Optimize Practice Alliance

Moore’s Law, Exponential Growth, and AI in Dentistry

Jessica Dean: So thank you so much, Josh Gwinn, for hopping on a call with me today, because I am very excited to talk about the special delivery this weekend. ChatGPT-5 just rolled out. And I just want to ask you some questions to sort of get the big picture and what this means for our clients and our partners specifically. 

So let’s start with the first question. As someone deeply rooted in dental entrepreneurship, of course, what do you see as the single most transformative way that ChatGPT-5 will change how dental practices operate over the next three years.  

Joshua Gwinn: Yeah, I think there’s a couple of things. And it’s interesting you said three years. So, to give you a grasp of how fast AI and model development is pacing, let’s start with a reference that most of us understand. 

So there was Moore’s Law in computing starting in the late 60s or so. And if you remember, you know, we’ve all heard these stories about how computers took up whole rooms inside the CIA or whatever, right? 

Pick your alphabet soup, but they all had these massive computers. And what they were really talking about in Moore’s Law, it was that computing power doubled every 18 to 24 months. And the basic standard is, say, 24 months, every two years. And what they’re saying is that the transistors, the number of transistors on the chip itself, doubled physically. 

So, you had twice as much power every two years. And there’s another concept called the other side of the chessboard, which is this idea of exponential growth. And what it is, is this: It’s a story from, you know, way back, way back in the day. And, you the king goes to the inventor of chess and he says, “Wow, you’re such a genius. As a reward for inventing chess, I want to give you, you know, your choice of any riches from my kingdom.”

And the inventor of chess, he thinks about it for a second. He says, “Okay, I want one grain of rice on the first square of the board, that doubles every subsequent square. That’s what I want as my reward.” And the king laughs. He’s like, “Well, that’s hilarious. Sure, you can totally have that instead of gold or whatever.” And he’s like, “No, that’s what I want.”

What the king didn’t realize is by the time with that exponential doubling, by the time he got to the other side of the chessboard, so to speak, they were talking billions of grains of rice, more rice than existed in the entire kingdom. 

So basically, he took all the riches of the kingdom by understanding this idea of exponential growth. AI is an exponential growth opportunity in our world. In a way that we have never seen. The only thing we can tie it to is this idea of Moore’s Law. 

Moore’s Law meant that computing power doubled every two years. Right now, the Moore’s Law of AI says that model development in power is doubling every 60 to 90 days. So, every two to three months, we’re doubling in power. And this is what we saw when GROK-4 hit. We’re now seeing it again in GPT-5. We’re waiting for the next anthropic cloud model and then LAMA and on and on. This is going to continue to grow and develop. So we’re here to talk about ChatGPT-5, but I wanted to just establish this idea of exponential growth and exponential thought. 

Practical AI Applications for AI in Dentistry

Now, there will come a time where the current architecture of models, which is a transformer-based architecture, can no longer really hit that 60 to 90 day exponential growth. There will be some slowdown. But what we expect is that model development will be so high by then that human scientists and engineers coupled with AI will come up with a new AI infrastructure that can take the models to that next phase, whatever that becomes, if it’s AGI or ASI, whatever it becomes in the future. 

So with GPT-5, though, the next three years, that long-winded story, we’re to say, man, I have no idea in three years. That’s the thing. I’ll tell you this. As a dentist or an entrepreneur, if you’re not using this in three years, you will be so far behind. You won’t catch up. So the things that I want people to focus on are today,

“What do I need to do today to get started?” 

Well, the first thing, you may already be using ChatGPT for basic things. You know, how do I make X recipe? You know, looking at, like you do Google, literally, like you’re Googling things. In my mind, that’s lazy prompting. And I really want folks to think about this, because if you’re doing deep-level prompting, you’re establishing the mindset of the machine. 

And I’ll give you a for instance. I’m taking my family to London and to Dublin this fall. So over the weekend with the new release, I brought up agent mode in which ChatGPT acts as an agent that has its own desktop computer and can run task programs, very complex tasking for you. 

