 Hello, everyone. My name is Marvin Bantugen and welcome to the May 2023 Hyperledger Financial Markets Mortgage Subgroup Meeting. That is always a mouthful. But before we get started, I'd like to express our appreciation to the Financial Markets Special Industry Group and the Hyperledger Foundation for their ongoing support and making this group possible. Okay, as always, please note that this meeting is being recorded and is under the umbrella of the Hyperledger Foundation. So we ask that everyone abide by the antitrust policy and code of conduct. The antitrust policy states that we avoid discussions of company-specific products and projects and pricing. We don't make negative remarks about other companies or their products. And the code of conduct means that we trust each other, we treat each other with respect. We never discriminate, communicate constructively. We fully support Hyperledger's policy of openness, equity, and inclusion. And for new participants, we welcome you. And if you'd like to introduce yourself in the chat, please do so. And if there's any specific areas of interest, please let us know as well. Here's our agenda for today. James Hendrick and I will, or excuse me, James will provide an update on the developments in the mortgage industry. I'll discuss AI and blockchain in the mortgage industry. I've broken up the AI and blockchain component into two discrete segments, what I'm calling first gear and second gear. The reason will become apparent once we get into the second gear portion and then a Q&A at the end. We always cover the slide in each meeting. This is to reinforce that we're all on the same blockchain journey, but we just may be a different path along that journey. For today's topic of AI and blockchain, this is particularly applicable since for most of us, we're at the beginning of the journey of understanding AI and blockchain. Today's session is part of a series on AI and blockchain that will focus initially on education and in the future sessions, we'll delve deeper into the topic and have speakers come on that are using AI and blockchain and or have more in-depth knowledge of it. For example, my good friend, Mark D'Angelo, has released an MBA column on AI titled, In an AI Reimagined Financial World, It Begins with the Consumer and will include a link to that in the chat. Also, I just want to mention to everyone that Mark, James, and I will be participating in an MBA webinar on May 17th. That's titled Leveraging AI Blockchain and New Technologies in Today's Environment. We'll include a link to that as well and that is a free webinar. It's free to MBA members and we'll provide a code for non-members so they can watch it complimentary as well. Okay, the next couple slides I usually go through pretty quickly. I always mention these for those that are new to the group. This slide provides links to different resources, the Linux Foundation, our mortgage subgroup. For example, it's the second link from the bottom. The mortgage subgroup link contains meeting notes for our group, recordings from previous sessions, and the curated articles about blockchain in the mortgage industry. These are great resources and we'll reference several of them during the course of this presentation. In order to access those resources, you'll need an LFID and this is how you get an LFID. I'm not going to go over it, but this slide is available and the information is available for you to get that ID. Another resource is, if you're new to Blockchain, these are some free resources that are offered by the Hyperledger Foundation. I highly, highly recommend these. I've taken these and they're fantastic. Okay, with that, James will walk us through the state of Blockchain in the global mortgage industry, the status of Blockchain and AI. James, take it away. Marvin, thank you very much. Welcome everybody. Let's go ahead and jump into the next slide. This is a Blockchain discussion, which is usually what we're together for each month, but we did want to talk a little bit about what's going on in the AI world. Some of it ties directly with Blockchain. Some of it's running parallel, but we have been starting to gather up with all of our Blockchain articles that we have available for folks, articles about AI in the industry as well. Do be taken a look at the Wiki. I'll cover that again towards the end, but to jump in today, we've got quite a few articles that I wanted to talk about. The first one from Finesa really talks about low management software. It's evolved immensely over the last decade, particularly when you compare it to the 30 or 40 years before that. It's estimated by 2032 that AI and automation market in the banking industry will be worth approximately a $182 billion industry. AI-powered software can analyze vast amounts of data to identify patterns, trends, anomalies that traditional systems are unable to detect. This provides lenders insights that enable them to make informed financial decisions with the client. This first article really covers three topics. It covers AI in the mortgage industry, talking about risk assessment, how it's able to analyze data from multiple sources, including credit reports, financial statements, even social media profiles, allowing informed decisions about approvals and rates, providing fraud detection, allowing them to detect and prevent fraudulent loan applications, chatbots and virtual assistance. I think you're going to hear that a lot in today's discussions. AI-powered chatbots and virtual assistance, they can provide borrowers with personalized guidance and support throughout the entire loan life cycle. And then predictive analysis, historical data and can predict future loan performances, help lenders identify potential risks, and proactively manage portfolios. This article does also talk about automation on loan management software, about faster loan processing, improved risk management, that personalized experience and better compliance. And then the article does also talk about blockchain. A lot of the similar topics that we've talked about month over month with this group, the benefits of decentralized lending, reducing transaction costs, increasing access to credit for underserved borrowers, smart contracts and how