Today we have Lior Weinstein on the show! Are you tired of the same boring tasks taking up all your time? AI is transforming how we work and think in Real Estate and in every industry – with amazing new tools being developed every day that allow you to take advantage of increased productivity. With Lior Weinstein’s expertise, learn about how AI can end tedious manual work while boosting business performance! Hear first-hand advice about why this technology should be part of every investor's toolkit.
In this episode you will learn:
- How AI is a reasoning machine, not a fact machine
- How AI will impact all industries, including real estate
- The increased productivity that AI brings
- How AI will end boring work
Table of Contents:
- Where To Listen To The Podcast
- The Power of AI in Real Estate Investment and Operations
- Unleashing Language-Driven Value With AI in Real Estate
- AI in Real Estate Revolution
- Enhancing Communication With the Help of AI in Real Estate
- Why Having AI in Real Estate Make Things a Lot More Easier
- AI in Real Estate Takes Center Stage
- Exploring the Privacy Measures of AI in Real Estate
- Unlocking Creativity With AI in Real Estate
- Exploring Opportunities and Partnerships in AI for Real Estate
- How To Reach Lior Weinstein
The Power of AI in Real Estate Investment and Operations
Darin: Lior lives in Atlanta with his wife and three kids. Lior is a serial entrepreneur and an expert resource in AI and performance-based marketing. He's built four companies and exited two. Lior's specialty is in simplifying complex concepts into actionable steps that drive effective results in the form of digital products, traffic generation, conversion, and software engineering solutions.
So just a little bit on how we know each other and then we'll get into it. So Lior and I are both part of a mastermind group called the Genius Network, and we met there last month, and Lior just has a ton of knowledge with AI. AI is such a hot topic now. I asked him to come on the show to share what he's seeing in the space and how we can apply it in the multifamily space.
So with that, I wanted to ask Lior, he's also invested in real estate. I wanted to have him share where he's invested in real estate and what his plans are going forward and then we will spend more time talking about AI. So with that, Lior, can you share with the listeners where you're invested right now, what your plans are for the future as far as real estate, and then we'll get into the AI discussion.
Lior: Perfect. I'm a principal in a few deals in the short-term rental space. Specifically, we're actually using a lot of AI to do a data play on finding the types of deals. And we're looking for areas where we can have some inefficiency on the cost of operations of the property, where the revenue is pretty similar to other markets.
AI in Real Estate Investments and Margins Maximization
Lior: I'll give an example of that. Meaning, the insurance cost is a quarter of another similar market. Or the cleaning cost is half of another similar market, while the actual traffic and the potential equity value appreciation in the market is still relevant. So one of the areas we're looking at right now and we're very active in is Palm Coast in Florida. Just to frame it, if you're on the beach in Florida like in Destin or on the South Florida beaches, your insurance costs are going to be $8,000, $12,000. Especially with what happened over the past year. In Palm Coast, it's $2,200.
So you're 15 minutes away from the beach, but you're a quarter of the cost of insurance. Cleaning cost in a competitive market like West Palm or Destin would be 350, 300 for a house. Our cleaning cost is 175, which adds up. That's how you make the margins. So we have an opportunity still in those markets to get 20 plus cash on cash returns even with a property management cost included in that.
And because we do our own property management when I say we, I mean we establish our own centralized property management for all the properties that we own. Then we can make 25%, 28% cash on cash before any appreciation takes place, even with the 7% plus interest rates that we're paying for the properties.
Darin: That's crazy cash on cash. In our world, in the multifamily syndication world, we're typically talking 5% to 8% cash on cash and double your equity in three to five years, but a bulk of it comes from the exit, right?
Lior: That's right.
The Entrepreneur’s Creative Journey: Fusing Coding, Marketing, and AI in Real Estate
Darin: So those are huge cash-on-cash returns. So before we get into AI, what was your background prior?
Lior: I am originally from Israel. So I moved to the States 12 years ago. So born and raised in Israel. Technology guy, I've been interested in money and value creation since I was a very young kid. I started trading stocks when I was in about fourth grade. Started my first company, an actual, equivalent of an LLC when I was 14.