So, what I said to it, this is my example of a not quite so lazy prompt. I said, “You are a European travel specialist, one of the best in the world, with a specific knowledge of the United Kingdom. I plan to take a family of five over the following dates to the following places. I want you to focus on flights with these specific companies, then focus on price, then focus on hotels with a specific brand, then a secondary brand near the following tourist sites. And in the end, I want you to build me out a 10-day itinerary that includes restaurants, that includes tourist attractions, and stops, mostly walkable with limited use of cabs, buses, public transit, or Uber.”

And then I let it run. And it asked me a series of questions: What kind of food do we like? How old are the kids? What are we trying to develop?

And then I hit submit and I walked away. About two hours later, ChatGPT-5 came back to me with a full itinerary as if it had been developed by a European travel specialist. 

So, I say all that to just say the possibilities of how entrepreneurs can use this, as long as they’re thoughtful about what it can do, what are its capabilities, they really are endless. 

Having said that, the next three years, how are they going to use this?

• Number one, agents are going to revolutionize your scheduling workflow. 
• Agents are going to revolutionize your RCM workflow.
• AI agents are going to revolutionize the way you communicate with insurances. 
• They’re going to continue to read radiographs and give you second opinion analysis on what you’re trying to complete for the patient. 

If you’re not using this already, it is deeply embedded in the dental world and is creating great, great achievements and growth. 35% more treatment planning effectiveness. Agents are also going to listen to clinical notes, and they’re going to input those notes directly into your system. 

They’re also going to be there for hygiene, for when your hygienist is doing things like perioprobing. They’ll just speak those probing numbers out loud. The agent will capture that and deliver it directly into the chart. Now there’s no need for that second assistant to be in there with the hygienist, etc. So, on a base level, that’s how AI in dentistry and agents are going to affect dental practices.

For ChatGPT specifically, you can take the agent mode and with the correct prompt, you can have it run a full site-by-site, competitor-by-competitor marketing analysis for your entire city. You can find out what specials every single practice around you are running. You can find out what the most competitive positioning you can have in the marketplace, and you can do all that within about an hour, as long as you understand the modes of ChatGPT and the specific prompts that you need to use to get there. 

The other big updates that I think are going to really push people forward, ChatGPT-5 now auto-selects the model that it’s going to use based on your prompts. So, if you’re lazy prompting, it’s going to give you the cheapest model that does the least, full stop. If you say, “Google restaurants near me,” that’s what you would put on Google, right? If you ‘ChatGPT restaurants near me,’ you’re going to get ChatGPT 3.5, and it’s going to give you a list that it pulled from Yelp or Google. 

You could have done it on any of those sites. If you say, you know, “ChatGPT, you’re a, you know, a huge restaurant advocate who likes to try all the newest and trendiest spots with a particular taste for Indian food, Mexican food, and Japanese cuisine. Find me the top five restaurants and rank them in order of newest, trendiest, and highest ranked.”

You’re going to get a much different model. ChadGPT is going to go to a tighter model, 4.0, 4.5, potentially deep research, depending on how deep it needs to go. And then it’s going to give you a very nuanced answer that’s going to directly hit to your taste, Indian, Mexican, Japanese. 

So that’s a huge development overall. And then another big stat that kind of popped is 80% fewer hallucinations from the 5.0 model versus the 4.5 model, which was already a pretty large jump in an overall kind of trust and truth, if you will. 

Enhancing AI in Dentistry, Not Replacing the Human Touch

Jessica Dean: Oh, that’s phenomenal. Speaking of dental practices specifically, there’s often a fear that AI will replace human touch in healthcare. I think we’ve heard those over and over again. How do you see ChatGPT 5 enhancing that as opposed to replacing the human touch in patient care? 

Joshua Gwinn: If it’s really well-used and really thoughtfully deployed, there are opportunities for the memory function to be able to give you a depth of knowledge about your patient. 

What’s interesting is the new token window. So token, think of it as your prompt window. The new token window and prompt window inside ChatGPT-5 in some cases can take up to 400,000 tokens. 

