In this Conversations with Pavan episode of Lykken on Lending, David Lykken sits down with Pavan Agarwal, founder of Sun West and creator of Angel AI, to uncover how artificial intelligence, blockchain, and stablecoin are reshaping the mortgage industry. With Angel AI already empowering hundreds of thousands of loan officers and realtors, and with billions of dollars in new global capital entering U.S. mortgage markets through innovative stablecoin-backed structures, Pavan explains why the old ways of doing business can no longer compete. This is a must-listen discussion on the future of lending, the rise of AI twins, and the new financial era redefining trust and transparency in mortgage finance.
[David] Listeners, we’re back with Pavan. We’re calling these sessions Conversations with Pavan. And why should you be interested in these conversations? Well, it’s because number one Pavan has got Angel AI. And I got to tell you a story listeners, because this weekend, I was celebrating a birthday with a client and a friend. We were out in the West Coast and having a ball and I had made the recommendation that this client consider looking at Angel AI because of legacy thinking. Well, I need this LOS, I need that. He’s just thinking legacy thinking and so he was going down a different path. Well, we’re sitting around dinner and I said, hey, what do you guys think about Angel AI? Any of you working with it? And one of the gals raised her hand slightly. She says, well, actually, I use it every single day and another will chime in and says, well, I do too. And the guy looked at her and says why are you using that? He says because of all the scenarios. Now, this is the number one mortgage loan originator in the nation. Celebrated. You can look it up and he’s number one. And so his name is Brian Cooke. I’ll put that out there. So everyone knows he’s number one. And he was not aware of that amount of that his staff is using Angel AI. One says, don’t tell Pavan this, but I use Angel AI for structuring all of my deals, and when I get an investor that’s turning something down or giving me pushback, I go into Angel AI, ask for the chapter and verse on these guidelines and I get them and then I forward them over to the investor. All of a sudden they go oh yeah, where’d you find that? You mentioned Angel AI? So, Pavan, your product that you’ve created is going to be used across the industry, and what’s so fun is it’s being used at a grassroots level. You have created a movement, not just a new technology. You created an entire new movement, an entire way of the way we do mortgage business, and it’s all new. Welcome back to the podcast, friend.
[Pavan] Good to be back. I always have so much fun with you.
[David] Well, these conversations are so important for our listeners because this is happening. It’s not something you’re trying. You’re not trying to sell it, you’re giving it away. So that’s the most important thing. I want listeners to understand that this is something. You go to ask angel.ai and use this product for free. What inspired you to do that Pavan?
[Pavan] What inspired to make Angel AI? What inspired it to be free?
[David] Well, both, I mean both. Let’s go and start with the making and then to then give it away that’s, you got over 200 million invested in this, if I understand correctly
[Pavan] Yeah, over 125 million R&D and 300, 400 million in marketing and branding and 20 years, at least 20 years of my life in it, and these things don’t get happened overnight, the key is I was trying to solve a problem Like the industry is just too slow, too difficult.
[David] And fraught with, fraught with air, fraught with air.
[Pavan] That’s what I was getting to. Too many people have lost their life savings because of the industry. We all lived through the credit credit mess in 2008. And we had to fix that. And just being in the industry originating loans, I see how difficult it is and I’m like, and my dad always told me he says, kid, you could automate it. Like, what’s so hard about this? It’s just making decisions and moving data from this point to this point. Right, just automate it kid.
[David] Well, and I’ve got told this story many times in the podcast where I got. It is worth real, condensed down. I met Pavan when he was 11 years old. I was in selling some government loans, got me loans to his dad, Hari, and I loved that man. He was a mentor to me and a dear, dear friend. And then comes this kid, 11-year-old, and he says meet my son Pavan. Fast forward, 44 years later. Now he’s 55, or I guess you’re a little older than that. We’ve been working together, running together now for a while. He’s built this system that is what his dad told him to do go build this, figure it out. And he has it’s absolutely amazing and it’s being used across the industry and it’s not done. That’s the thing is, you’re continuing to push the envelope to what’s going. I want to get into so many things in this podcast, but the first thing I want to talk about is that you recently had a valuation on your business. I’ve got a copy of it. I’m not going to give the exact number out, but it’s extraordinary and I’d like to talk about it. It’s tens of billions of dollars.
[Pavan] Yeah, and you know, and it’s not that surprising if you take a look at comps and where other AI companies that do a fraction of what we’re doing are being valued at and it’s not about what we’re doing, it’s about the technologies, how it does it. Right, because most AI companies are what we call not really AI companies or API companies, meaning they don’t really have any AI, they just connect to different AI providers and claim to be an AI company. Okay, so we actually have over a hundred and I think over 105 now patents, worldwide patents, plus a completely novel AI model, right, so? And we just had a great new release which is super fast, always accurate, as always accurate and trusted and backed by warranty. And we encourage you to go to Ask Angel and try and mess it, and ask scenarios, try to confuse it, try to trick it. We love that because that’s how it learns and that goes back to your other questions. Why is it free? right, because we need a massive number of people hitting it and using it. Okay, and to get smarter that way.
