nerds, and we love this stuff. And I embrace it. And I look at the world of mathematical way off very often. And now it's just so natural to go into the AI stuff.
It's just the next time. Yeah. And you know, the other thing with ophthalmology, where we have an advantage, I mean, this is where the IowCELTS is the amount of data we have. And it's not just textual, textual data or numerical data, which is what I'm all about.
Oh, CalXR, but there's so much imaging data. And that is so rich with potentially developing data sets. So, and we're like a primary care type of specialty. So we just get it on our patients.
We don't have to call radiologists. We just get an OCT, get a topography, do fund this photography. And so we have access to so many, so much data and therefore so many tools that we can use to help patients. Right.
I mean, it's gotten to the point where certain things are certainly better than humans. Like I example, give an IowCEL calculations is that, yeah, a human, I can optimize a single A constant for an IowL. But that doesn't make any sense because how would you treat a 28-dipter, hyperopics small I, the same as a 10-dipter, a very large map, you cannot. But you know, we can use an AI version to calculate, all right, sort of optimizing just one A class.
We can do 10 ranges of axial length, 10 ranges of AC depth, 10 ranges of K values. Well, that's a 10 by 10 by 10 or a thousand point matrix. I can offer a thousand points in space instead of just one A-cuff. And then the other thing, now if you go to Image and you said, crazy part is, you know, I can look at a fund this photo and say, this is probably a young person, like 20 years old, and this is probably an 80 year old.
Well, you have a year or two, which is crazy. And AI now could look at a red name where you tell you, is this a male or female? And I have no idea how to do that. Exactly.
It was crazy. It's absolutely, and that's a whole field of oculomics, right? So, oculomics for your listeners who may not, and I'm sure most real listeners know, it's the field of ophthalmology or of eye care where you're looking at, so it doesn't have to be a fundist photograph, it could be other things. But typically it's a retina photograph and you're diagnosing other non-extraocular issues, like cardiac disease or neurologic disease.
And it really started based off, you know, early, late 2015's, lab and Google that started doing the sex prediction based on a fundist photograph. We still have no idea how it actually figures it out. And some people are just like, well, maybe it's the diameter of the vessels, maybe it's some type of pigment in the background, but we're like, I still have no idea, you know. But a computer with 99% accuracy figures it out.
It could talk about cardiac prediction, Alzheimer's prediction, kidney disease. So think about all the things you can now do with a retinal photograph. And it's just fantastic. So that's, you know, we sort of like, when people talk about oculums, I'm like, you know, in med school, they talked about the eye as a sentinel disease.
I mean, that's probably like, there's so many health things that you could do with the eye, but we're talking about superhero type abilities. Right, I mean, it's just, it is really mind-blowing. I mean, the amount of advance we've had, even in just the last five, 10 years, is just gonna be just amazing. Yeah, yeah.
I sort of think about AI in sort of three buckets generally. When I go around talking, that's the benefit of doing multiple talks, but like just to sort of like build a framework. Cause things are complicated. There's like AI as an assistant, right?
And this is where I think it's like kind of like the low hanging fruit where a lot of people, I think this is where, if you're an ophthalmologist in private practice or in academics, this is where there's a lot of tools out that could help you, scheduling assistance, scribing assistance, really sort of helpful things. Like imagine your best scribe, just having an AI version of that. Then there's AI as an agent. And this is doing your stuff that you know, you do, but it's offloading that, like diabetic and right now, they're screening, right?
I think that makes a ton of sense. I mean, you don't, there's a burden of disease out there. We don't have enough ophthalmologists, so you can eye care providers, to look at all the retina photographs that we need to do. Then there's AI as the superhuman.
And this is AI that we can't do what it's doing. And this is exactly what I was talking about, the oculomics stuff, the cardiac prediction, the neurodegenerative prediction. And especially as we get therapies to hopefully make a dent in Alzheimer's, imagine going to an eye care provider in your 60s and 70s and getting maybe an early diagnosis. And then hopefully there's something you can actually do about it.
So that's where I see sort of the three buckets of AI as it pertains to eye care. And I think that's been very helpful for me in sort of evaluating new technologies as they come in. Like which bucket am I putting this in? Right.
I think the amount of events we're going to see in the next few years is going to just keep increasing very, very rapidly. Yeah. And I think the next point of doing these super human things, I mean, like, we're using this AI now in certain parts of say, the robotic company. Yeah.
And some of the predictive things are like, wow, we shouldn't see that coming. And then even the next part too, you'll switch against the robots is, right now we use the robots to kind of emulate the way humans are doing the surgery. But what if we didn't have those restrictions? Like, what if I could just have like, okay, I'm used to making decisions here because that's really how I have.
What are you gonna do anywhere? What if you don't need to do just, you know, these same techniques we're doing now? What if it's, we make a tiny opening in the lens capsule, we have a little propeller blade that emulsifies the whole nuke and you suck it out like a milkshake and maybe that's different. Maybe fill the capsule bag with the gel polymer instead of, you know, these acrylic I-wells.
I don't know. Yeah. I mean, that's where creativity comes into play and it starts coming up at novel type of solutions. And that's like a great deep, deep blue example you gave it.
If you listened to interviews from Gary Casterall, he found that computer to be, it wasn't doing classical chess techniques. I'm kind of like a chess nerd. But he wasn't a classical chess technique. It got him because it was started doing very, it was orthogonal to normal chess.