they can automatically, without intermediaries, reduce time and cost of processing loans, and then improve data security from providing a decentralized, tamper-proof system for storing sensitive loan information. So, fantastic article. It does cover a lot of topics, and I highly recommend taking a look. The second one from Unite AI really discusses AI and machine learning in the financial industry, providing various examples. This article, too, does talk about conversational chatbots and how AI-driven chatbots and virtual assistants, they offer 24 by seven assistants, any from everything from account balances to recent transactions, and customers can even send funds using conversational language with these AIs. Customer sentiment analysis. Oftentimes, we have difficulty combining customer sentiment into our platform and our survey analysis. AI is capable of analyzing what customers communicate, can pinpoint even the emotions they're expressing in real time. And these systems can alert lenders, customer service teams, so that they can resolve issues more proactively with the customer. Credit worthiness for a thin file or no file. AIs can give a clear view of a customer, especially when they have that limited credit. And while not mortgage specific, the article does walk through a use case in the automotive industry, demonstrating how AI is used to ensure compliance with CFPB requirements. So, I thought that was fairly significant as we're always constantly dealing with the CFPB. Our third article to open with, this one is how technology is revolutionizing the mortgage industry. Fairly large article. It actually breaks down into eight completely different technologies and sections. Four of the sections are really relevant to what we're talking about today. So the first section on AI virtual assistance, streamlining those mortgage communications, providing that personalized recommendation. The second section is on blockchain technology, talking about enhancing security and transparency while reducing fraud and how it can lower costs in both the origination and the servicing arenas. And then big data analytics is our third section. Excuse me. Talking about informed decision-making, how to analyze market trends to identify emerging opportunities, how to track changes in property values, identify opportunities for high potential growth, and to be able to predict future default rates. And then later on in the same article in section seven, they do talk about machine learning and predictive analytics in mortgage lending. How machine learning can analyze vast data sets to predict the likelihood of loan defaults, repayments, and other critical outcomes. So again, another great topic, fairly lengthy covering a variety of different sections, just so you guys are aware of some of the other things that they talk about, e-signature, digital closing, robo-advisors, and integration of other financial services. Marvin, let's keep moving right along to the next slide. So our next article, this is actually an article we pulled out of LinkedIn. It's by Frederico Tagliani. He's talking about the emerging trends in mortgage industry, is the use of large language models or LLMs. LLMs powered by AI to automate aspects of the loan processing workflow. They're able to analyze data such as customer financial records, credit scores, and identify patterns to predict future outcomes. By automating a lot of this loan underwriting process, LLMs can significantly reduce processing time and improve accuracy. In fact, a study by McKenzie found that the use of machine learning algorithms, including LLMs, can reduce processing times by up to 50%. In addition, this article also goes into blockchain and how blockchain can assist companies by streamlining the loan process workflows and by automating various components such as credit assessment and loan approvals. Accenture found blockchain in the industry can reduce processing times by up to as much as 70%. And then the article also goes on to discuss how open banking can reduce the risk of fraud in the loan industry by another 20% as well. So great article coming out of LinkedIn. Definitely recommend taking a look. Talking about some real-world applications. Our fifth item is Provan Argelal from SunWest Mortgage. He's talking about their AI product called Morgan. So Morgan is a 24 by 7 real estate personal assistant that's accessible via basically a chat interface. She makes buying a house as easy as having a text message conversation with the borrowers. And borrowers are able to talk to her just like any other chat system. And she delivers guaranteed responses within minutes. You know, one of the interesting things I found about this article is we hear talk in the industry about how AI can potentially have this unconscious bias that it influences into their results. Well, one of the things they talk about with Morgan is almost the exact opposite of that. When the Urban Institute did a study, they determined that African-American borrowers in the United States had twice the decline rate of their counterpart white American borrowers. When Morgan runs through the process, it helps eliminate a lot of that additional paperwork of cost. And then it starts to look at all buyers as being equal. So it's when SunWest implemented Morgan and brought rates down to be equal to the or the decline rates to be down equal to the national average of white Americans as well. So SunWest is really touting Morgan and how every American has the same opportunity to buy a home. Within this press release, there is also a link to Havan talking with, I think it was Channel 5 News out of Las Vegas about Morgan. So if you get a chance to take a look, jump into that, the actual interview segments only a couple minutes long. And then our very last one. So for those of you that are regular attendees, you've seen on our Wiki, we provide everything from written articles to YouTube videos to blogs. We want to provide you a wide sense of, hey, where's this information available? So this very last one that we pull up on how AI and data are changing the game for mortgage executives, it's actually a video blog from Trust Engines YouTube channel. Their guest, Steve Brown, who's a technologist, he