[David] Yeah Well, and it’s continuing to blow me away as to how good it is and how fast it’s going. You touched on something there. It’s a private model, not a public LLM. A large language model which is fraught and you also talked about many other quote-unquote AI companies or AI initiatives within the mortgage industry is connecting out to the various, you know, large language models out there, the cheap ChatGPTs of the world out there, and they’re bringing back erroneous message or erroneous answers and responses, where yours is not doing that, because you’ve been building this for over 20 years on actual decisions that your underwriters made for over years and perfected and perfected it.
[Pavan] Yep, and that’s the key. I mean, when you talk to Angel AI, it’s like you’re talking to tens of thousands of underwriters and loan originators at the same time. It’s the best of their knowledge coming into it and that’s the hard part. It’s the data science, it’s the collecting of the data and organizing it and then the second thing is that we build a custom neural net. It’s our own AI from the ground up that delivers these answers and it’s only because of that we’re able to deliver accuracy and reliability. So, when you use Angel AI, you’ll see at the lower right-hand corner of every message that it sends you, it’ll have a little shield with a check mark in it and that’s your assurance that answer, there’s a warranty on that answer. So, that means, even if it got it wrong and you submit that loan to SunWest, okay, we’re going to honor it and fund it, and that may mean that we scratch and dent the file.
[David] Yeah, yeah, and it’s important to note that you guys, you know, Celligence is the company that owns Angel AI. And you also have SunWest that your dad started and that you now own since the passing of your dad. Right and so and go ahead.
[Pavan] Just wanted to follow on to that, just to tell you why I can do that warranty Okay and provide that level of no one can do that. All of the mortgage bankers listening to this, I dare any of them to say anything. My underwriter says I will back 100%. He just can’t do it.
[David] No one, yeah, but you are Now that’s I should. We should clarify that’s before those loans are being submitted to SunWest, but the whole system is you’re not going to warrant a loan, that’s not going, you’re not buying. Loans are being sent, submitted to Sunwest, but the whole system is you’re not going to warrant a loan. That’s not going. You’re not buying because you’d be. Yeah, I mean. So that totally makes sense, but it’s still the same system. The same system that’s being used by those that are not brokering the loan or funding the loan through Sunwest is the same system that’s being used by those like on that scenario, where they’re actually brokering the loan out to another company and it was getting turned down, and so they went into the Angel AI and captured why it should not be and chapter and verse and send it out to the lender and they go oh my gosh, you’re right. Where did you get this information? And that’s where I think that’s the most important part is you said you touched on it. You’re trying to create something where it is being perfected, always learning, always perfecting, and it’s got amazing accuracy already. That is all well and cool, but that’s what’s really attracting the valuation, because there’s nobody out there. There’s many others out there that are doing similar things. Some of the largest two wholesalers out there are doing this and we all know who they are. I don’t want to mention them by name, but they’re relying on the API route where they’re pulling in on. They say it’s proprietary, but they’re relying heavily on other non-proprietary LLMs or large-language models, the ChatGPTs and the various others. Talk about that real quickly if you yeah, so it’s a contrast.
[Pavan] Yeah, just let me just first start with the result. The result is in the last seven years since we’ve released the product in production, we’ve gone through 27 various agency audits Freddie, Fannie, Ginnie, HUD, VA, right, and audits that’s like three, four a year, okay and zero, literally zero, loan defects, right. Zero buybacks
[David] For anyone. I’ve owned three mortgage banking companies and to have a technology or a decision with zero defects over that long a period, it sounds kind of too good to be true. I’m sorry, but I know it to be true.
[Pavan] Yes, it is true. And this is what’s wrong with plugging into a Gemini or ChatGPT or something. And, if you can, I get brokers and loan originators asking Gemini or one of the systems that are built on Gemini, a question, and it gets it wrong, and then they ask Angel AI the question, and it gets it right, and then they challenge us. It says, well, Gemini says you could do it right, and then we pull out. You know, it’s just, it’s just wrong. Right, and it’s always on. You know, Gemini, these systems can give you the broad answer right, it’s when you read it. Yes, mostly it’s right, but the specifics is the devil’s in the details and this industry is about the details, right. So those systems will guide you where you need to go, but it cannot take you across the finish line. You can’t put your money on it. There’s no shield on those systems saying and no, anybody putting those systems out cannot say whatever it says, I’m going to, I’m going to follow, I’m going to honor it.