Like it was unusual. It was create. And the way he described it was like talking about a gifted surgeon, doing things in a way that he couldn't figure out what was going on. But it was just, it was plotting out so many moves in advance.
And so can you imagine, yeah, doing, you know, maybe we've been, maybe there's a more efficient way to do it or there's a better way to do it that we haven't considered. Certainly there probably. Right. I think the neat part here, there's some fear that I've among off the mall, just whether it's AI stuff or robotics, stuff that you're going to be replaced.
I think it's an unfounded fear. I understand why you feel that way. But I don't see that happening at all. I think there's, it's just going to certainly just augment you.
If you look back at generation or two, you think back, okay, the genius of the macula, Godfather of the macula is Don Gas. So when you look at the gas atlases that he made, like 20, 30 years ago, he drew pictures that look like OCT scams because he had this crazy ability to look with a 78-doubtful lens of the macula and he had the OCT built in his brain. And I would always see his images and then I'd see the patient in my clinic and I'm like, I don't see that. So then we got an OCT machine.
It made me like, wow, now I have diagnostic skills equivalent to Don Gas. That's right. But now the next step is, how about treatment algorithms? There's some people who have incredibly bright treatment like, let's say the glaucoma.
Well, no, we should start with this, then prove with this. And if you have this much progression that you should do, there's some, they're genius glaucoma experts who have these algorithms just in their brains, in their back of their mind, in their heart, and they just know, well, I don't have that. But now with this AI thing, I can have the same treatment algorithm. And all the same scans and I can have that person's brain.
Yeah. You know, there's this kind of like maybe overused buzzword of a democrat AI is a democratization tool. And it does sort of do that. It democratizes the expertise.
Like you, like imagine a robotics surgeon that was trained off Uday Dev. Like you could bring that level of surgical expertise to patients around the world or diagnostic abilities for glaucoma or retina disease. And you know, I think that the holy grail is giving the absolute best care to every single patient. Because as you both know, you know, we're all different.
We all have strengths and weaknesses. Some, you know, some are very good at certain things and some are not as good at certain things, be it diagnostic stuff or surgical stuff. But why can't we elevate everyone to bring excellence to every single patient encounter? I'd say, I'd say this stuff will make the rest of us as good as the best of us.
You should be a buzzword, look catchy. You should trademark that too. That's good. That's a good one.
Right, they'll just increase the bar for everyone, make everyone be better for sure. Everyone, like everyone now, with an OCT machine, you have the diagnostic ability of the world's best macular specialist on gas. Yes, in your hands, exactly. And now everyone has it, doesn't think anything of it.
Right. So that's, I think that's a fantastic use case and when I talk to ophthalmologists, they're worried about being replaced. I think that's a concern. I think it's totally unfounded, especially when, like there's so many problems, the great thing about ophthalmology, again, another great thing is that we just see so many problems in clinic, from efficient, it's just like basic stuff, like surgical coordination or efficiencies and flow of clinic.
If you're in a busy clinic, it's really, can be very frustrating sometimes. And there's so many AI tools that are out there that can, you know, it hasn't all come to fruition, but it's coming and we could sort of see the riding on the wall that will definitely help our flow. Right, I think the low hanging fruit is, yeah, I did start incorporating like a scribe and like then the AI scribe, who's ophthalmology, which is ophthalmology trained. So he understands, you know, our buzzword, pseudoxfoliation doesn't come up as like a question mark, you just understand what that means.
So I think that's step one. I think the other one is with patient interactions, I mean, how much time do you spend field and phone calls for patients or what kind of simple things? When you have an AI agent kind of go through the tree with you. And so I think it's, I think we're so busy.
And my point of we're using a couple of tools, some of them are third party, some of them we've developed ourselves. One is a tool that does pre-authorizations for every single patient. It will, it's an AI voice bot that will call the insurance companies to get a verification of benefits before they come in. Wow.
You can, you know, dealing with the insurance is a big pain point and we have a stat, you know, there's a big staff to like deal with that. You're paying them and they could, not that you'd actually replace them. And that's what they're all, everyone's sort of being worried about being replaced, but then you could have them do higher levels of things because the AI is not going to do anything, everything. It will actually triage to you when the AI voice bot gets, you know, it gets hung up, but it will call the insurance company and say, and at the end it was great.
It says, so, you know, miss so and so, how did you like, you know, how did you like speaking with me today? And it's, you know, most I'm like, it was fine, you know, but it's great. It was, it's a fantastic tool, improves our efficiency, increases our, you know, our reimbursement rates. And so that's a great tool that we're using.
Yeah, for sure. I mean, we're going to have more and more of these. Also, you can develop apps now with AI. It's just so easy to do any of this stuff.
Yeah, that's right. The problem that I don't have enough hours in the week to play with all this stuff, like I was playing with Nano Banana, which is Google's AI image. Yeah, yeah, I know exactly. I just, I spent hours like, oh my God, what am I doing?
I'm having too much fun. It's, it's, there's so many fun tools. I encourage it way beyond chat. You've been getting the Nano Banana as a good one.
You know, there's a lot of great image development stuff where you're like typing a prompt and just makes a beautiful image. I mean, you could use that for your own clinic to make advertisements, make videos. You know, we developed a couple of tools in the house for voice bots, for it to help our surgical coordinator. You know, I was just talking to her and I said, you know, how much time do you spend, like dealing with phone calls that, it turns out the patient or the family read the preoperative stuff, like, you know, the drops after.