That would be your entire patient record for all your patients, let alone for one patient. You could input all of your patient records and have it create a very customized patient communication plan to where because you’ve gathered little bits of data, it can give you a communication plan to where you could literally, when they walk in, say things like, “Hey, it’s so good to see you. How’s John and Tina? How’s your wife? Janet? How was that vacation you guys took? Did you really enjoy that?” 

And you can show a depth of care and personalization for each patient that only the best practices do today. Now, it happens, but it is so rare. And the absolute best relational customer service patients or customer service practices are already doing this. ChatGPT can just put it all together for you for every single patient, give you scripting for your receptionist, and you’re off to the races. 

The other thing that happens when you do it and you’re that thoughtful is that you start creating ‘yesses’ for each patient. The other thing you could do, which could be interesting, is create little interviews for your new patients when they come in so that you gather tons of data and knowledge about them that ChatGPT can then review each time they come in. 

You’ll have this ability to review some of that data. each day and it can continue to build that relational communication plan as you go each time. I think that could be absolutely crucial if you really, really were thoughtful and wanted to build some really neat, tight relational pieces with it. 

Getting Started: First Steps with ChatGPT for Dental Practices

Jessica Dean: Well, then kind of getting down to the nitty-gritty then and specifically speaking to dental entrepreneurs, what would you say could be their very first, most simplest, maybe most practical first step that they can take to start integrating ChatGPT-5 into their business? 

So we’re not talking like, you know, big AI-driven software. We’re talking about really using ChatGPT-5 as a sort of getting your feet wet. 

Joshua Gwinn: Yeah, I’ll throw a couple of things out there. And the first is the most obvious to me, but it may not be for others. Sign up for the pro plan, pay the $20 so that you have access to the best models all the time. 

You can run on the free plan, totally fine. If your practice isn’t generating an extra $20, though, we should talk about many other things Optimize can help with. 

• But pay for that pro plan so that you have access to it. That’s number one.
• Number two, schedule time every single day, 30 minutes, 15 if that’s all you got, to sit down and use the tool. 
• Number three, get some training, even if it’s just YouTube videos. Get some training on what the model’s capable of, what the whole LLM, as it were, the whole ecosystem that goes around ChatGPT, what is it capable of? 

Get that training and use that 15 to 30 minutes a day just to practice and work through things. And then I’ll give you two easy things for your business just to try. 

• Number one, run the marketing analysis that I just spoke of. Give your zip code, do a great prompt, prompt out a marketing analysis and see what the practices around you are offering to new patients. It’s very interesting to find out.
• And number two, upload a 12-month financial and have ChatGPT act as a financial analyst and review your financials for trends and areas that you could be more efficient to bring more profit to your bottom line. 

Start there. If you get through those two things, you will see so much value. One, just the financial review is worth way more than the $20 you paid. And then secondly, once you see those things, if you just implement a couple things off of that, you’re going to see some movement in your practice regardless, and it’ll start to change things. 

Hidden Game-Changers in ChatGPT-5 for Dentists

Jessica Dean: Okay. So you’ve been at the intersection of AI in dentistry for a couple of years now, pretty solidly. What is one sort of behind the scenes ChatGPT-5-capability that you think most people don’t know about or something that could be a game changer for their practice? 

Joshua Gwinn: Yeah, there’s, man, there’s, I think there’s quite a few. I’ll say this. How about automated patient recall campaigns? You can upload all of your patient data into ChatGPT and it can build out a full recall campaign. 

• You can build an agent that will then auto-text all those patients on your behalf.
• Another thing it could do, it can produce multilingual education materials. If you need your patients to be able to read or write any language, it can flip it into different languages at any moment. You can use that.
• You could use it to draft insurance prior authorization requests. It’ll pre-autopopulate using your patient’s data and your treatment plan and just upload those. And it’ll build your pre-offs.
• It could audit your compliance documentation, upload your whole client’s plan and let ChatGPT have a swing at whether or not you’ve got compliance up to snuff for HIPAA, for OSHA, etc. 