covers six different topics and how mortgage lenders can take advantage of them. He actually wrote a book called The Innovation Automate or Ultimatum. And within that book, he talks about AI, the internet of things, autonomous machines, next generation networks, augmented and virtual reality, and then blockchain and distributed ledger technologies. The actual YouTube video, the blog that you'll watch, it predominantly focuses on the AI environment. He does really a great job of covering what AI can and can't do in our current environment and how AI at a high level works behind the scenes. You know, AI is such a bleeding edge technology right now that we're really not sure where we're at in the lifecycle of things. You know, are we kind of at the top right now and we're about to plateau for a while and what AI is capable of doing? Or, you know, more than likely, are we just scratching the surface? And we're going to see, you know, a lot more happening in the next several years. If you're a betting person, I'm betting on the second one, but you know, we'll find out as we move throughout 2023 and 2024. His real big message out of this is that AIs are coming. And the key point is learning how to integrate AI with human interaction. Not so much that you're replacing jobs with AI, but you're augmenting human jobs with AI. Human interaction provides that trust and that empathy that current AIs just aren't capable of delivering. And, you know, I've actually used AI, chat GPT, barred a couple others in order to do some analysis and pull some information together. And while it does a fantastic job of pulling a lot of information together in a short period of time, you still wind up having to massage the information that's coming out of it based on the particular tool that you're using and really, you know, again, how young the technology is in the environment. He talks about how the biggest skill of the 21st century is really going to be, you know, about how to use these AIs and how to program AIs, how to take advantage of them in the business environment. He discusses different tools for mortgage lenders to check out everything from copy AI, Jasper, chat GPT to barred. So if you do have an opportunity, take a look at it. It's about a 45 minute long video. Marvin, let's go ahead and move on to the next slide. Yes, and just, you know, another reminder of the Wiki site out there, all of the articles that we've talked about today, you'll find accessible on the right hand side. Alma just dropped the Wiki link into the chat. So feel free to click on that and make yourself a favorite. Also over on the left hand side, you've got recordings from all of our previous monthly sessions for the last little over a year and a half now. In addition, over on the right hand side where I've got the articles, I keep about three or four months of the most recent articles that we've reported out on. You can find over on the left, there is a link for previous mortgage blockchain articles. You'll find all the stuff that we've covered over the last year and a half since 2021 within that section as well. And then we've curated a whole list of other articles that we've got available should anybody be interested in, you know, looking at different use cases companies that are doing, looking to build a, you know, an argument for why to move in this industry or to just in general, get yourself educated. And then lastly, in the upper right of this page, everything Marvin talked about, about how to set up a LFID in order to log into this site. You've got links directly there that'll show you how to do that and then how to get added on to the meaningless. So I will go ahead and take my lead there and pass it back over to you, Marvin. Thanks, James. That's great information. I love getting up to speed on what's going on in the industry. So let's now talk about AI and blockchain. As I said at the beginning, I've broken this up into two gears. Let's go into the first year. The purpose of this presentation is to provide an overview of what AI is, its risks and its uses. And then we'll talk about how AI can work with blockchain in the mortgage industry. What is AI? AI is the ability of machines to form tasks that typically require human intelligence, such as learning reasoning and problems all. AI can be divided into two types, narrow and weak, which we get AI, which is designed to perform a specific task and general or strong AI, which can perform intellectual tasks that normally a human performs. There's also a wide variety of applications. And James alluded to some of these. We have icons for all of the key ones at the bottom of this page. I'll only go over a couple of them. The leader that we've all mentioned, chat GPT. Next, there's BARD. Some of the internet connected AI applications include Microsoft Bing AI, Perplexity, UChat, Koala Chat. If you wanted an AI for content writing, there's Jasper Chat. Chat by copy AI, Chat Sonic. If you wanted something for messaging, there's personal AI and there's more. This is just a sampling of the different AI applications. We've all seen some of the dystopian future where AI has run amuck and turns on mankind. We're all familiar with Skynet, Hal 9000, Winter Mute, Ultron. These are all just examples of AI gone bad and I can see James smiling on that. If someone knows the reference to Winter Mute, pop it in the chat and you get extra points. These are just a few examples of AI gone bad. Aside from potentially bringing the apocalypse, other risks of AI can exclude and include some of the things that we've listed on the page. Job displacement, James mentioned that, privacy invasion and abuse and bias. James mentioned some of those as well. We're going to come back to this slide later on. Despite the risks associated with AI, there's still a myriad of uses and benefits. The possible uses span a breadth of fields such as healthcare, transportation, education and entertainment. AI can help diagnose diseases, improve traffic flow, personalize education, and help develop software programs, games, and so on. It can also be used for scientific research, environmental monitoring, space exploration, just a lot of things that we can do with AI. Specific examples of AI applications, we