Hi everyone, this is Dr. Kaila again for Signal and Symptoms podcast. Today's going to be a little light. We want to just basically discuss more of what the state of AI is, and I'm going to tell you my journey. The idea was to understand what we in the industry term called fourth industrial revolution. You guys are seeing in the whole media that AI is a bubble. AI is a bubble. But I just want to take you back and I'm going to ask the first question uh which we call big ideas and I always do this research. I think this is always nice. Not that I'm asking you to invest in this company or whatever but this has AI is accelerating the performance doubling rate closing in on the end of chessboard. What I'm saying is that that what we have done is productivity, productivity, productivity. So let me first ask my expert here the first question. Think of 1950s 50% actually 40 to 60% depending on where you look at it statistics agriculture was done by human beings as a matter of fact in United States everyone was busy doing agriculture producing food now only 2% of the population is actually doing what agriculture. We went from this was what we consider depending on which one which book you read second industrial revolution or third industrial revolution and Now even in agriculture like the tractor doesn't need to be moved by a human being that 2% is at risk. So I want to ask Harvey that in the advent of the state of AI that we are in, take us back take us back from the industrial revolutions that we have gone through from electricity to uh um you know motor cars to agriculture and the shift of human potential that has changed including so number one human working that are in that particular region. The second biggest thing is the productivity increase. And then the third thing honestly it's becoming a little scary for me. So soothe me up that how things are going to go because if AI is going to take over this much the productivity increase with AI is going to be insane. What do you think that how the state of AI is today in that length from an industrial revolution change to what do you see in 5 years?

Yeah. Tough man. and you always and I want everybody to know he never tells me the questions so these are all real thoughts so it's really tough with that said you know I like taking it to the next level let's think about neuroplasticity right why do I mention even that statement is because there's a shift going on and it's just not beyond what we're doing and how we're doing things but our brain is totally shifting and so to answer your question I just want to say the way you will do things in 10 years 20 years Even though your kids will be older and be doing maybe the same thing. Let's say he or she your kids are doctors because your brain has been formed differently, you're going to use the tools of tomorrow differently and they will use it differently. And as a result, the output will be differently. And so I want to make sure you understand that because what will happen in the future, you even though you and I love AI, we will use our creativity. We will use the way our brain and the way we form things differently from your kids and your kids may start allowing things that you would never allow because of their neuroplasticity. So I just wanted to start with that.

Now let me just go through a few more uh updates from other important publications and then uh ask uh very tough questions to Harvey. So one of the things is that that things people who do not read this is an artificial intelligence index report. I think this is my very important reading. I read this multiple times. I have read this at least three times already. It comes out essentially within the Q1 of 2025 and it is you know gives you the what we call top key takeaways . uh what they said is that that interestingly we are breaking through all the benchmarks. So they have to create new benchmarks. The second thing is this that AI is increasingly embedded in everyday life and that the choice that they made was that FDA approved in 2023 enabled devices just six in 2015. So what they're saying is that that what you're seeing is that real world classic application is going and that on the roads driving no longer. So I just want to make sure that I let uh Harvey uh sort of comment on this and then given that he's tra he's an international traveler international speaker by the way he's going to have a TED talk one more I think he already has four so one more is coming to comment on this uh for us.

Yeah so think about it um you know when you used to come into the United States when you travel it was uh paper right if you didn't have a paper that was it now you can use your mobile and now take that to the next level. When I come into United States, it is crazy. I'm literally in the United States within seconds. They I literally push a kiosk, it takes a picture of me, I'm in I'm part of the global re-entry program and then a TSA agent looks on my face, it's you're Harvey Cast. I nod my head and I walk in the country. That would not happen 10 years ago and that's making my life easier. Same thing I'm noticing in international travels. all these countries, Singapore, uh in fact, Singapore was the first time I had seen this. I I walk into TSA and there was literally maybe three uh agents equivalent for about 50 stamps and then they they go through these little doors and and I'm able to come in. But that's just simple AI causing applications for us. Uh, real quick, I want to also say as these uh, uh, autonomous vehicles become a thing, one of the main things as any of our doctor is me signing off for vouchers for people to be able to get back home or Ubers and whatnot. Now, autonomous vehicles, we will have some set up for a radius of the hospital that will go get our patients and come back.