Those are four that I guess pop to the front of it, but there are really so many more. The way we like to think about it is this mentality of AI first. And what I mean by that is, specifically, “if we’re paying a human to do it, could AI do it today?”And could it be more efficient if it did? And then trying it out.

And it’s not to replace people. We keep saying this. It’s not about AI in dentistry taking humans’ jobs. It’s about AI enhancing what humans do so that we can give the type of service that we all want anytime we go anywhere. 

There is a golden age of service coming our way where we’re going to have somewhat flawless execution on the data and numbers side. And the humans in the loop who are watching the data and have to be paying attention to that can spend 95% of their time being good people. 

Guiding Sherpas through the process. You know, we know a lot of folks come in and they have dental anxiety. What a great way to get rid of that by having your own personal concierge walk you through that’s not also thinking about scheduling, automated recall, RCM, answering the phone, all the million things that we don’t want to be thinking about. That’s going to be a big part of it over time, I think. 

The Competitive Edge: AI vs. Non-AI Practices

Jessica Dean: I am going to throw a question at you now that was generated by ChatGPT. So I’m going to get you to take on this because it’s a little bit interesting question. 

If two dental practices are identical in every way, except one uses ChatGPT and the other one doesn’t, and this is in their operations, their marketing, you know, front to back, everything, how would that AI-enabled practice look different after one year of using ChatGPT?  

Joshua Gwinn: Oh, 10% to 15% profit margin, without a doubt. And again, when you say ChatGPT, ChatGPT, I’m taking that as enhanced AI, right? ChatGPT in this sense is just the tool we’re talking about, but any of these tools are all working off of large language models, be they ChatGPT, be they Claude, be they Llama, you name it, either way. The point being, if you take the enhancements that are available, I believe there’s 10% to 15% margin and a much faster cash cycle. 

So speeding up that cash so that there’s more cash available for you to do anything from distributions to more investment in growth, development. And the way we like to think about it is, depending on your goals, you can use AI to develop your own virtuous cycle of, you know, of growth and development inside your practice. 

Use these tools to enhance. And if you were running a 15% margin, that’s a good business today and paying yourself a fair wage. But if that were 25% in a year and that 10% becomes your investment. The game changes pretty rapidly on that course. 

So, that’s really how I would think about it is the opportunity, and it’s not necessarily in losing people. There may be some inefficiencies in your system. I can’t speak to that. I’ve seen practices that are highly overstaffed. But most practices are normally staffed and have pressure because it’s hard to find good people. 

This can buttress and enhance what those great people can do. And if you’re down a person, maybe it doesn’t hurt so bad because you’re able to really still push the support necessary and hit the functions as need be. 

The practice without it, it’s going to look the same, Jessica. It’s going to be the same as it always was. And in three years, that practice is going to be obsolete. One thing that we are seeing, 65% of patients are now hyper-aware of AI radiography. So, they’re asking, “Are you using AI to check these, these x-rays?” And if you’re not, they are going to start finding practices that are, that’s going to happen. 

Jessica Dean: Wow. 

Joshua Gwinn: That will change, absolutely change. The patient loyalty factor, especially with the younger generation, is going to be loyal to the practices that are using the most enhanced tools to make sure they get the best care, straight up. 

Lightning Round: AI in Dentistry Insights

Jessica Dean: Yeah. Yeah. That makes sense. Okay. We’re going to do a lightning round of questions. Okay. Let’s do it. Ready? 

What is one word that describes ChatGPT for dentistry?  

Joshua Gwinn: Revolutionary. I mean, it’s going to change everything. It already is. For those who aren’t using it, it’s already changing everything. One more thing I want to say, Jessica, like, load your denials from insurance into ChatGPT. Let it tell you what the problem is, and it’ll fix your denials. Then resubmit. Boom, you’re done. 

Jessica Dean: What? 

Joshua Gwinn: Like, people spend hours and days on that. No one wants to read that crap, load it up, have it, fix it, and then move on. 

Jessica Dean: That’s brilliant. 