covered them on the previous page. They can perform virtual assistant tasks, language translation, self-driving transportation. Really, the thing that I want to stress on this slide is AI is still in its early stage and its future is uncertain as James alluded to. If I were to be melodramatic, I would say that AI could either lead us to a utopian society where machines do all the work or a dystopian society where machines control humans. It's going to be up to us to determine how AI is developed and used. If you feel that the preceding overview of AI was too light or mediocre, guess what? I use chatGPT to write that presentation. This is a product of AI, AI and blockchain first gear. I wanted to demonstrate what chatGPT can do, how to leverage it, and its limitations. But note that the graphics for the AI slides were generated by an AI application and I thought they were fantastic. The text supporting the slides were generated by chatGPT and although I did some editorializing, I tried to abide by the original language as much as possible. Now let's shift into second gear and let's do a deeper dive. Okay, as I mentioned, this is going to be our second gear. We're going to start off with just AI or chatPT primer. Just to get us all up to speed. Granted, there is more than chatGPT, but it's the most popular, so we're going to focus on that. As James mentioned, he's used Bard quite a bit. I haven't. So James, if you want to include any comments as we go through this on Bard by all, by feel free or anyone else that's a part of this meeting, by all means feel free. The purpose of this is for discussion and get everyone up to speed. Okay, chatGPT, that stands for chat generative pre-trained transformer. It was developed by OpenAI and released in November of 2022. It's built on top of AI's GPT 3.5 and 4, which are foundational large language modules. James has already used that term. LLMs, such as GPT, are trained on vast amounts of text data. For chatGPT, for example, there are 176 billion parameters. That is a ton. It uses a training data set of 570 gigabytes of books, article websites, and others. But the data set used only went to 2021. Okay, keep that in mind. We're going to come back to that because that's pretty important. Okay, also, chatGPT, in how it answers its questions or queries, doesn't use the internet to locate answers unlike other AI assistants like Siri or Alexa. It constructs a sentence or response word by word selecting the most likely token or word that should come next. So, that's predictive. It's based. It uses this prediction in its training. In other words, chatGPT arrives at an answer by making a series of guesses. That's the most simplistic explanation I can provide, which is part of the reason why it can give wrong answers and present them as completely true. Also, chatGPT uses a combination of what's called supervised learning and reinforcement learning to fine-tune GPT and we'll dive into those in the little bit as well. It's a little bit as well. And then the last point, basic chatGPT is free, premium $20 a month. Okay, so on this slide, I just wanted to talk about some of the chatGPT considerations as I've been using chatGPT. And most of these considerations are applicable to all AI applications. So, James, again, feel free to include any commentary. First and foremost, as a user of chatGPT and other AI, learn how to do prompts. Okay, those are queries. Write a prompt like how you talk to a person. Writing a prompt is more than just asking a one-sentence question. It involves providing relevant background to provide context for asking a query. For example, when I first used chatGPT to create that presentation on AI, I wrote a prompt that said simply, quote, create a presentation that explains what AI is, its risks, and how it can be used. And if you thought the presentation that I just gave was simplistic, that was even more so. It produced an overly simplistic presentation. Then I rewrote that prompt to be, from the perspective of an AI expert, create a presentation that explains what AI is, its risks, and how it can be used. This produced a more detailed and informative presentation. But, and this is a big but, if a question requires a deep level of expertise, chatGPT may not be able to give a solution or answer that meets the standard of a specialist in the field. So, we just saw an example of that. That initial presentation, although informative, it's not going to be deep and it's not going to be at the level of a specialist. Also, chatGPT can generate wrong answers. And for heaven's sake, don't ask it to do math. Okay. I've done queries online about what chatGPT, almost everyone says, don't ask chatGPT to do math. Okay. We talked a little bit about the data set that chatGPT uses and the fact that it's all limited to 2021. Recall from the earlier presentation that chatGPT was released in November of 2022. And if you go to that presentation that I just made using chatGPT, it made no mention of chatGPT in its discussion of AI. The one slide that included discussions about the different AI applications, that was me editorializing. I added that additional content. The original presentation made no mention of chatGPT, BARB, or any of the other ones. So, if you're doing anything that is time sensitive, time specific, definitely keep that in mind. If your query requires the use of any critical data or events that is post 2021, you will get an incomplete or even an incorrect answer. Also, chatGPT gets its information from human content. So, think about this. That means it includes all of the fallacies and foibles from human content. It includes all of the websites out there, but it also includes Reddit. It includes Wikipedia. And it includes 4chan. And for those of you that have been on the internet, you guys know a lot of the errors and information, negative content that's on Reddit, 4chan, and sometimes on Wikipedia. So, that's included. So, keep that in mind. Okay. We are keenly aware of information security and session management as you use chatGPT. Your conversations with the chatbot are used to improve and train the AI and improve future responses. That's the training that I mentioned on the earlier slide. The open AI privacy policy demonstrates