Cause they'll call and say, oh, when should I start my Vigamocs? Oh, when should I do this? And it's like, well, it's on the paperwork. But why not have a voice bot free up our surgical coordinator?
So you're not dealing with those types of phone calls. And, and the funny thing, patients love it because they could call anytime and the, our voice bot surgical coordinator never gets tired of talking to them. Right. Right.
Right. Right. I would just, another one I want to kind of incorporate is, you know, the text messages. Yeah.
So you do the patient surgery and then like, you know, a few hours later, a text patient, okay, at this point, you should be experiencing this. Let me know if anything is unusual. And like, even, you know, just follow-ups. Yeah.
And I'm, I'm surprised. I'm gonna say shock, but I do surveys. I've done surveys every year to about 300 ophthalmologists where I say, you know, are you using AI tools and what are you using? And, you know, unfortunately it's 80% or plus have not really used anything in their clinical environment.
And I even say most of them haven't, don't even use JFQPT or Claude or Gemini or whatever on a, on a routine basis. I mean, and for me, I use that stuff every day. I drive when I'm driving to work. I sometimes instead of the radio, I have JFQPT and I'm just talking to it with various ideas I have and it's talking back to me.
So I'm just like, how are you not using it? You gotta, you gotta start using it. Right. And it's getting, it's really getting better and better.
I've often used, you know, one of these LLMs like Chad GPT to help me write an article. But I always have to go back and reread it because it's like quite there. It's just not quite there. Yeah.
It's, it's getting better. I mean, we don't think it's getting better though. I mean, I find that it's not quite there. And, but the prompting makes such an importance.
And the more you use it, the more you upload from the documents you've written in the past, you can really match your tone in a way that's pretty darn good. It's not gonna be perfect. And that's where your expertise comes in. And that's where you went to always be needed.
But I will say I'm getting increasing. I, every new generation, I'm like, oh wow, Chad GPT 5.1. It is, I see it's, it is getting pretty good. Right.
Like in the earlier ones, we're kind of easily fooled. Really easily fooled. Oh yeah. But now it's, I think it's a lot, lot better.
I mean, I remember when it first, we first had access to Chad GPT, I saw some examples where I tried. Some were like, how many states have the letter R in their names? And it'd make up some random number. And I'm like, listen.
And they'd listen and it'd be like Iowa, Texas. I'm like, wait a minute, those are wrong. Yeah, yeah, yeah. Oh, my party is correct.
Oh, you're correct. Okay, we'll fix that. Yeah. Yeah, no, I mean, you definitely have just been basic IQA on this.
But I've been, it's just, I'm increasingly impressed with how much better and how fast, how much faster it's getting and responding. And, you know, maybe I'm sort of biased, but yeah. I mean, as long as you use some basic QA on it, I think you should be, you know, you should be fine. I do say like when I read other people's posts or papers, I'm like, you use Chad GPT, didn't you?
And so I think now the weird thing is like, we're entering this world where if someone has a typo in their email, I'm like, oh, wow, this is actually like, they actually wrote it themselves. They didn't use Chad GPT. Like I love it. They're like a sense of humanity in there because there's like errors in there.
And, you know, it's not exactly formatted, right, et cetera. And so now, you know, if I send out an email, maybe I'll put it like an extra period in there, just like make a little error. So if people will know that, you know, I did, there was some human oversight or some human involved in the process because imperfection, you know, is the type of beauty in itself. Right.
Yeah, for sure. Why be savvy, right? Japanese kind of a beauty of imperfection. That's right.
There you go. Yeah, that's good. So no, I think that's certainly important. The problem is I sent an email now.
I'm just wondering, Google already just summarized it for me in the AI. I'm like, do I even read the whole email? There's another question. Now the other problem I think we're gonna get to is someone estimated at this point, all new content on the internet, like a larger percentage of it, maybe even approaching half as AI generated.
Yeah. Well, don't get to the point then, where just AI training itself on its own product. Yeah. Yeah.
We want to create the AI on our human created content. Not to just train it on itself over and over again. But now we're getting to that point where maybe it's getting trained on stuff that's just not right, that it created itself. Yeah, and it turns into this cycle where it's just and it's spewing out more AI based content that's trained on itself and who knows what.
So I think human ingenuity, human creativity is always gonna play a role, right? You're gonna always have to have that to train these models. Like people call it, I think it's not like a dead internet theory, but most of the traffic online is actually bots. It's not actually even humans.
It's like 80% bots rolling around. And so yeah, I think this is a real problem. I don't know if anyone has the answer yet. And so I think, you know, what are like, and so I have young kids.
And so sometimes I think like, what, talk to my wife about this all the time, like what should we, what skills should we be teaching our kids in school and beyond to be living in this AI world, right? Cause we're kind of blessed in their sense that like, we're in this transition period where I think we're gonna be fine because they're gonna need us, but they may live in a world where who knows, right? And so I think one of the things that I like doing is I think daydreaming is gonna be the thing. Yeah, yeah, yeah, created and specifically connecting unrelated, totally unrelated things.
Things like sitting under a tree and an apple hitting his head coming up with gravity, right? Like, hey, I could never do that because it's so unrelated. It doesn't make any sense, but that's what I think we should be talking to our kids about and really having like, I think the best thing is just to do nothing sometimes and just let your mind wander in ways so you can make these connections. I don't know, am I like smoking something?
Is that what would be? No, no, that's just fantastic. Because if you think back, I'm about a decade older than you. When I was in school, they made us learn how to write in cursive.