[David] Yeah Well, that’s the part that a lot of people just have to be honest with you.. So, listeners, this is free. You can go out and try it. If you don’t believe us, try it and use it. Many others are and literally the adoption rate of on you send me the numbers regularly it’s through the roof. Give us some insights. Give our listeners some insights of how many people are coming on board. I want to create some FOMO.
[Pavan] Yeah, so you know, this quarter we had a 13% acceleration. So that means every month, 13% more users were being added than the previous month. So users being added every month, but it’s accelerating. The growth rate is accelerating. The rate of growth is accelerating. So we have something like 230,000 users right now and I think last month 15,000 or 16,000 new users were added.
[David] That’s staggering. And it’s at the grassroots level. That’s the part that’s so exciting.
[Pavan] Yes, that is a single month. I mean, you know, the biggest mortgage companies don’t have more than a few thousand loan officers, and most of those are loan officers. There’s loan officers and realtors, right. So we’re growing in every month more loan officers and realtors on the platform than the biggest brokerages and biggest lenders have in their entirety.
[David] I’m imagining this conversation with a loan officer is having with a realtor and the loan officer is saying well, we’re struggling with this aspect of the loan, we’re doing like that, and the person goes well, and the realtor then comes back and tells the loan officer what the chapter and verse is and why they should be able to do it, because the realtor is using angel AI and the loan officer is not. I share this. Yes, I’m your friend with Pavan and, yes, I believe it, I’m promoting it. Yeah, yeah, yeah, all that. But the reality is, listeners, I’m trying to save you in embarrassing moments because it’s happening where more and more realtors are starting to use this system and they’re going to start quoting chapter and verse back to you and your underwriters if you’re not using it. So I don’t know if you want to endure that embarrassment or not, but when you have this kind of growth, I’m not going to be embarrassed by it. But this is opening up so many new things. Pavan, not only are you working with continuing to enhance and expand and improve the brain that’s working on this and it’s all proprietary to you again but you’re also expanding. You’ve expanded to Angel Twin. You’ve expanded into AI, brought AI into capital markets. Rob Christman posted about this yesterday in his blog. He said AI in the capital markets. He put a question mark. Yep, SunWest and his affiliate, Celligence, the creators of the advanced Angel AI technology. Celligence, the creators of the advanced Angel AI technology, set up SunWest Investment Trust, SWIT and a strategic expansion of AI-powered mortgage-backed assets into international markets, and beginning with Japan. Pavan what he’s talking about here is so exciting. The implications are so extraordinary. Let’s talk about this. Let’s dive into Japan now is interested in this. Why?
[Pavan] Well, again, we talked about this a little bit, about how 2008 happened, and Japan had his own version of credit crisis as well before that. Okay, so there’s the billions, hundreds of billions of dollars of capital ever since these disasters have happened, that’s sitting on the sidelines, that is, refusing to go to work in the fixed income markets. They’ve been burned not once, but twice. Okay, and so Wall Street can’t see these people. No one has any view into these people. They’re out there. There’s a lot of this money sitting out there. People no one has any view into these people. They’re out there. There’s a lot of this money sitting out there, but no one has any view into it. Okay, so what we did was we talked to them. We said look, you don’t have to worry about getting misled by rating agencies and investment banks and due diligence companies and all these other layers in between that have misled you in the past, because now you can talk to the AI and tell the AI exactly what you want, what your risk profile is and exactly what kind of loan products you want, and the AI will originate those. Okay, and because the AI is so accurate, right, it’s like you’ve got. You’re assured that whatever guidelines you ask for, you’re going to get. So we’ve tapped into billions of dollars of new capital that wouldn’t. That was sitting in the sidelines before.
[David] And I wouldn’t have gotten a who got burned. They got burned in the US market. In the case of Japan, they got burned in the US market where they invested in mortgage backed securitiesapan. They got burned in the US market when they invested in mortgage-backed securities. Then they got burned in their own market. And they’re going until we get a system that can make this a bulletproof. We’re not coming back in, but they’ve seen a bulletproof system in what you’ve created. Am I hearing you right?