I've been going at it ever since. I had a big affiliate marketing business and ran a lot of traffic. In the early and mid-2000s, I went on to develop high-frequency trading algorithms for a large hedge fund for a couple of years, then built technology companies. Built four, and exited two. The focus is I've always lived between the CTO side of the universe because I started and was exposed to coding from a very early age, but marketing as well. So my world of marketing is more performance marketing, running traffic as opposed to brand marketing.
I lived between those two things. I think I'm like most entrepreneurs, I like takeoffs more than I like landings. Zero to one, let me create it. I'm a very creative guy and I love business and I'm very curious. So it doesn't matter to me if we're innovating in the trash hauling space or real estate or some neural network. So as long as I get the chance to build new stuff and have a vision and do some Cambrian explosion in a new market, then I'm happy.
How AI in Real Estate Is Revolutionizing the Industry
Lior: Now, I have three young kids, currently seven, four, and two. I live with my wife. We've been together for about 16 years in Atlanta, Georgia currently. A lot of my business life is oriented around how can I create cool things with great people while having a great family life. So balancing those things out as opposed to when I was a young gun, working 18 hours a day.
Darin: You're still pretty young. Come on now.
Lior: I mean younger gun.
Darin: So AI, you can't watch the news or read anything without hearing about AI. You shared at the mastermind with a bunch of different people here what you're seeing in AI. Can you share, just first start at the 30,000-foot level, how is AI impacting every industry today?
Lior: Yes. I'm very happy this is happening right now, and I'm generally in the category of tech adopters and passionate and optimistic about tech. I'm not in the technophobia category. So that's a big caveat to this conversation because depending on who you talk to. Most of the stuff you either listen about grimacing and how the world is going to end.
Darin: Some people are really scared of what's coming.
Lior: Some people are really scared, but I would say you have news clippings from the 1800s about the industrial revolution and how people were like, "Everything is going to replace humans." Historically, that never happened. That was always more beneficial than anything negative. AI is very different though, I would say, this particular revolution because not just the advancements in terms of the core technologies, but the products.
Leveraging AI in Real Estate for Better Neural Networks
Lior: Meaning, you have a lot of science over the years that gets discovered that's profound like Nobel award-winning type science. But the translation from that core innovation to a product people can use takes a long while. The first digital camera actually invented within Kodak was in, I think, late '70s, something like that. The first few pixels, but before that turned into a digital camera that completely destroyed the film industry in that perspective, it took decades. It took a long while. AI is not dissimilar. People have been working on AI challenges I want to say maybe since the '50s or '60s.
Forgot if it's really '50s. It's definitely the '50s or '60s, but maybe even as early as the '50s, and literally calling it artificial intelligence, meaning artificial intelligence research. Probably the same things you would hear people talk to now, that's how they were talking then. So it's one of these things that overnight success that took 70 years because now not only has the technology matured, but the product unlocks for the use case has been discovered.
So I've been using AI tools, and when we say AI in this particular context I mean the type of AI that leverages neural networks. The same basic concept that most AI systems have, and that's different than machine learning. We can talk about that as well if it's interesting, a little bit different than machine learning. Been around for many years. I've been using those, say, in the computer vision category since, again, the 2000s. In financials, has also been existent for a very long time.
Unleashing Language-Driven Value With AI in Real Estate
Lior: What happened with OpenAI and then successfully building up GPT and eventually launching ChatGPT. And with their GPT-3.5 model now GPT-4 is that it's relevant for anyone to use, and that's incredible. If you wanted to use artificial neural networks in whatever productized form a few years ago, you needed to be a developer. Not necessarily an expert developer. But you need to be able to run code and also have a use case where that type of impact and computing insight actually changes something, makes something better, faster, cheaper.
With a product like ChatGPT, because they use a large language model, which is it's one type of AI that's different than image generation, that's different than music, that's different than video. All these things are getting lumped together right now, but they are pretty distinct technologies. It's incredibly useful because it uses basic language that anybody can use at any different level.