US still leaves production in top AI models, but China is closing the performance gap. Now I'm going to be honest with you. uh we actually did talk to Alibaba from a statewide perspective to just to understand what is their capacity and everything and given that we are an international company in the US of course remains on you know our partners we have Google cloud and our partners and they actually host our vision models and everything but what but for the rest of the world we do have options and I'm just going to be honest with you Alibaba which is a cloud service and they have two things behind them the moonshot AI which is Kim K2 and the Quinn model which is Alibaba by itself or onetenth of the price and the performance gap is really not that much. What what do you think the future is uh in terms of this?

So you have some big companies in the United States that are charging 10 times more and from an economic point some of the things that we you and I are doing in healthcare do we need a limousine for a certain simple task or can we do a Chinese model or a non- US model that may cost a tenth of the price or some cases a hundth a 100th of a percent of the price from an economic point it makes no sense to do that um and as this gap gets uh closer and closer to reality and gets to a point where I needed to do this. The technology is already past that in six months and a year. So why switch models? Keep something. I don't need that extra push for some things. Um so I see fast forward what's going to happen. I I really think it's going to push the US to lower their prices because of the competition and that goes back to capitalism which is ironic, right? Because we're saying China is pushing us to lower prices uh in a around the by way think of China uh in a different light. So it's interesting.

Now this is another one. Global AI optimism is rising. So this is exactly as you told last time and people are investing in my company in you know in our joint ventures and everything which are AI based is a significant amazing optimism that we have seen but more importantly a deep regional divide remains and this is it in countries like China, Indonesia, Thailand strong majority CI products is more beneficial than harmful. In contrast, optimism remains lower in like Canada, United States, Netherlands. So what are your thoughts on that in terms of what why why people are perceiving it differently in the first place and how should we move the perception if we need to?

Yeah, good question. I I I say this a lot. I I really think it's about the culture. It's about the settings. In healthcare, we're very traditional. We take our time. We're slow adapters. We're not sure there's a cost issue. But in other parts of the world, they use technology openly. When I was in Singapore, I was shocked how many simple vending machines that would make you fresh orange juice were all over the place. And they would fill up these vending machines with fresh orange and you would walk by and they would squeeze it for you, create it. And this autonomization is like everywhere in Singapore. And so the culture accepts innovation. And as a result, Singapore is really big on AI to the point where the government, this is why I work with the government, I work with the Ministry of Health to help them regulate and they're having a hard time regulating because the people are pushing this technology and pushing pushing. They're like, "Oh my gosh, we we we barely can keep up." Whereas in the United States, we have a different type of culture and it causes that resistance. And that's why I feel like we start seeing some countries saying, "No, no, we're all in." um, and some others are like, "No, no, no. We we're skeptical."

uh and that's part of by the way US government for the first time actually invested in Intel and Intel as a chip manufacturer is extremely important for national security and they're developing and going to be hopefully competing with Nvidia and the idea was that that US government is now has a 10% stake in Nvidia which is really really odd and I don't I don't know why this dude so but if you look at Canada, China, France, India, Saudi Arabia is projected to spend hundred billion dollars on these initiatives. So number one is of course investment is also what I ask but also on the footnote you know US government actually introduced 59 new related regulations more than double the number of 23 issue by twice as many agencies. So that is actually going on as this. So how do you take both of this together Harvey ? on they are increasing the investment and also increasing the regulation and what is what what is it pointing to is it pointing to protectionism or what is it?