[Pavan] That that is correct? because they looked at the AI. They look at how well it does it. It’s like I’m going to get exactly the product that I asked for and nothing else. So if I asked for the wrong product and that product doesn’t perform, that’s on me. I am in control, me, the guy giving the money. I’m in control and obviously part of that trade is because they’ve seen the valuation of Celligence, they’ve seen the valuation of this technology and you know, no one’s going to put in a billion dollars, and right now they’ve committed a billion and they’re scaling that up to 10 billion. No one’s going to put in that kind of money right into this, unless the company that was behind it is worth many multiples of that. So that’s kind of how it all works, right. You got to have the foundation, you got to have the net worth and then you got to have the product, right, and so the good news for America is all this money that’s sitting on the sidelines is rushing now into the US market. And we just kicked it off in a small way with the HELOC, with our new HELOC first lean HELOC DSCR product, and we got a whole bunch of new products coming out, the full spectrum of non-QM and essentially these investors can design their own product. They can decide one specific investor. Think about the level of control and disintermediation. I could be an investor with a million dollars and I could say I want to invest it in Florida on condos only, which is probably a terrible example. I want to invest it in Florida, condos and a 12-month bank statement and he can tell the AI this is what he wants and that’s the product that his money will go into. Just think about that. And then he knows that product, the product that his money will go into, right, just just think about that. And then he knows that that product has been properly underwritten, that everything’s done exactly the way.
[David] And that’s so much of what happened in the last cycle, when we had the housing crisis. Many of the products were created, but they were not underwritten to the guidelines.
[Pavan] It happens all the time.
[David] Yeah, you and I have lived this. We’ve lived this for a long, long time. So talk about SunWest Investment Trust. So is that what’s drawing in the capital from Japan?
[Pavan] Right. The money goes into the trust. The trust acquires the mortgages that the AI generates and gives to the trust, and the AI only generates what the trust directs the AI to generate.
[David] Wow, so you are creating a new Fannie, Freddie literally is what you’re doing.
[Pavan] It’s like that. I would say it’s a new form of we’re creating a securitization market without the securitization overhead and maximizing the yield to the investor. The securitization overhead and maximizing the yield to the investor. I mean, let’s face it, the whole point of securitization, the whole point of rating agencies, is that, this is too hard for an investor to understand how a mortgage works. Okay, so I’m going to rely on the triple A stamp or the double A stamp for the agency.
[David] And we saw how well that worked in the last housing crisis.
[Pavan] And it doesn’t work right. They have no skin in the game. They don’t, they don’t care, these want to get their fees.
[David] Yeah, okay and like the traders. The traders are doing that. They trade. The traders are doing that. They trade. They’re not in there guaranteeing or with a care of. Are you really doing that? They get the, the rating agencies to put a stamp on it and they trade it out, and that’s why we’re in. This solves a lot of that because of the intelligence that’s going into this and creating this, because and that’s what’s drawing in the money. It’s very interesting.
[Pavan] Yep, I mean you, could you know the way the’s drawing in? the money. That’s very interesting, Yep, I mean you could. You know? The way the industry in the past or still today solves these kinds of problems is that they buy insurance policies called credit default swaps, and they have different big name underwriters and so forth. And these are all fees and fees to give the investor some assurance that the product is going to perform. Okay, but this is all just, it’s like you have a rotten cake and you put all layers of icing on top of it to make it look good and hopefully you can, it goes down without giving you heartburn, right? That’s what they’re doing.
[David] Yeah, I love what Jay Brickman said one time. Jay’s the old MBA, the chief economist. He’s been a good friend for years and years retired out now. But he says you know what CMOs are and CDs and what we created. He says it’s analogous to a trash compactor. All you do is you dump trash in there and the nice thing is these securities are a trash compactor but at the end of the day it’s still compact trash. It’s trash. And what you created is something that is not trash. It is written to a standard that is absolutely guaranteed because of the technology using. The implications of this Pavan are so profound and what you’re going to be able to do as a result of the SouthWest Investment Trust that you’re going to be creating. But I want to continue on talking about your vision. We could spend a whole time. Go read Crisman’s article. You’ll find it in the August 11th issue of Crismain’s blog. Loved what Rob does there and he creates. He called attention to it, and wisely so. So, I want to talk more about the development, but one before we go there. I want to talk about policy, because you created, put together a one day seminar. My good friend Greg Sher went to it and many others did the U S strategy for AI and crypto. Talk about why you did that. You are becoming a leading voice in this industry for financial services, and what’s going on with AI and crypto?
[Pavan] Yeah, I don’t want AI and crypto to go the way of social media where you have a couple of big, huge monster companies who think they know better and telling us what we should and shouldn’t say. So, the same thing with AI and Crypto. I mean, if you want free speech to be controlled, if you thought it could be controlled by social media, imagine what would happen with AI. Okay, it completely take over. So the point of the conference is to say and this was right at the launch of crypto week for congress is to say, hey, let’s put in. Yes, crypto needs some regulations. Let’s put in, yes, crypto needs some regulations, but let’s be smart about the regulations. Not do it based off fear. Right, then you get bad rules. When you make rules out of fear, they become bad rules. and bad rules create monopolies and then monopolies then control your life. So what we want is good rules that spark American innovation, that spark American businesses, that decentralizes, that creates a lot of businesses doing and competing with each other, and not centralized monopolies. And when you have a lot of different people doing different AI and controlling their own AIs, then you have freedom and free speech, and then that’s where AI and blockchain need to combine together, because, ultimately, AI is only as good as the data that it’s given, and data is free speech. If you don’t control your data, you don’t have control of your speech. If someone else can tell you what is in your personal vault, what information represents you, someone can curate your information. You’ve lost complete control of your free speech. Okay, so that’s what this whole conference, was about, is what the Angel AI technology is about is so that you the data that defines you and who you are is personal and it’s in your, it’s in your data vault and only you can control it. No, no third party, no government, no corporation can decide what goes in your data vault.