that the service saves pretty much everything. Okay. When you type information into a query or use additional information to write a query or prompt, this includes personal information you provide, conversations, personal information collected through social media, log data or metadata, including your IP address and browser, what device you're on, and more. So, it uses all of these. Keep in mind that most websites have similar policies, but the risk is even greater at chatGPT. So, remember when you're interacting with any web service, including chatGPT, you should assume that any information that you provide is going to be public. Nothing is private. Okay. So, let me state that again. Any information you share with chatGPT to generate a response is saved by chatGPT and then used to train chatGPT and potentially answer future queries from other users. Okay. I'm going to get off my soapbox. Hey, Marvin, if I can jump in with a couple of thoughts here too. I think you've hit some really key things. It's data and data out. So, the responses you're going to get are only good as the prompt that you're putting in, and it's only good as what data that the AI is feeding off directly. A couple examples of this for the group we played around several weeks ago with taking chatGPT and barred AI and feeding them the exact same questions and getting very different responses. One would generate a response that's three or four paragraphs. The other one generated a response that was two or three lines. And it's really a matter of, like we said, once you generate it, you need to get in there and you need to massage it more. But it's also based on the question you asked. It's one thing to say, hey, what are the benefits of AI? Start with that in an AI generator and then add to it. What are the benefits of AI in the mortgage industry? What are the benefits of AI in the mortgage industry to help reduce risk? And you'll see how the answers start to change as you get more specific in your question. You are going to get more specific out of the answers. As Marvin mentioned, I played with barred AI a lot. We do a lot of Google shop stuff with our clients. Barred AI is still really in a beta environment. If you want to check it out, I think you go to barred.google.com and you can sign up to be a part of the beta test program. I signed up a little over a month ago. It took under a week before I got my access, but now I'm able to access it via my phone. I'm able to access it via my computer. Some of the things that are a little bit different about barred, so one, barred is free. Barred, when it produces your response, it gives you three variations on the same response. When you go through those three variations, it's basically all the exact same content, but one might present it in a bullet point format. The next one might present it in just a paragraph format. The key thing is to get out there, start playing with them. I envision at some point you're going to see barred directly integrated into the G suite of applications, much like Microsoft is working on a system called co-pilot. Co-pilot ultimately is going to be integrated into Microsoft Office, so you can access it via Word, Excel, access, outlook, pretty much anything that you're looking for. Thank you for that, Marvin. Oh, no problem. That's fantastic information. Let's continue on. Hopefully, I haven't completely turned you off from using chat GPT or any other AI application. I'm sure you've all read about the excitement around AI and chat GPT, and that's why you're here today. In my personal opinion, this excitement is well founded. I do think AI will change the world, and it seems that everyone is jumping on the AI bandwagon and dropping it into the marketing or pending it through their name. I included a quote here from Jenny Johnson, the CEO of Franklin Templeton Investments. AI and blockchain will be the two biggest disruptors to any industry. I don't have a crystal ball, but I do see the potential synergies between AI and blockchain, and I think those synergies can benefit mortgage companies. Let's review some of those synergies on this slide. For example, one of the most consistent criticisms of blockchain is its processing requirements and transaction speed. AI can be used to detect network congestion and empower the blockchain to adapt dynamically and maintain peak performance. You can use AI to bolster network integrity as well by detecting and preventing fraudulent activities. I think that's definitely one of the key advantages of pairing AI and blockchain. Also, you can enhance smart contracts with AI. You can adjust smart contracts, allowing them to adapt to shifting market conditions and assimilating new information. You can use AI to evaluate and interpret smart contract terms and independently evaluate and potentially resolve issues before contract execution. AI can also vastly enhance the decision-making process for smart contracts by capping into predictive analytics. This is something that James mentioned as well. AI can scrutinize vast data sets, identify trends, patterns, and potential risks that could affect a smart contract's outcome. I think that the potential synergies from AI and blockchain are just being identified. I'm sure as we go through and companies start to pair the two together, we're going to start to see more of the potential of the two technologies. How is AI being used with blockchain in the mortgage industry? Here are three specific examples that I've identified. The first one is Redfin and Zillow. I'm sure everyone has used one of those websites to do a property search. You enter your criteria for a home or property, and these websites generate a list based upon that criteria. By using a chat GPT plugin, these websites can identify and suggest homes and communities that would not have been included in your search. If you have a very definitive criteria, it will provide a more expansive population by taking your interests into consideration and potentially identifying a home or community that you didn't even consider. I think this is a really interesting use of chat GPT. There's also Big Purple Dot. They are a mortgage real estate customer relationship manager management solution, and they've