Because I was looking at you guys, because back in the old days of a quill, you didn't wanna lift your pen, you wanted to write the whole word without lifting your pen. That's cursive, okay. Well, now my kids really didn't learn cursive in school. It's kinda useless, it's the point.
Well, then similar, like, you know, we spent how much time learning how to do long division. Yeah. And it's like, okay, well, you don't only have to do that anymore. If you're, it's great to finish out the concept of it, but the concept is probably more important than the execution.
Right. So just like, and I had the very strict parents going up, right, like memorize my time tables up to 20. So I still know them for 20. Wow, all right.
You know, biddattas. Because, you know, 16 times 19, you never know, they may be like, you know, you have to know that you're getting an answer within five seconds, see, I'm glad you know that. So, right. So, but I think getting with the chains the way we're doing, those kinds of skills are just not that useful.
Right. It's just, it's so yeah, it's not for inclusive. And then even this, the things that you could delegate for a computer to do, we should probably delegate. Yeah.
But the nice part is, yeah, you can now start an entire company just in your head. Yeah. So it's really, it's a different world for sure. Yeah.
It's, it has a zone set of challenges, right? Like, I think for highly creative types, stuff like chat, like, chat, like a godsend. Because I literally stay up at night. I'm like coming up with ideas after ideas, asking it to evaluate it.
And so, I try to encourage with my kids, just, you know, if you have ideas, like, I think AI is going to be a great friend to those types of people and to the people who use it. I think, you know, crawling under a rock and start trying to fight it and resist it. Obviously, there needs to be regulation. I think that's the obvious things.
It, particularly when it comes to healthcare and these types of things. But I think it, for kids and for students, this is a, I think, on net, a benefit done properly. But that's something I really, I'm really worried about. Because sometimes I think, well, maybe it is helpful to the brain to know your multiple casing tables up to your 20s.
I mean, maybe there's something going on in your brain that like, it just primes you for other types of difficult cognitive tasks. And so I, I don't know. It seems kind of like a shame to throw that out. But you're right.
Like, I think cursive writing, multiplication tables, long division, conceptually makes sense. But I don't know, I don't know if anyone has the answer. We're all just, we're all just trying to survive, brother. And we know we're just trying to figure out the best thing to do.
Yeah. I think mental stimulation is kind of the key. So on my Instagram page, my Instagram feed, one of the things I actually follow a lot of, I like essay team math questions. You know, like, wait a minute.
You're just like an old man. You too. So I get those feeds too. That's funny.
Yeah. I love the essay math questions. Cause they're all like, they're all like exponent based on this. And that's good memories.
And like, oh yeah, that's how you do the math. And so I like that because I like the mental stimulation of that. Yeah. I promise I won't sit for the SAT anytime soon.
Yeah. I did that 40 years ago. I'm good. But I think the stimulation is kind of a key.
And I think, yeah, you're right. Teaching our kids be more creative. Let your mind wander. What do you think would be interesting to you?
That's probably the most important thing. Yeah. And let me ask you, what are you telling residents now? Cause you talk to a number of residents or early surgeons earlier in their training, or maybe medical students.
Right. Do you specifically say, do you talk about specialties? Do you talk about, hey, you know, I don't know. So I don't teach residents one on one anymore, but I travel a tremendous amount and talk to young animals all the time taking countless selfies with them.
My love of these guys. Cause there's so many as you ask. I have only come my three takeaways for them are, number one, you've got to learn every single day. You've got to learn something every day.
I think that's kind of the key. And it's why I do cataract coaches a video every day. You're gonna tell them, just give me 30 seconds of you seeing my insta or maybe three minutes of you seeing on YouTube. I'll teach you something today.
Yeah. The other thing is you got to, that's room one. World two has got to embrace new technologies and new techniques. You can't kind of be afraid of that.
And part of the new technology and techniques is gonna be robotic things, gonna be AI things, et cetera. And the last one was like, yeah, you're the future of ophthalmology. You got to question everything and challenge everything. And so those are kind of my take home messages.
And that's after I give a lecture, which I'll actually do one early, early tomorrow morning for McGill. We're doing a remote for 30 my time. I'm doing a seven, 30 McGill time. Okay.
But yeah, that's the, I show the robotics stuff, I show some of our AI projects or lens calc stuff, the patterns we got. I'm not that special to be honest. I'm just super hard working. And I'm just kind of getting really kind of passionate or kind of stuck in my head and pursue things with relentless drive.
That's kind of my good skill. Yeah. But those are the take home words for the young ophthalm, yeah, you got to do teach everyday, something. Well, you say, I don't know anybody AI.
Well, today is an eco-start run. Yeah. One thing a day. And then yeah, embrace new techniques, new technologies.
And then yeah, think differently, challenge yourself, question everything. What do you tell them about specialization? I don't, I don't only mean like specialties in terms of ophthalmology, cornea right now, I'm also talking about, you have to broaden your mind a little bit outside of even ophthalmology. I mean, because there's so many problems that we deal with.
Like, you know, all these low-hanging fruit insurance verifications, everything like that. I think that's a big, you know, I don't want to say like ophthalmology where myopic in that way, no pun intended. But, you know, there's a world beyond just that. I mean, there's so many tools that you can use within AI to help with other problems you have.
So you have to sometimes like go beyond that. But what do you, what, and then what do you think about that? And what do you think about like specialties? And then specialties in general.