[David] Again, implications of which are profound, what this could be and there’s others that have been working on this project, but no one quite from the same angle that you have been taking on. So kudos for you setting up the US Strategy Conference. I know you’re going to be doing more of those and we’ll put this out to our listeners. But go register. Go check out the U-S-S-A-I-C dot com. You can register there.
[Pavan] Well, actually the conference is over, so we you can’t.
[David] That link is done, okay.
[Pavan] No, no, but you can go to USSAIC dot com and all of the content, all of the panels and every conversation that happened there is on the site is published, Okay so you can listen to it, to all of it, so it’s free resource for the world.
[David] Oh, it’s good, All right, so very good. Greg Sher, who is there with you, just couldn’t say enough about it. He called me. He says Lykken, you’ve been telling me about Pavan, you’ve been telling me about Angel AI, and I respect you, but it’s not until I went there did I get it and then he sat down and had breakfast with you. His mind exploded. He says this is changing our industry. This is not just a little shift, this is a tectonic shift.
[Pavan] It’s not just the crypto week and the stable coin bill that passed is called the Genius Act. This is not just changing the mortgage industry, it’s changing the banking as we know it In five years’ time, if it even takes that long. Most community banks are out of business. Most banks are out of business. They don’t exist anymore.
[David] I’ve got a number of community bankers that are listening to my podcast, that I know of that are going to hear that and go. What did you just say? I’m thinking of Carol, who works at one of them. I know she’ll hear this, so I mentioned her name. So she’ll be retired by then. But are you serious?
[Pavan] I’m 100% serious. In fact, if you’re a community banker listening to this, write a thank you note to Elizabeth Warren because she extended it, you would have been done in a year but because she made sure the law had a provision in it, so now you can live a couple more years. okay. So basically the law, as Senator Hagerty and Senator Tim Scott had proposed it in the Senate, was that it would allow the stable coin issuer to pay you interest on the money that you deposit with the stable coin issuer. Okay, so that would would if that had, if that clause. Elizabeth Warren said, ok, that’s going to go. Ok, that was a compromise and the bill passed ok. So when you remove that clause, ok, this was lobbied again. The ABA American Bankers Association lobbied really hard to try to kill this bill. Ok, so the compromise was that that piece was removed. Because imagine, if you can just buy stablecoin and get paid 4%, 5% interest by the stablecoin issuer, why do you need a savings account? Why do you need a checking account? Community banks are gone. All the deposits would exit community banks overnight.
[David] Yeah, very, very interesting Again. I want to continue this. I’m planting the seed so people can start to be aware of all that we’re talking about. Go ahead.
[Pavan] There’s some massive global implications. Let me tell you why the stablecoin bill was so important to America. Because what it does, though on the flip side, yes, we save the community banks, but I think the greater good for the country is saving the US dollar. The US dollar is under attack. I mean mostly for our own doing right. We loaded it up with, you know, 35 going heading towards $40 trillion of debt. Right, it’s kind of makes a dollar questionable, right, when the debt to GDP is blown way past 100%? And where are you going to get $35 trillion to pay back these treasuries? So, the whole value of the dollar is being bought into question right now and now with the stable and then you got BRICS that have created their own payment rails, separate from SWIFT, separate from the dollar, and they’re trading with each other outside of the dollar. That represents, you know, three quarters of the world’s population. So you’re really now, okay, at risk of losing the dollar as a reserve currency. So, but with dollar on chain, stable coin basically says the dollar as a reserve currency. So, but with dollar on chain, stable coin basically says the dollar cannot-.
[David] Yeah, so that explains. I wanted you to explain this. So it’s taking the dollar, the US dollar, as we know, the greenbacks, and putting them into a blockchain and that’s called stablecoin. That’s going to allow our currency to be the dominant currency.