integrated chat GPT into its marketing platform. This allows chat GPT to interact and streamline the sales process by using natural language processing. In other words, their AI assistant understands questions from clients and responds in a more human-like fashion. One of the things that James mentioned was using AI to provide quick accurate responses to queries, and this is exactly what VPD has done. They could provide quick accurate responses to clients so that loan officers, real estate agents, and recruiters can spend more time building relationships and closing more deals. I haven't actually seen this, but I've reached out to Big Purple Dot to get a demo, and I'll see if I can get them to attend one of our future sessions. Our last example is a company called Intane. Intane is a blockchain-enabled structured finance platform. They use AI and blockchain to provide an automatic integrated solution to connect all interested parties involved in a structured finance. This includes issuers, underwriters, verification agents, auditor, servicers, and so on. Starting with servicing conventional loan securitizations, Intane partnered with WSF, the fourth biggest trustee in the MBS environment in 2020. Currently, more than $6 billion in assets have been administered using their Intane admin tools. I think this one is really interesting. I hadn't heard of this one before, but this is really interesting. I'm sure people in capital markets would be interested in taking a look at that or researching it. Okay, I know I went through that information quite a bit very quickly, so let me pause there and see if we have any questions. Okay, I do see a question in the chat. What is the expectation on the decrease in high cost to the mortgage customer from AI slash blockchain technologies? Jeff, that is a fantastic question and really cuts to the core of why we're here. I don't know. I'll be absolutely frank with you. As I speak with different technologists, right now the expectation is using blockchain and loan doesn't necessarily decrease your cost, but it does cut down on some of the additional steps. For example, Redwood Trust has shown that using blockchain, you can take out 100 basis points in audit-related and re-verification expenses in securitizing a mortgage. That's one example. There is no data available yet on pairing AI and blockchain technologies. I think people are still exploring that and as soon as we do get any of that information or even a lien or a hint and we'll definitely share it with the rest of you guys, because that's why we're all here. Let's find out what AI and blockchain can do. We can't answer that right now. Yeah, Marvin, he posted a follow-up question as well, Jeff, regarding the decrease in the length of time to acquire a mortgage. Very similar, Jeff, while there isn't definitive studies out there as of yet, the areas that are working within the blockchain environment at least have been touting everything from a 50 to a 70% reduction in the processing time to set up and fund a mortgage. Again, as we continue to see more develop in the industry, more companies getting involved will be gathering up information like that and bringing it to this group. Okay, thanks for that. This is Jeff. I think those answers are good. Thank you. Absolutely. Yeah, and one of the things that we're trying to do as well is there are a couple LOS, it's important to sales systems, that use blockchain and I've seen a couple demos where they have significantly increased the length of time to process a mortgage. There's one company out there called Vesta. We actually spoke with that company at the mortgage, the MBA technology solution. They are a really interesting solution. We'll see if we can bring them on board. Vesta, V-E-S-P-A, they're a really interesting tool. They said they were able to significantly increase the amount of time to process a mortgage. I don't have any of that data, but we'll see if we can dig that up. Okay, any other questions from the rest of the group? I see that Mark D'Angelo is on here. As we mentioned earlier, he has released an article on AI through the MBA. In Mark, did you want to speak a little bit to that article and the series of articles that you're producing for the MBA? Sure, thank you. I appreciate the opportunity. This is going to be a series over the next probably another two to three months. There'll be about four parts to this article, but the first one was released really talking about focusing AI, not from the technology, not from what do we do with it or the blockchain, is the strategy and understanding how does this impact your customer? To take Jeff's point, let me throw a camera up here if you can see my ugly mug, the real challenge will be is what are those decreases? Marvin, I think these new solutions will actually decrease that amount of time. It may automate the complexity, but if you look at the idea of what this article that I released and where the industry is, we're still talking 45 to 75 days to get a mortgage from the customer contact to funded to ready for the capital markets. If you look at that, could AI and blockchain create immutable assets that are more valuable to the capital markets on the back end, not just the GSEs, but I think that will be the real challenge. If we didn't have some of the regulation, there are some theories, there are some academics that would argue if they were to reimage the supply chain of the mortgage processing that we could take those 45, 75 days and probably make it under 15 definitely, but some people are even advocating for a week. But again, 15 pretty much is the regulatory burden. There are compliance issues, there are time differences, so there are some things that have to change and this is where the regulators come in even though the technology can do it. And that's really what this article is all about, understanding the customer, understanding their pain, understanding the industry challenge, and then mapping that into the future ecosystems of what our data is telling us and what we can do with it. So it's a different look that is not technologically focused because again, we talk about the different types of blockchains, we talk about the different types of AI solutions, whether