Like with ophthalmology. And do you think this is going to impact their selections with regard to that? Yeah, I think it will for sure. I think probably the most important thing in choosing your specialty is, what do you love?
What can you be excited about? What does not work to you? Like if you want to sit here and you want to talk about like retinovide doctoring drums, to me, that's a lot of work. I'm not going to be excited about it.
I'm just not. But you want to talk about some crazy cataract case and the bads on your lap of the in this traumatic. Oh yeah. Yeah.
I'm all here. Let's do it. So finding what you like. The thing is you don't have to ultimately choose your final outcome now.
You can always change gears in the future. The first, like six years of my practice, I did everything that was working with you. I did general general ophthalmology. I did kids.
I did basic service surgery. I did lids surgery. I did diabetic PRP laser. I did obviously cataract surgery, glaucoma.
I did some corneal. Whatever I could do. I did everything. But then I finally, after a while, I actually really enjoyed this stuff with anti-segment surgery.
I don't really want to do PRP. The story there was I had a patient where I did this cataract surgery. They don't know cataract from glaucoma from diabetic or not. And they said, oh, it was an easy surgery.
But for them, they thought it was a laser. Oh, it must've been a laser for the cataract. It's like the year 2001, 2003 or something. Right.
And he tells his friends, oh no, go to him. He does this laser. I see great. I see a lot of people who are looking for it all.
Well, it's under his friend comes in with bad proliferative diabetic or an apathy. So I got to go to town and do some, you know, thousands of people out of PRP. And that gets his, oh my God, it hurt. And it's the laser.
It is. He docks it in a laser. It hurt. And I don't see any better.
Maybe I should be referring out all the retina to people who love retina. That's right. Yeah. Yeah.
That's really funny. My view on this sort of specialty versus generalist is that I have a tendency to believe that AI in the future, it's going to reward generalist. I don't mean by general optimologist. I just mean generalist in the way you think because it will reward type connections.
And so to have connections between like, even your Iowa accounts, like there were, there's how is this about the model to spec then who are doing cataract surgeries and using Iowa cogs. Why, you know, how did you come up with that? I did as well. You had a friend and you had this and you probably had like some weird connection.
You thought about AI, you saw deep blue and then you thought, hey, like there's a connection here. There's a connection from what I do in my domain to connection to something else that's apparently not related, but there's a definite relationship. So that's what I think, I think AI is going to be because you have the expert to just like ask these questions like, Hey, like, imagine chat, GP existed in 1999, you can like, like I just saw deep like, it's our connection to using Iowa. Can we use AI cry all costs?
And so I think that's going to, I think generalist in the idea of sort of the way I'm thinking about is well read, you sort of access to multiple different types of knowledge, you know, you're not just so focused on one particular thing, they're able to sort of draw connections out from why disparate view. So that's, that's where I think about the future. And yeah, I mean, I don't know if it specifically applies to a specialist within ophthalmology, but in terms of innovation, that is really synonymous added about. So that's, I'm trying to encourage people to be read every read a lot of stuff.
Right. So maybe instead of doing what we did, which was you go to undergraduate university and even ultra specialized thing, like I was microbiology and molecular genetics, which of course, everything I learned back then is completely outdated and useless. Yeah. Maybe I would have been better if I had like a classical education of the humanities, a more liberal arts education, where it's a lot more thinking and discussing and more the philosophy and more that, you know, it's funny, maybe actually better for the future.
So maybe when your kids go to university, maybe they should, you know, consider well around education. You know, I was a philosophy major and I did a lot of political philosophy and an ethical philosophy. So yeah, I kind of maybe are like preaching my own, you know, to my to my self care a little bit. But yeah, I think that and that's part of what we do at that our podcast, AI and iCare.
We really, we intentionally it's it prides connect the three O's of iCare because ophthalmology, sometimes we have siloed off. So we connect the optical optical companies, optometrist and ophthalmologist. And I know, you know, sometimes there's we don't see the same things the same way, but you know, we're all in the business of helping people see better and we all need each other. Right.
And so that's one thing. One of our mandates is we want to publish for everyone, everyone in the whole industry. And so that's been a real it's been great because we have optometrist and ophthalmologist and people from industry all coming together saying AI is going to be so disruptive that we're all going to need to work together because there's yeah. I like the team approach.
No, no, no, the truth is there are other players out there outside of the iCare industry that are coming to potentially eat our lunch, right? You know, think about large companies in the bear like, you know, meta and Google and their smart glasses and there's especially for diagnostic optometry and iCare, there's a potential future that we could be not be that involved in primary iCare in the same way. Right. Right.
Tell us more about your podcast. We'll put a link down below in the description. You can check it out. Yeah, it's published by Jobson, which is owned by WebMD.
It was started by myself and Dr. Scott Morris is an optometrist based in Colorado. We only launched a year and a half ago after a conference on vision Expo West, which is in Vegas. And that's really a conference geared toward the optical industry.
We thought at that time, because I had worked very closely with on optometrics, sort of industry related projects, that we need to incorporate all of iCare, including optometologists and optometrists. And we can't be so silent anymore. Because in the future, we're going to have, you know, you can imagine going to your CVS, your Wal-Mart and getting an eye scan and then being referred out and where inoculomics and robotic surgery. And we just wanted a platform to educate our stakeholders, those three people, to really educate them on what the basics of AI and iCare are.
And so we have a podcast and one of the great things we do, that we have an educational component. The first five minutes, we talk about, hey, what is AI? What is a transformer? And we're getting pretty sophisticated.