[Pavan] Well, what happens now? Because it’s on blockchain and it’s on wallet, anybody, any individual, bypassing all the institutions, all third-party government controls, any person anywhere in the world can own the US dollar. Okay, and so most developing countries have currency controls, you know, Brazil, India, China, they all have currency control. Russia, they all have currency controls. But now those specific citizens, okay, and I don’t know whether it’s legal or illegal in those countries, but those specific citizens can just, on their cell phone, get a blockchain wallet. Hopefully it’s an Angel wallet, because we have a blockchain wallet built into Angel AI. Get a blockchain wallet and go buy some stable coin, and now they own the US dollar. So forget about central banks owning US dollar. Individuals will own US dollar. Will this do central banks owning US dollar? Individuals will own US dollar.
[David] Will this do away? with central banking as we’ve known it.
[Pavan] I think, if any, I think, this is a big challenge to central bankers everywhere else. It’s a huge challenge and I don’t think they know how to. I don’t think they figured out what they’re going to do, how to combat this, because, let’s face it, it’s still America and our number one export is culture and rule of law and trust and faith. Right, so everyone wants to be long the dollar at a grassroots level. Yes, everyone wants to be long the dollar.
[David] This allows them. Stable coin allows them to do so, it’s fascinating. I’m going to put in a plug for a book I just read a couple of weeks ago and it’s called the Tower of Babel and it talks about Babel, Switzerland and the bank that runs the international settlement, the Bank of International Settlement, BIS Bank of International Settlement, which all the central bankers belong to, and it’s this quasi behind the scenes what controls all of the money and everything’s about ready to change dramatically. So fascinating.
[Pavan] There’s no need for international settlements when you get the US dollar on a stablecoin. Right, it’s like it’s, so I don’t even know what the Euro is going to do. Right, what’s the EU going to do Right. The tax rates are so high. It’s such a mess in the economies over there. Wealthy people are leaving and want to want their money out of Euros as fast as possible. Easiest way put it into stablecoin
[David] Yeah, it is, again, implications how long before you could actually do this with stable coin? Can you do this now?
[Pavan] Well, the laws passed because tether was grandfathered in, so USDC and USDT is already there.
[David] Okay right very interesting, allright. Well, I want to continue to move on because there’s so much in the way of computer science developments that are going on. You introduced me to a guy by Yan, and how do I say his last name?
[Pavan] Yann Lecun
[David] Yeah, this is really interesting. He’s the recognized as the godfather of deep learning. Talk about this guy and what he has recently refocus, or his focusing on now. Talk a little bit about that. This is important listeners to pay attention to this.
[Pavan] it’s his research that has led to modern day LLMs and ChatGPT, Gemini and all this stuff that you use, and he was recently at a talk, I think like two weeks ago, and he said, I think he said something to the effect of I’ve moved on from LLMs, they’re done.
[David] We’re just moving into it. He’s already moved on. These are done, wow.
[Pavan] He says it’s now a commercialization, there’s no more. There’s nothing more you can do with them from a research perspective. Now leave it up to the industries to commercialize them and get it out to work. It’s not my problem. So he’s moved on to creating true reasoning AI, an AI that works more like the human brain than what LLM is a step forward in that research towards the ultimate goal, which is what they call singularity, which is a truly thinking AI. So put it in perspective. A human child by the time you know he’s five, maybe he’s exposed to you know a few hundred thousand pages of of text, right, If even that, you know how much? How many words, how many sentences has he actually ever heard or interacted with by the time he’s five? But by the time he’s five, he’s speaking cognizant and speaking about complex concepts. But an AI? If an AI was exposed to the same amount of text, it’s not going to be able. It’s just going to create gibberish. If you train AI on such little amount of text, it’s going to produce very little output, if anything meaningful, okay. So that’s where Jan is looking at it and saying look, if the human brain can do it with such little data, okay, if we’re really going to make a reasoning machine, we need to make a machine that could also do it with such little data. And so when you really dissect it, it’s how does the human brain do it versus how does current AI systems do it? Well, the difference is that the human brain finds broader patterns in the information it is getting and it trains itself on the broader, large patterns and not on the specific word by word. So it doesn’t need as many words, it just needs enough patterns.
[David] Okay, so concepts Interesting. So where are we heading with this? What does the future look like?
[Pavan] Well, I’ll put it this way I mean your brain, that’s sitting on top of our shoulders, that it uses I don’t know 10 Watts. To forget something, something tiny amount of power like that, right, right, and microsoft’s uh, uh, ai systems. Are he the bill gates wants to buy an entire nuclear power plant, he wants to reactivate three mile island to power, you know, millions of gpus, hundreds of thousands of gpus. Right to do, right, okay, but it’s not a fair comparison. Right where, because it’s still a machine, and a machine can retrieve and restate exact text from anywhere. A human brain can’t do that. So, yes, there’s a difference, but it’s not thinking and the human brain is thinking.
[David] Right, yeah, so he’s creating something that we think is what you’re saying.