they're generative, whether they're using an LLM or if you're, it's kind of funny, chat GPT and open AI, they would advocate that the days of LLMs and chat GPTs is done. They are saying release four, maybe release five would be the last LLM that would ever be issued that the real focus and value of something like a chat GPT, the generative AI discipline, would actually go into industry clouds and systems that are focused, maybe they're enhanced by something like the MBA or different industry groups, that are ring fencing these assets that are more industry focused and that technology, those solutions, those disciplines that capability is only applied to those industry realms, not the LLMs that we think about from tell me how to make a souffle or something silly like that. They're not going to be very large, they're going to be industry focused. And as examples of this, we're seeing this already from folks like Oracle, we're seeing this with their CERNR acquisition and since they have that software, they're creating LLMs that are focused on an industry basis. If you can call them an LLM, they're very specific, they're very pointed. And I think that's where this whole idea of AI is coming to be. James, Marvin, you actually talked about some of the vendors. If you look at the, since I worked with private equity and VCs, the funding last year alone, 2022, was nearly $30 billion for AI, generative AI solutions, and was spread across nearly 260 different companies. And every week, every week, there are nearly two dozen applications that sit on top of chat, GPT, that are introduced for a commercial offering. And so you start to add this stuff up, you start to look at where this is going. And what this may mean for a mortgage market is that the idea of how we did FinTech, how we did RegTech. And again, FinTech and RegTech in the last, what, 15 years, guys, about 15,000 different solutions. How are you going to make sense of that? And we cannot sit in a world where we have 15,000 AI. So this is some of the more fundamental questions leaders have got to ask themselves is how do we deploy this? To your point, AI is not commercially off the shelf. It's not like a FinTech solution. These are truly customizable, learned type of solutions that can typically go bad within 12 months. 85% of the AI implementations in 12 months have no validity after they've been implemented. This is some of the challenges that data is driving is digital information, new ecosystem. So it's a very, very complex discussion. There is not one size that fits all. This is not a prescription module where a vendor comes in, maybe you don't want to mention the bigger vendors, but we all know who they are. You've got all the software you've bought the Indiana across the mortgage supply chain. What do you do with it? AI is going to completely disrupt that. And I think that's where, you know, if you look at the article I started as saying, this is all great stuff. We could have lots of good discussions on it. We can go on for days, but it begins with the customer and it's knowing what your customer is, how to reach them, how to make sure that what we're producing for them is not biased. It's not, you know, we're not using training data for our AI solution that is going to cause non-financial inclusion. You know, we talk about financial inclusion, but to restrict people to only give, you know, good interest rates or good solutions to people that really don't need it anyway. AI is more about the inclusion. And it's also about revenue generation. So, you know, there are so many different things here that it's going to take a strategy of AI. It's going to take building blocks of AI to create the mortgage industry's kind of next iteration. And that's what this article series is. It's taking a broad brush, giving some specifics, and asking some of those probative questions saying, before you start to implement something and then find out that your ROIs aren't there, what are you really trying to do this? Where's your customer? Where's the industry going? And how is it all playing with the digital ecosystems that are coming to be? That's more than you probably ever wanted to know, but that's where we're going. Thanks, Mark. That was fantastic information. I appreciate you sharing that perspective with the rest of the team. Any other questions or comments from the rest of the audience? We do used to have a couple minutes out there. Any other questions or commentary? While we're waiting for any additional questions, I did want to remind everyone that Mark, James, and I are going to be on a webinar hosted that is with the MBA, the Mortgage Bankers Association, on May 17th. I think we dropped a link to that in the chat. So that is a free webinar for MBA members and Zeventus is providing a code so that it can be complimentary to non-members. So by all means, join us if you can. And I believe that recording will be available as well. Let's see if there's another question. Okay, we just dropped it into the chat again. I do want to mention that in the next hyperledger meeting, I am going to try and get one of those technology companies. Yes. Is there a question? Yeah. Hi, this is Sanjay Nishankir. So I am the CEO of Intane. So I mean, one of the companies that you mentioned in your presentation. Oh, that's fantastic, Sanjay. Welcome. Welcome. Thank you. Would you be able to? I mean, I absolutely didn't know that you're going to mention it, but it's great to hear that you actually mentioned some of the work that we are doing. So that's great. Thank you very much for that. Okay, you're welcome. You guys have a very interesting solution. And if you're available later on this week, I'd like to reach out to you and speak with you more about it. Sure. So we have been doing this. I mean, obviously, we are talking about chat activity in the last half a year or so, but we actually worked on this product at least four years back. So we have been building on this, you know, primarily from a business problem perspective. But yeah, that's that's a topic for another day. I just had one question. Maybe, maybe someone can respond to that. So just based on this chat, CPT, is there any, any real life case study in which somebody is actually using