That's a good question if you think about it to go into the weeds what what kind of regulation is it is and and I'm playing the de devil's advocate are they regulating to the point where it helps the big guys like open AI and the companies that are uh really out there and they're protecting their interests or is it regulating to protect you and me, the small guys starting up simple AIs. And I think there you'll see is it true regulation or is it more protectionism towards their own investment if you think about it that way too. Um I'm not saying one way or the other. I'm just planning to see to to really look at it and dissect it. Uh when it comes to regulation, we all know that the big three entities are the G uh the GDPR in Europe for healthcare regulation, AI, and then the second that you start seeing more and more is California. They're doing their own thing and then the rest of the United States. So, it's interesting to see where these regulations play because as you know, if regulation becomes too uh broken up in pieces, then someone that's trying to create an AI now has to figure out 50 100 different laws and how to apply those. So, it becomes a little uh I think it can slow down the innovation.

I'm going to switch over again. This is very important read that you guys need to go over at some point in time. And then I'm going to quickly go how organizations are rewriting the capture value and the state of AI is concerned. And this is uh from McKinsey report. I would really appreciate that you guys review this as well as you go along. I'm just going to capture just one moment over here. Risk and defa two are the most centralized elements of deploying air solutions where techn is exactly becoming a really uh important thing . uh we actually talked about last time with our uh Dr. Kayat who's here u Harvey Castro's good friend that AI governance is extremely extremely important but I just want to make sure that Harvey your opinion it seems like this is a industrywide challenge corporate uh governance.

Yeah. Yeah. It's um you know we're in the wild wild west uh days right now. What do we do? How do we move forward? and how we play chess is so important to create those initial correct moves because there's so much at stake obviously in healthcare we do this wrong and we're hurting people and we're making the disparities worse or can we use AI to help decrease those disparities and help humanity so we we must get this correctly uh everybody's talking about the governance how can we create this and there's several organizations out there that are guiding uh states and national and different governments on how to best create the best uh practice.

Last two questions. So this is in air street capital Nathan Benick. We're going to reach out to him to actually discuss this whole thing. But I just want to make sure that you guys understand this resource as well and then continue to read on these important topics as we go along. I'm just going to go ahead and uh and talk about this. Think before you speak. I wish I did that. Is that one thinking ignites the reasoning race . uh I I disagree but anyways uh because chain of thought reasoning manis has been doing it before long before this and again uh within you built that coot which is chain of thought as compared to this and everything. So I'm just going to give my opinion uh how we built right AI agents and then um essentially it's a very sexy you know word at the end of the day it's it's what the mode is a workflow right where we're building an amazing workflow for acute stroke or acute coronary it doesn't matter if you use an AI agent in sequence in parallel or whatever it's it's the end goal is to solve the goddamn problem. but again your thoughts on this because um uh you know this thinking has been really off for me. What has your reading been in terms of thinking as an important factor in the LLMs?

Yeah. No, I think it's important u I I think at the end of the day we're again wild wild west there are all new different techniques how can we get the best output uh as these models and and the chain of thought itself gets better and better the output will get better. Now you and I as doctors and you especially as a your background you look at thinking very pure like a scientist you and I know too that these are just algorithms and they're just fancy math equations. So is that really thinking? You can make it a point that it's not really thinking the way you and I are thinking about thinking. But with that said, I look at this as just a different way of asking the question and and instead of for layman's people asking a question, having a the chain of thought, having the prompt created in such a way that it extracts the best output is what ultimately you and I want. And call it chain of thought today, call it something else tomorrow. It doesn't matter in the sense because there's different techniques. Just as an aside, you'll like this. I I did this hack and you'll love this hack. uh Google came out with a cool paper. God, I think it was in May about prompting. I took that paper and you can do this yourself and I use it as a GPT in the background to help me optimize my prompting because it's taking all of their research in prompting and it giving me a heck of a prompt.

Well, uh today's episode was uh short and sweet, but we discussed what the state of AI is. And I'm going to end it at Harvey's note that we have to continue thinking about thinking in AI.