First of all, let's maybe go back about what money is without going into a crypto conversation, which we can do as well. The only value of money for you, for Darin, the person is what it can do for you when you give it to other people. You're not going to pay yourself.
Darin: Exchange of value.
Lior: You're not going to pay yourself. If you clean your house, if you code your own thing, you're not paying yourself. There's no point in taking your own money and paying it to yourself. The only value you can derive from money is to give it to other people to do something for you, to build you an iPhone, to clean your house, whatever it is, give you a pita with falafel. That's the only value of money.
Language-Driven Value Exchange Empowered by AI in Real Estate
Lior: By far the only way we communicate with each other, the most common one to value exchange is through language. I call you and ask you for things, I text you, I write you an email, I post something. I use language to activate other humans. If it's by request, if it's that's how I give them value, if that's how I get value from them, I use language.
With a large language model technology like this, that's super accessible, you can already start to imagine. Well, by definition, if this makes any language better, any language communication better, and we can talk about all the ways it can. It means it literally touches any value exchange, nearly any value exchange that humans have, certainly the most common ones.
So if you think about your day in the life of your listener, wake up in the morning and all the things you do for yourself, you go make yourself breakfast, whatever. These are the things you do for yourself. But as soon as you talk to anyone, as soon as you engage with another human, language comes into play because you're articulating your thoughts in a linear way.
You're putting them in a sequence, in an order of words, in whatever language you're using right now to convey it to another person because you want something from them. You want them to know something or you want them to do something. That's what you're doing. As soon as that gets into play for your day, then a large language model can do it better.
Enhancing Communication and Efficiency With AI in Real Estate
Darin: I'm laughing because I'm thinking as you're saying that, so my wife and I, we've been married 20-some-odd years. But during different times, we've been to counselors and I remember thinking to myself, "Can I just have an app so that I can say what the counselor is telling me to say at the right time?" because I always goof it up. So that's just a thought, a personal relationship, but the communication happens with how you talk to customers, how you talk to new prospects and customer service. All of that is communication.
Lior: That's right. When I give talks about this, about ChatGPT and AI. I have this talk I give right now about AI models and methods, how to think and use AI so you can scale and do less busy work, scale without hiring and do less busy work. I basically say in that talk, I start to say, "Hey, this is not going to be a 20-minute flash bank of all the ways you can use ChatGPT. Because you can go to YouTube and TikTok, and that changes every day.
To be honest, that's more entertainment than education. Because your ability to ingest 50 examples and then retain it and put it into practice and retrieve it when you need to is generally limited to just not relevant." So most of my talk is focused on helping people think about this better because if they know how to think about this better, they can take advantage of it with whatever tools come into play in your industry. So you're in real estate, it's not very helpful for me to show you how to build a grant proposal for a government grant, most likely.
Unlocking New Possibilities With AI
Lior: Maybe marketing, maybe it is relevant, but it's not relevant for me to teach you the performance marketing side of how these tools can be used. Because you're not in the performance marketing space. But maybe some marketing, maybe you're doing a podcast intro or you're building the landing page about a deal. You're trying to explain to people, again, communicate with your investors. You're marketing a new syndication or maybe you're giving them investor updates and you want a certain tone.
You want it to be more formal because it's not so good news or you want it to be more informal because it is good news, and so on and so forth. So any single communication activity, these tools can accelerate and make it better because it's like having a 200-person IQ available to you in pretty much anything that requires language. The rate of performance change and how better these tools are on a weekly basis is second to none.
So I was saying the difference with these tools or these sets of technologies versus other explosions of technologies that we've seen in the past is the rate of adoption. That's generally when you have people that are more on the technophobia side and they're fearful about jobs getting displaced or replaced and so on.
The big context there is time. So when digital word processes go, "Oh, let's talk about computers," most people don't know before computers came, there were humans whose job title was computer. That was your job title. You computed. The whole staff in NASA. You sat with a pen and paper and computed. That's what you did. Then machines that did that came on and they were called computers because they did what the other computers did.