Now, what is hallucination in AI models? What does that even mean? And so I've gotten a lot of really great feedback. So I would never have looked this up myself.
I'm so glad you guys are talking about that. So we encourage everyone to please give us a look at AI and iCare.com. And I'm looking forward to speaking to other great, and thank you so much, Uday, for being on our podcast. For sure.
For a really great number of really great guests that we're excited about. Yeah, I think it's going to be transformative. I like your team approach of using kind of Biscuit, all the 3Os for essentially helping people see better. I mean, that's where it's going to go.
And I think now we're finally seeing more traction. Google Glass didn't really work 10 years ago, but now you see meta with the clever partnership with the Ray Band Glasses is really kind of picking up. See, I've seen patients on my clinic wearing them. Like, hey, whoa, okay, you're wearing these.
And they're like, tell me how great it is. And yeah, I think we're going to get it's going to be a game changer for sure. But that is approaching on our field for sure. It is sort of encroaching on our field.
And I think we need to be involved. I think patients are going to come desk and say, hey, doc, like, what about all these VR headsets? I mean, virtual reality, AI, how does it impact my eye sight? What should I be doing?
What should I be worried about? And then diagnostics that may be on those glasses themselves, you know, looking at, you know, looking at author surface disease, your blink rate, and then also doing academics on the surface of the eye. Google also released a paper recently not looking at the retina, but looking at anterior segment photos and being able to track your glycemic index based off your conjunctival photos. How?
I'm struggling for your listeners that again, the same with how to figure out your, you know, your sex and male or female. It does it. We're not sure. There's a little bit of a black box, but it was able to.
They have a paper that they published. And so you can imagine a future where people like to take a selfie of their eye or their face and being able to get all sorts of systemic information. Wow. Yeah, for sure.
How do you put two together? Yeah, obviously the retina photo, but I didn't even think about the anterior segment photo. Yeah. And all sorts of other things, looking at the lid movement to determine early onset Alzheimer's or Parkinson's.
I think those are all sorts of things that are coming that it's going to interface really interesting with the optimal and the optometrist. So when they have these Ray bands on, they may get a notification in the future saying, oh, you know, I got a notification. So as I care providers, we really need to be on top of this and understand how these devices work and also interface with these devices because it's going to be a big, what a great world to be able to catch things early and to educate our patients. And things, you know, when they have devices that they need some help with.
So I'm really excited about it, but I think we should approach it with some caution and help these companies really do it in a patient-centric type of way. Yeah, for sure. I was just, you got me thinking even just our eye exam now, and Milk Health's changed over the last 20 years. Now, essentially every patient is going to get an OCT, macula, OCT, optic nerve, you know, and every one of my pediatric cataractors gets that.
And obviously majority of the time we don't build for it because there's no disease or you check anyway. And I just automate the entire exam, but I don't even need to be there. Couldn't I do anti-segment photos? Right?
I couldn't I do the posterior segment scan with OCT, whatever else. I can do a wide, wide image retinal fundus photo. I mean, do I need to see the patient person? Yeah.
At our podcast, we talked about the future of the eye exam and what that's going to look like. What? Well, it's got, you know, you have a beautiful office there I'm seeing on video. And it's kind of funny that, you know, the last time I saw my primary care doctor, because I did one of these concierge things, I went in, got all my testing, then I went to his office.
And we talked like two normal human beings at an office with a nice desk. There's some books he asked, you know, I had, I had a couple of water talking about my health. Why can't we do that? Why can't we elevate the care and ophthalmology?
I can imagine a future where you, you know, you have, they come in, they see your myself. And the same thing with screens all around with not just foot lamp photos, but a highly, with high fidelity images in potentially in 3D and the companies that are working on it of your foot lamp, of your cornea and your segment retina all around you. And that Iowa Callettes, you have your, you know, your AI allettes with a, with a grass and talking them in a normal way. And the human with no screen, no slit lamp in the way, and talking about their options.
I think that is just a higher level of care that AI will enable for the AI example. Wow. Yeah, but that's going to sound like a lot of out of pocket. I think you're very prepared to the patients.
Now that we, the era of the sub $500 cataract. May, yeah, maybe, but I'm not sure that's actually the case. You could line up all the devices in a row. Patient comes in to all the devices and they go to your, actually makes it easier for the surgeon.
And then to the nice off, they talk to you like a normal, like a normal human with screens around. I mean, you just have a nice fireplace going and you'd be talking to that. I just, that's really super exciting because it's kind of crazy. I mean, it's still up to an amazing device, but I wonder that it's weird.
You're like, you know, you're like 10 inches away from their mouth and then you come back and you start just, and they just had a light in them. And if you think about from the customer experience, like they can't see, they're like half blind or they're dilated and you're talking to them about all these options of cataract. I mean, it's complicated stuff. Isn't there like a better experience?
But I agree. I mean, costing is a big issue, but. Well, imagine this. Imagine they come in, do all this automated testing on these fancy machines, run to back the entire eye, the way you name it, the works.
Then you have the patient interact with an AI agent to explain kind of their needs for their vision, their vision, their personality, their tolerance for this. Let's say we don't have an economy lens yet. And then okay, then the AI determines not only what would be probably given the interaction they had with the patient, what's the best lens option? Maybe the dominant eye should be a monofocal, thorac, lens, amortia, the other eye should be a multifocal lens, triphonic, whatever it is.