Joshua Gwinn: So anyway, that’s a must-do.  

Jessica Dean: It’s must-do. Yeah, absolutely. What’s the biggest myth about AI in dental practices? 

Joshua Gwinn: I think the biggest myth is that it’s going to, and this is not just dental practices. It’s everywhere. It’s so scary that it’s going to either destroy everything, we’re going to hit this dystopian landscape where only the haves are going to survive, et cetera. 

And I don’t believe that to be true. I believe we have the potential to hit a golden age of society and a golden age of dentistry and, you know, have a clinical tool next to you that is enhancing your knowledge base. 

Right now, and this is across healthcare, it’s really across almost everything. But in healthcare specifically, if I go to a general practitioner doctor who’s never heard of a specific thing and I have that specific thing, I might die because they have no idea.  Now, this tool has heard of all the things. And so it’s going to aggregate data and tell my general practitioner, hey, check out this thing you haven’t heard of. 

Or send this guy to see someone who has so I can get that specialty care I need. In dentistry, it’s going to be the same thing. It’s going to be this idea of you have a multitude of ways to treatment plan and sequence care, that is the most efficient based on your inputs. 

Do you want to be most efficient by time, by visits, by income? How do you want it? And then ChatGPT will tune it up, or any model will tune it up and give you the feedback as to how to make this patient experience the best it possibly can be for this treatment plan. And if you’re really communicating and engaging with your patients, this is a conversation you’re having together, and you’re tuning the model to their preferences and yours. So that you have this experience that you both really enjoy and it’s super interactive. So I think that can be absolutely critical. 

Jessica Dean: One task that ChatGPT can do today that still blows your mind. 

Joshua Gwinn: Oh, loading up insurance denials and getting me paid, so I don’t have to deal with it. That part is so critical and it’s so frustrating, especially for those of us who hate to do double work. “I did that. I submitted it. It got kicked back. I don’t want to do it again.”  

And I see practices, because of that one little mental tick that they have, certainly do. I’ll do an amazing first job. At least, I think it’s amazing. And when it doesn’t work, I’m done, right? You need to bring in, bring in, call on the cavalry. Somebody else has got to support.  

And in insurance, I’ve gone into, I mean, tens of practices that have stacks and stacks of denials. And they just go away through untimely filing and they lose all that money because they don’t want to go back and redo what they already did. They’re like, “I’m working on today’s stuff. I don’t have time for that.” The idea that you can load all that up and get answers quick and reprocess claims.  

You need to think HIPAA. You need to disidentify data. Make sure you’re not putting HIPAA data in, but that’s a big part of it. And for me, I mean, it does a little bit blow my mind.  

The second, and this isn’t ChatGPT related, it’s AI related, is that it reads radiographs and color codes the types of material in the teeth. So, you can see the layers of the tooth very, very clearly. So, when someone says, “This is a cavity that’s just barely into the enamel,” well, I can see the enamel. And then when it says, “but now it’s into the dentin,” well, now I can see it in there. Or if it’s into the root, I know whether I need a root canal or not. “You can’t fade me. Yeah, like I see it. It’s in the root. Give me a root canal or don’t give me one because I can see it’s just in the enamel.” 

 And the idea of trust, in my opinion, actually goes up dramatically as a clinician when you’re showing them the proof. Yeah. You know. The doubt goes away. So, you know, training plan, acceptance goes up. It’s been proven. We’re seeing it across the board right now. 

Jessica Dean: And similar to what you were saying about the rare disease example, it’s sort of a democratization of information, right? We’ve been in this information age for a couple of decades now, but it has become more and more tiered. And this is truly a democratization of all information, which is… 

Joshua Gwinn: I will tell you, and again, this, I hope, gives some sense for how important I believe this to be. China started training all their children on AI starting at the age of six. That started this year, this school year, all of their children. America’s not doing that currently. In the new AI action plan that came out from the administration a few weeks ago, there is some of that baked in. 