AI in Real Estate Revolution
Lior: So same thing, ATMs, automated tellers because you had tellers and now you have automated teller machines. So every time you have this new introduction of technology, it takes a while for that adoption to come in, and the transition from a pure economic perspective. When we talk about mass population and so on, has been very positive overall.
With this particular set of technologies, the adoption rate, because it's 2023, internet is available to nearly anyone in the world or cheap for generally little to no money. And news spreads fast. Bad news spread faster, unfortunately, but news spread fast, and because the product is very simple to use, it's a chat, most people know how to chat. Certainly, in the internet space, chats have been around since you had terminals and IRC, if people remember that, internet relay chat.
It's an interface that's very easy for anyone to use. Even my seven-year-old can use a chat interface and he does. So because the technology is, one, very widespread, it spread very quickly, fastest growing product in history, million users in five days, more than 100 million in 60 days.
Darin: That's crazy.
Lior: Nothing even came close. The most viral app you can think of didn't come close to this, certainly, not a viral company like Facebook and so on. So fastest growth in history and it's super cheap. So a lot of people, just to give you a sense, even with the expensive models right now, you can think about it like the core technology driver of the system that you're using.
AI in Real Estate Redefining Possibilities
Lior: The cost for me to use it, meaning as a developer, if I wanted to use the API, so they're giving me a basically retail cost, so not their cost, my cost to pay them to do the thing. It's about two cents for 7,000 words. That's like a book for a dime. You can write a whole book for a dime. That's, again, quote, unquote, "retail". Meaning, you can run your credit card right now and pay them and get access to that.
So the cost is so dramatically cheap, and we're talking in May 2023. I can't even imagine even in August if people hear this, months from now or a year from now the cost is probably going to be 10x down. So the economic opportunity of leveraging the technology is unprecedented.
Darin: Let me jump in because I think that there's certain people that are really getting into it and there's certain people that either haven't even tried it. Or I think a lot of people have gotten on and just fooled around with it and they're like, "Wow, that's really cool." So here's the deal, and I probably fall into that category you can go to, what is the website? Is it just ChatGPT?
Darin: Created by OpenAI, right?
Lior: OpenAI is the company. I think chat.openai.com.
Darin: If you Google that, you go to it, there's a free page that you can sign up for. It requires you to sign up, but then you can put in just simple queries and it is crazy what comes back.
The Impact of AI on Real Estate
Darin: So then what Lior is working on with people is really taking that to a completely different level. And how do you leverage that technology to maximize the profits and the efficiency for different business industries. Here's an example. I had a friend over who is an IT guy. He manages a bunch of coders he pulled up the computer in front of me and he said, I'm just going to ask ChatGPT to code this type of program.
He gave the instructions and it came back and spit out the coding for that. And he looked at it and he said, "Wow, this is pretty good." He said, "You know what? My guys would probably need to tweak this, but this would probably save them hours because they would all of a sudden have a starting point to tweak from." That's a real-life example in just one company.
Lior: That's actually a good segue. So one of the things I like to talk about, first of all, I explain to people. "You're not going to be replaced by AI, but you will be replaced by people using AI," and that's a very different mental model to how people are imagining these technologies.
There's a tendency to project into the future, do Star Trek-type futurism, and just imagine this all-in capable genie that can just do everything and replace entire systems. At some point, there's a different discussion about what's called artificial general intelligence, which may be a whole different type of discussion. A better model is to think on how AI technologies can accelerate humans.
Breaking Language Barriers With AI
Lior: There was a great article by the New York Times just a week or two ago and the title was brilliant. It said, "It's not the end of work, it's the end of boring work," because what this set of tools, especially that product, and there are more products in the market besides that. That product is so good. It's like having access to an expert on pretty much any topic you can imagine, but also that expert being able to talk to you at your level, and that's a big deal. The fact that you can literally tell, "Can you explain quantum physics to me as a five-year-old?"
Darin: That's crazy. You can. First, you could ask the question, "Can you explain quantum physics?" and then you could change it, "Hey, can you reword this based on a five-year-old?" and it'll change the wording. That's crazy.