some of these LLMs in terms of integrating with a blockchain backed application? I mean, obviously, you talked about four or five aspects in which this can be used. But is there any real life use use case, which has been done in the last six to nine months, where some of the LLMs can actually has actually been used to improve the smart contract efficiency or to train some of the other other applications. So I'm specifically for a blockchain backed application. I just wanted to know if someone actually has an idea on that. I have been researching that Sanjay, I've gotten whiffs of a couple companies that are bear that are keeping that information very, very close to their best. So the answer is I suspect yes, I don't think we will know definitively until they release their products so that they don't lose any market advantage. Okay, okay. Yeah, so I mean, the question comes to my mind because of the simple reason that we also, you know, within our team, we have been thinking about how best to use it. So we have been able to use chat GPT or co-pilot only in terms of our coding tasks where we use ReactJS or we use NodeJS in terms of writing applications. So we use co-pilot extensively for that. But then as far as LLMs are concerned, we have not been able to find out in what way we can actually integrate all of that. But that might be because the smart contracts that we use in our application, in our platforms, do not, I mean, these are all very large value transactions, but they do not have so much of data. I mean, like any typical structured finance transaction. So that's the context of this question. But yeah, I mean, we'll probably look for similar kind of applications and then see how best we can utilize it. Thank you very much. So Sanjay, this is Jeff on the call. I asked a question earlier, but since you're there, I got a question for you. So since you're out there already, if you're a company, company A, and I'm going to put in chat GPT, I'm the mortgage group, are you looking at ROI to be from a risk reduction standpoint, in other words, less the faults on mortgages, because I'm going to get my risk down on a loan, the optimized or cost reduction. In other words, I've got people here that process lots of paperwork, I can cut those steps out. But currently, you have any idea, anybody on the phone currently, what is the feeling of the market industry as far as the number one reason why they put AI in their business? I'm not sure if I understood your question properly. Can you just come up with the question again? So for business, what's your mortgage? So I'm a mortgage lender. What's my reason to bring in AI? Is it to reduce risk, in other words, giving out loans that people can pay? Or is it purely, I can reduce my costs? Oh, yeah, absolutely. Reduction of cost is the primary reason why there is a reason. Yeah, so why someone will use AI? I mean, I can just briefly in two sentences, I can tell you as to how we are using it as part of our application, which is one of the stronger selling points. In fact, when we take our platform to our clients, in fact, that's one of the lowest hanging fruits, the uses of AI in our application, which actually fascinates people and because of which people actually get into that. I mean, in just two sentences, what we do is we say that before we get any mortgage data into our blockchain platform, the data has to be digitized. The data has to be sanitized so that we have the authentic and correct data which we take into blockchain because once we take it into our ledgers, we cannot change it. We cannot delete it. So that's where we actually use AI to read all the contract documents, the loan contract documents and we extract specific data parameters which we pick up and save those or keep those in a blockchain ledger so that from that point onwards, that data is immutable and it's transparently available to the participants on our network. So that's a pretty useful scenario where we have been using AI and like I said, it has been one of the biggest selling points and which fascinates people to get into the platform. So and like I said, I mean, LLM is something that you're still looking at, but this has been pretty useful for us as far as adoption of our platform is concerned. Interesting. Thanks. One other comment I would give you is what some of the capital market people are looking at, especially in kind of the post GSE or the post banking regulation is how do I generate additional revenue and can AI help identify where those are? So if you look at the business models and how they're changing in some of the recent bank failures, AI has got another interesting role to play that is really just evolving because efficiencies, I think everybody recognizes the question is how do I actually adopt to change a consumer and generate new revenue if AI is the equalizer? Yeah, Jeff, have you heard about Redwood Trust White Paper? It's titled building a mortgage blockchain ecosystem? No, not that one. Not at all. Okay. If you pop your email in the chat, I'll send you a copy of it because this is something that they actually mentioned this in or talked about it in at the last MBA conference where they were able to identify 100 basis points and savings in the securitization of a portfolio of mortgage backed securities. It's really something that it brought them onto the map and it really outlined the benefits of blockchain. Not AI per se, but blockchain and how by putting that information onto the blockchain, you maintain transparency, immutability, all the benefits that a blockchain would provide and provide significant savings. I don't recall if it had any time savings, but in terms of we having to input the stare and compare that everyone talks about, it's a really good white paper. I'll email it to you after this call. Okay, that was great. Thank you. Okay. Okay. Everyone, we're two minutes past the hour. One last call for any questions or comments. Okay. I want to thank everyone for participating. This has been a great call. We love the interaction with people. Definitely invite you to attend the May 17th webinar and also our next hyperledger meeting where we will delve deeper into AI and blockchain. Thank you, everyone, and have a great day.