[Pavan] Right right, he’s creating something that will be thinking is what you’re saying, right? Right, he’s creating something that thinks. The ultimate goal is you’re getting closer to thinking right when you’re dealing with concepts and principles like that and you’re using less power and you’re getting closer to thinking. You’re getting closer to the human brain, the less power that you use. The more power that you use, you’re getting more of a machine, just a lot of fast decision-making machines.
[David] Well, I’m going to fast-forward into what also came out in one of those conversations this weekend when I was with Brian and his team. There’s a guy by the name of Ryan there and he says Dave, I really think the day is coming where loan officers are not going to be needed. And he says I want to continue to make as much money and prosper during this time. But I really and I totally disagree to them. I said that’s not true. How the loan officer works is going to be what’s changing? Not that the loan officer is going away. You’re the creator of Angel AI. You’ve already done so much. What is the future for our industry? The various positions with our industry. Let’s start with loan officers.
[Pavan] Yes, if you keep doing things the way you’re doing it right now, yes, you’re going to get disintermediated.
[David] Disintermediate is a fancy word. It means you’re out of a job. Yeah, so that is. So it’s requiring us to change, but how are loan officers going to have to change you?
[Pavan] Have to think if you’re doing three, four, five loans a month, right, you have to change. You have to think after doing three, four, 500 loans a month, each loan officer, you have to think differently, right? So it’s about relationships. It’s about getting your message and your brand out and connecting, and that’s why we created Angel Twin. So the Angel Twin is a AI replica of you and that is doing most of the heavy lifting, but you still have to be there and get out in front and promote what you’re doing and communicate, because you’re, as a loan officer, you’re an expert communicator. You’re an expert community builder.
[David] Right. If you’re good, if you’re good as a good loan officer, that’s exactly right yeah.
[Pavan] And you’re an expert relationship builder. Okay, so you’re going to build relationships, you’re going to build communities, you’re going to, you’re going to gain trust with the communities and when, and, and then, when someone actually needs to do a transaction, you let the AI do it.
[David] Yeah, yeah, and the people it craps their brain on that point and that part is changing. So he’s right in the sense that the job is changing, but at the end of the day, there’s still a human element that is desperate because it’s the most important transaction of anyone’s life at the end of the day. I want to look Pavan in the faces and but what I’m doing here make sense and, yes, you could have, as they said. Well, you just said that it’s going to have human reasoning very shortly. It’s going to have emotions. There’s still, it’s still a machine. At the end of the day, humans want to talk to humans to get the validation done and they’re just going to be able to do it out of fact. But you say, are you using Angel Twin? So let’s talk a little bit about angel twin. We’ve talked about it before. It is a replica. I just you sent me before we did this uh recording. Today you sent me two videos and it was sam who was holding the phone. I think it was Sam that was holding the phone and talking to. By the way, Sam does the best mindset conferences anyone needs to go to. They need to start going to Sam’s conferences. Kudos, Sam, for what you’re doing, and you teamed up with him early on and you find a way to bet on the right group and the right people. Pavan, that’s just so cool. But anyway, Sam is looking at, is holding an Iphone or an Ipad or something like that, his phone, he’s holding his phone and he was asking Angel Twin some questions, it one of which was what am I wearing?
[Pavan] Yes, it’s super cool.
[David] And he responded back with the action, the camera analyzed what Sam was wearing and came back and brought it back. Then he says what is behind me? And the guy you could, you could tell the process was system, the avatar, whatever you say was moving his head around. So he could tell you was thinking. And then he came back and said you’re, you’re sitting in front of the wall. Behind you is a wall of a really nice, and he described it. It was in a hotel ballroom, it looked like, or something like that, someplace in Las Vegas, and it was described it. You have a light, it has light that’s shining up. It was amazing, Pavan, and these are the things that scare a lot of people in the industry and they go there, you go like and you just said we’re not going to be neat where you were going to be needed. But look at this if you have technology that looks like me, talks like me, that could do this, the answer is who’s running it? and the answer is you’re running it.
[Pavan] It’s you and this is where the whole blockchain part comes from is your data? Everything about you is stored in your wallet on chain and only you control. Everybody has their own AI. Okay, now this is really exciting because this is so I gave you, David. I think it’s the first time I actually gave you a sneak peek on what’s going on in the lab. So you got to look under the hood first person outside the company that’s seen stuff like this. So there’s a big NARAP, National Association of Hispanic Realtors. Their annual conference is coming in September 18th, 19th in San Diego. I encourage you all to look it up. We’re going to be there. We partnered with NARAP and we’re running their AI and crypto parallel track for two days. Okay, we got some really amazing, just world-renowned speakers coming to that track top of the field when it comes to AI, crypto and real estate and tokenizing real estate. We got some speakers who are at the leading edge of tokenizing real estate and tokenizing real estate. We’ve got some speakers who are at the leading edge of tokenizing real estate Because, remember, the SEC right now is working on the final rules of tokenizing real world assets. Okay, when those rules are issued, then there’s the end of REITs, because everything will be tokenized. It’s so expensive to set up a REIT Could you imagine how hard? how much? money goes into setting up a REIT. Now, you don’t need to. You could just tokenize it and off you go.