Then it does the AI calcs for you with a precision you could never achieve on your own. Then you go to the surgery and then you've got the robot assisting you to do the surgery and you program the robot to be far more precise than a human, far better reaction time. You're even program the robot, hey, don't let me come within five microns to post your caps no matter what. You have a saber surgery, you put the lead, the whole thing.
Look at that. The whole thing end to end. But here's the beautiful thing about that. The surgeon, the human is still involved.
In my mind, there's actually more connection. You may spend actually more face to face time with the patient in high quality time, not just after the sitlamic abrely and then you send them off to a coordinator and they don't talk to you again and after surgery is super busy. I think this is a higher level. I think patients would prefer that.
A man with surgeons would. I think this is better overall care and more personalized care. I'm super excited. I love the way you talked about the end to end solution.
I think that's the only thing. I'm only supervised by us. When I flew back on a plane, I was in Brazil a few days ago, I had a great pilot on the plane. Obviously very talented.
But I also had autopilot and radar and GPS and paid the computer systems with extra redundancy. I'm glad I didn't have just the human pilot with none of the other technologies. I want it all. You want it all.
You want it all. It's a place to human. It gives you comfort that there's a human there and it gives you comfort that there's all these. I don't give the pilot extra credit for flying without those types of other additional tools.
I want you to have that. As a patient, I want my surgeon to have all those tools. Even if they're the best surgeon ever, I want them to have those redundancies in there. Right.
I guess back to my OCT exam. If you're an officer like me who's 50-ish or older, you did your training without an OCT machine. Yeah. Those ones have a Jeff by the way either.
Right? Nothing. But aren't you happy now that I have Don Gas's diagnostic ability in this machine in my clinic that scans the eye and tells me, oh, that's the little matter hole. It's grade whatever three.
I'm just so happy that I don't have the natural innate ability to develop a Don Gas brain. He was in exception. He was the top of the top of the top. So this is, I think, going to be the same thing with AI.
You're going to have the ability to have these incredible diagnostic, but also, therapeutic treatment regimens and protocols that you couldn't really do easily on your own. Yeah. The key is not thinking AI as artificial intelligence, but thinking it as augmented intelligence. That distinction is important, especially to bring everyone on board because I know there are several.
I know many healthcare professionals who view this with so much trepidation and I get emails from them because I know this. Like how can you be doing this? Is it going to ruin the patient, physician relationship? It's sacred and you're putting a robot in AI in between.
I think the examples of airplane safety is a great example. And I think elevating the patient experience and saying this is just another tool to augment our intelligence in profound ways because we want to bring everyone on board. This is not meant to, this is like as we mentioned earlier, this is about democratizing all the tools that we have and bringing, as you say, everyone's like the highest level of potential care. Right.
I don't give my office models extra points. If you diagnosed my Mac good thing without an OCT. Well, what is the OCT? Let's like, don't for sure.
Exactly. I mean, I get you're really good at the 78 and I after it and it's a little lamp, but okay. You know, the old school frugal ends or a contact land. Okay.
Okay. But you can use those CTs. Yeah. So I don't think the OCT has replaced us.
No. All this. So I don't think augmented reality augmented intelligence, I like. That's a good, good new AI version.
I think that's not going to replace us either. No. No. So I'm super excited about all the tools we have in the future for augmenting all our abilities.
So I look at it and then part of our podcast and our publication is just sort of spreading that gospel type of thing and getting everyone on board because, you know, we're all going to do this together and we need every, you know, I care relatively is actually a small segment of the overall, you know, overall healthcare, but it's an one and we need to all work together to bring the future we want to actually realize that future together. So that's what we're trying to do. Yeah. So I tell the other people to embrace new techniques and new technologies.
This is a key example. Yeah. And I think your cat our coach is so good at that. You know, there are like so many different tools and techniques that you talk about that.
I think for a lot of people like, Oh, maybe do it. Well, I saw what they do it. I saw one of his videos hit with a guest surgeon to it. And one, one of the autonomy is that there's so many different ways to do.
I didn't think about like the technique, you know, rotating on second. I love those videos are rotating on something like I don't do that. But wow, that's that's doable. And there's a benefit to doing it in a certain way.
And I just love knowing that there's so much variability that that is out there. And I love learning about different tools and techniques that people are using. So thank you for doing that. I think it's kind of.
Sure. Well, Kelly gets back to your point about the creativity. So the spinning of the nucleus, we call it the rosetelli spin after net the rosetelli amazing surgeon, but he that was born out of necessity. So he's an ultra high volume surgeon.
So in an eight hour day in one operating, he'll do it like a hundred or more fake. Wow. Wow. And one and one and one.
Wow. And he's worked for the public council in Brazil, he does this huge volume of surgery. So he'll do 500 cataks a week, he'll do 20,000 cataks a year. He just he does cataks every day all day.
That's it. No pre-opnoposto. That's what he does character. He's method idea was, well, spending the nucleus so many times that I break up the adhesions of the cortex to the bag so he gets the entire nucleus out with just the fake of probe.
There's no need to use my a probe. Yeah. And then he's hydro implants the I will and he's done. And so it's born out of necessity, but it's to your point, it's his creativity because I would have never thought of that.
So you've got to be creative. So not going back to your kids. Yeah, you got to encourage the creativity, the thinking. Yeah.
Yeah. And I think there's a time where you sort of the phrase in necessity is the mother invention or innovation. That's exactly it. And I think the other thing is you have to be on the lookout for problems, like problems that you can that are and we were stronger problems left and right.