However, having said that, I have reached out and found programs through… through some of the leading AI labs for my own children, and I am prioritizing my six-year-old and my 13-year-old getting trained up as fast as humanly possible, because I fear that AI ability to use it, understanding of, development and being able to conceptualize what’s possible is going to be the most important future skill. 

Because to your point, the data is going to be available for all of us. The democratization of data will be complete at some level. When that happens, the real skills you’re going to need are prompt generation. Can you prompt effectively to get out the output that you want? 

• So are you giving it the input it needs to give you the appropriate output for your solution, whatever you’re trying to complete? 
• And then number two, do you have a true understanding of how to layer tools together to develop the capabilities to get done the things that you’re trying?  

If you have that understanding, once people have that understanding, the game changes. Those who never get it or see this as a fad or a bubble or all the other things I’ve heard, they’re really going to struggle. 

It’ll just happen to them. A lot of times, I think in terms of dividing the world into those who make things happen and those who have things happen to them, we’re about to see a whole lot of people have things happen to them in ways that they cannot comprehend. 

And only those of us who decide to go and learn now and figure out the process and what’s coming are going to be able to make things happen in this new world. It’s going to really compress that ability. 

Jessica Dean: The more things change, the more they say the same, right?  

Joshua Gwinn: 100%. It’s always the way you respond to the adaptation that makes or break your mindset. Charles Swindoll, right? Life is 90%, you know, 90% how you respond to the 10% of things that happen to you. 

Jessica Dean: So, absolutely. Absolutely. Absolutely. One last question. What is one mistake, or maybe there’s more than one, but one mistake to avoid when starting to use AI in dentistry? 

Joshua Gwinn: Specifically, in dental practices, be hyperconscious of the following:

• You need to disidentify data so that you’re not breaking any HIPAA or any compliance, and understand that what you tell ChatGPT is not protected information. Let me be crystal clear about that. There are no legal protections. For instance, some people use ChatGPT as a therapist. If that information gets subpoenaed in a lawsuit, they can have it. There’s no legal protection. There’s no therapy protection. There’s none of the things that we enjoy today with our professionals. When we feed them into ChatGPT, they [our right to privacy] go away. So be very conscious of what you’re saying. And I was listening to one of the AI Lab CEOs, and he said specifically, say things like, “my friend, blah, blah, blah,” like you would talk to, you know, someone back in the day, and they’d be like, “Oh, your friend.” Speak like that, because that keeps you legally compliant [protected]. You’re not admitting anything. You’re not doing anything. So that’s really what I would start with is, hey, there are no legal protections here. That’s a big piece.
• And then secondarily, if this is a clinical decision you’re making, double-check your work. Double-check your work. Make sure that this is rooted in sensibility. If it’s something you’ve never heard of before, I probably wouldn’t recommend you trying it for the first time on a live patient. I would make sure that you’re doing all the work to double-check. I mean, there’s some great, great, great clinical research out there. And perplexity, actually, you can tune that model to just go search PubMed and give you answers directly from only clinical text inside. 

So there are ways that you can still use the models to double-check for accuracy and make sure that you’re really getting to a place where you’re getting the best data possible. 

Jessica Dean: Wow. That’s good advice. Well, thank you so much, Josh Gwinn. I really appreciate your time today talking about this.  

Joshua Gwinn: Yeah, my pleasure. Let me know. I’m always here for this. It’s good stuff. Awesome. Thank you.  

Conclusion: The Future of AI in Dentistry

ChatGPT-5 isn’t just another tech tool, it’s the beginning of a new standard in how dental entrepreneurs operate, grow, and care for patients. From boosting efficiency in scheduling and billing to enhancing patient trust with personalized communication, the opportunities are endless.

The key is simple: start now, learn daily, and think AI-first.

Ready to See How AI Can Transform Your Practice?

At Optimize Practice Alliance, we help dentists navigate the future of technology with confidence. Book a free 30-minute consulting call today, and let’s analyze your dental practice’s AI usage and uncover how you can leverage tools like ChatGPT-5 to grow faster, operate smarter, and deliver unforgettable patient care.