[David] A lot of people are listening to that and go. I have no idea what you’re talking about, but tokenization anyone’s been in the industry any amount of time knows what a REIT is a real estate investment trust but they’re going away because of the tokenization.
[Pavan] Token, tokenization and what it does is because tokenization is so simple and low cost. Usually you have to have a minimum size couple hundred million dollars minimum to do a REIT. Now tokenization I could be a small investor with two rental properties and I could tokenize those two and create my own little mini REIT and start trading it.
[David] Yeah, again, the implications of everything you’re built and building, Pavan, are so profound for our industry and beyond financial services and beyond. It’s extraordinary, and I’m looking at the time that we’ve already spent on this. I’m going like we’ve got to wrap this up. But, Pavan, I want our listeners to continue to press into this. Don’t be afraid of it. And you and I you send me so many things. We’re on WhatsApp together and Pavan, our listeners you can’t believe. I don’t know when this man sleeps. He’s sending things constantly to me and it’s just an ingesting of so much information. Our world is changing, listeners, and you’ve got to get on board. You can do what 15,000 other people did last month it’s go to askangel.ai and start using it, and that’s how you get into it. But this also, you’ve got to start somewhere. You just got to get in, start playing with this thing and that’s how you get into it. But this also, you’ve got to start somewhere. You’ve just got to get in, start playing with this thing and that’s the best way to do that. And there’s a whole lot of wonderful people there. You have your Angelistas all over the San Juan, Puerto Rico, that are supporting that. I was explaining this to someone over the weekend. I said Pavan has Angelistas that will make phone calls for you. He goes yeah, but how much does that cost me? Nothing, he’s doing this. It’s staggering, Pavan.
[Pavan] I get asked the question why do we have people making those phone calls and not AI? It’s because your clients are way too valuable, Because they’re still. If a realtor is going to trust us with his client list, right, just a slight mistake from the AI can blow that up, blow up that relationship, and so we’re still. We want to do a lot more of these and get a lot more data before we think it’s ready for an AI to completely do talking to your clients like completely on his own. And so and this has been a really important part of the whole process is to do it right, to deliver a high quality angel twins to you, high quality AI phone calls to you, right. To do it right is you have to invest the money in the data science and you have to invest the money in collecting the data right. So, with humans doing it number one I assure you have really high quality calls being made and realtors and loan officers that are using this. Their minds are blown by how many leads they’re getting. And, remember, these people are massively efficient because it’s an AI that’s guiding them, so they’re AI-assisted. It’s a real human voice talking to them, but it’s an AI-assisted human, so they’re AI assisted. It’s a real human voice talking to them, but it’s an AI assisted human voice, so they can make thousands of phone calls and so my cost is like almost it’s very, very little my cost. So I get high quality service delivered to you at the very low cost and I’m building an amazing training data set from which I will have the best twin generator, the best AI voice generator and being able to pick up on the subtle inflections and nuances of customers when a customer is cold called, how they respond, how best to react to them. It’s got to be able to pick up on all those subtleties so that you get, you don’t burn any customers and you also have very high conversion rates. That’s the reason we’re doing this.
[David] Yeah, and the core word there is trust. What wins the game is operating with trust, and when people can start trusting a system like this, it’s going to change the game. It’s not going to put you out of business, but it’s going to require you to do business differently. Pavan, thanks so much for joining us again. We’re going to have more conversations and sharing this. That’s what we’re going to start calling these Conversations with Pavan, because there’s so much developing and happening. Kudos to you for what you’re doing. It’s honored to be and a chance privilege of knowing you for now, 44 years plus, and working with you now. What a joy and delight, so exciting.
[Pavan] Thank you, david.
[David] Keep sending it over. We’ll get you back on here and keep educating our listeners. Those are the doubtingting Thomas’s out there and I know who you are. We chuckled with you, but you’re coming over and it’s happening. Keep paying attention to it, Pavan, thank you again for what you’re doing for our industry and for the world.
[Pavan] All right, cheers.
[David] You bet.
Important Links
Pavan Agarwal is a renowned leader in the mortgage lending industry and a pioneer in bringing artificial intelligence to the financial markets. Agarwal serves as the President and CEO of Sun West Mortgage Company and Celligence International, LLC.