Like there's so many problems like challenges. And that's an amazing one. He just I mean, he was sort of forced to do it. We have different problems like in our clinics and insurance preauthorizations.
That's like that's an inefficiency that AI can help with. And so yeah, that is a fantastic straight 100 more than 100 a day. That is that is 20,000 a year. Hey, send him a robot send him one of those robots.
I think you're the question. Could I sit and supervise four robots? I could watch a screen with four robots operating saying good, good, good. Oh, you know, pause that robot for a second.
Let me see. Okay, we can resume or let me take over. Yeah. I mean, it's just going to you're going to get there.
You will imagine you're doing them. Let's say you're doing a full thickness transplant, okay, according to transplant. And you say to the the robot, okay, put a suture exactly every clock hour. Make it exactly this length.
Make it 80% t-shirt. Yeah. Yeah. It's just like done.
It's like what? So what do you how are you going to solve the training problem? Because you know, I wonder in a world of robots, we may need more gifted surgeons because the cases that they're going to get maybe even more. How are you going to do that?
How are you going to train these surgeons? Right. So that's what I say. I said, you know what, you better get step up your game because what if the robots good enough to do all the easy cases?
Right. Right. So that's going to be an issue. I think we're going to have to really think about that in a society.
I mean, this is not just outside of ophthalmology too. Like what about the tough stuff? Who's going to who's going to train in a world that's beyond after, you know, we're done operating? Who who's going to do who's going to do the hard cases?
Right. Sure. For sure. And then, you know, but I think we still always need surgical judgment.
So yeah, initially the robot will emulate us, but I think we're going to come up with newer techniques where it's just not what we could have done with our own hands. That's right. So right now, if you think about when you do a surgery, well, I like to sit temple for catheter surgery. Okay.
I like doing my sit chat. Why? Well, that's where my hands are. Yeah.
And why don't I make, you know, parainomani syndrome? Well, I have two hands. What if I had three hands? Three hands, right?
What if you're what if you're a retina surgeon and you're operating with the you're doing a three port part of the track to me. And what if you have a light pipe and it just follows you or you're just saying your infusion port, you have four ports. You're going to have an infusion in one. You can have a light pipe in the other.
And then you program the robot like wherever I move my two instruments, it moves the light automatically. Follows you. Right. It wouldn't that be just much easier for you?
Right. It goes back to your earlier point about coming up with entirely new types of solutions, right? We assume the way we're doing is sort of like the best way. It's really just the best way so far.
And this is just another innovation. I wonder if this is just, you know, fake emulsification back in the day. It was not received an enormous of ways initially, of course. And there may be an analogy here.
You know, for the young people who may not know, Kalman was really ostracized, criticized. He was like a bad guy, a villain. And yet that's how we do certain now. He said, what are you doing?
We don't need this. Same thing you're saying about AI. IOLs. IOLs were so controversial.
Do you know when IOLs first came out of academia across the US was famous for saying IOLs are picking time bombs in the eye. You should never put an IOL in the eye. This is the reality truth. Taking time about them.
Well, academics fought like no, IOLs should not be legal. You're damaging these eyes. And now you wouldn't even think you're doing this really that in Iowa. Yeah, so that was literally the saying was ask us to see North America.
And I'm not going to be doing the first robot. It's so negative. Yeah. Yeah.
And it's in an off-the-mallaby subreddit and the text when the people writing it, they're clearly off the mallages. Like seriously people, and no, I'm not going to reply. It's what forwarded me a link to this Reddit sub thread and like how I'm the bad guy now because I did the World Free for Body Care. No, that's not.
This is the future, I promise. Yeah. This is no different than Kalman doing fake or the controversy about putting an IOL in the eye. 100% agree.
And I've received, I think it's a far minority. I think most off-the-mallages are super excited about these things. They see it. And look, we see it and we feel it too.
A long day and no are. Your back is hurting. I mean, like back surgery, you know, you can't play pickleball anymore. I mean, it's really impacting your life.
This will increase your longevity. It will increase your joy and happiness in the in the OR. I think that's the promise. So I hope those negative views will become increasingly a minority.
I think they already are, especially as they start using those tools and reading about it and learning about it. And I think that's part of what we're trying to do. Yeah, I think it's exciting time. And by the way, for the full disclosure, yes, I have financial disclosures with robotic category, robotic ice-worthy company Horizon Search Group.
Also AI in off-the-mallage with LensCalx. That's another advanced Euclidean solution to the company. I have disclosures in this specific field. Yeah.
So. And likewise on the robotic stuff with Horizon. But you know, there's so many different tools. And I think it's going to become a crowded space.
And I think it's good for the industry and good for healthcare and good for ophthalmology. And so again, thanks for having me on. I really appreciate it. This has been so much.
Oh, that's a pleasure. So much to learn from each other. Now I'm going to put the links down below for your AI and off-the-mallage podcast. And then again, I'm going to start thinking about it more as augmented intelligence, not artificial.
I like that. Thank you so much, my friend. I appreciate it. Thank you, Dave.
I really appreciate it. Remind our listeners and viewers, remember we got a new podcast every single Sunday on Apple Amazon Spotify, Google, and where you find a podcast, you find the category podcast. But of course, we have a new character video every single day, probably 20, 100 videos by now and counting. And of course, check out our sister channel, retinarounds.com, as well as the retinarounds podcast.
And we can also include our retina colleagues in our fun discussions. So until next time, I'll catch you later.