Lior: That's right. It'll change the wording because it understands the semantic universe of a five-year-old and the grammar universe of a five-year-old, and also make it accessible in other ways. One, people don't know. People are not aware that right now ChatGPT has the single best machine translation product and it's basically free. Way in order of magnitude better than anything else, and it works.
If you go to ChatGPT right now, you can type in Spanish and you can just use it in Spanish or Portuguese or French or German. It's incredible because, for a computer, a human language is just a language. It's not different than computer code. Just a language, and it's incredible with that.
Enhancing Communication With the Help of AI in Real Estate
Lior: You can make things accessible in more than just a level of complexity but the format and style because that's really how we communicate. If we did this interview right now in Hebrew, it wouldn't be a very effective interview. It doesn't mean that I'm not giving the right answers to your questions, but you're not understanding them because you don't speak Hebrew.
Darin: You know what? It has a lot of applicability because it doesn't matter what industry you're in, right?
Lior: That's right.
Darin: If you've got five attorneys that you're going to talk to, well, one attorney, the way they communicate with you may feel safer and the right person to work with versus the other four. So they can use ChatGPT to help them to create wording that matches their style.
Lior: I'll give you another. I was just coaching an entrepreneur that I've been working with for a few years. She has a tendency to be a bit too direct, too straightforward, and some may be crass.
Listen, let's write what you want to say in your normal way of saying it. Don't edit yourself, but go to ChatGPT. Just write it. Just say, "I'm angry at so and so. You should have done this, you should have done that." Just write it in free form and then ask ChatGPT, "Can you rewrite this as an email, as a formal language, or an informal one?
Language Adaptation Using AI in Real Estate
Darin: Tone it down, make it a little softer.
Lior: Tone it down. By the way, you can tell it, "Tone it down, soften it." You don't need to think like a programmer about what are the right perfect words to say. Just use it as a natural language because that's what it's good for. Her emails and her employee responses to her in the past couple of months have been incredible. She gets to say what she wants to say, but she gets to say it in a way that's productive for her team.
So just having another layer of changing how you communicate, and I'll give you a non-business example. I was with my son. I give this example all the time. Four months ago, I'm with my son and sitting down on the couch playing Minecraft. He loves Minecraft. He asked me, "Dad, how do I get to the Woodland Mansion?"
I'm like, "I don't know," that's not it, but as soon as I said that, I'm like, "Oh, you know what? Hold on a minute," and I opened ChatGPT and I put it, "In Minecraft, how do I get to the Woodland Mansion?" and sure enough told me, "Yup, open the chat, forward slash locate space structure space mansion." I'm like, "Holy heck," and it worked. It was amazing, and same thing.
Darin: What did your son say to you?
Lior: Well, he was just happy that his father saved the day. He doesn't have insight of what just happened.
Darin: No, but you got a win with your son.
The Advantage That AI Provides
Lior: Well, I'll give you another one. So I have two boys and a girl, and my boys are Power Rangers, Minecraft, and classic interests. At night, I love to read them bedtime stories, but it's the same bedtime story, whatever books we have. Now, I go up, and I also don't have an hour every night to do, so I go up, open ChatGPT and I say, "Write me, whatever, children's story. I want a bedtime story. It's going to be 15 minutes. Make it exciting. Make it have at least two drama twists, and make it a theme of Power Rangers and Minecraft. My son's names are so and so. Make sure they're included as characters in the story."
Within five seconds flat, I get a 15-minute, which we speak about 150 words a minute. So it knows what 15 minutes is in speech. We read about 300 words a minute, we speak about 150.
So I get the format that I want, contextualized to the situation instantly, and that's a big deal, the fact that it's instant. For the same use case, if I had an employee that was an amazing, brilliant copywriter and I call him, "Hey, John, can you write me a quick bedtime story?"He'd be like, "Sure," but he'd get back to me after a few hours or after a few days, right? The fact that it's instant unlocks the value that it's right now. I can express myself and I get the turnaround right now. That's a very different value set.
Improve How You Present Yourself With AI in Real Estate
Lior: Now, can I write every day a children's story for 15 minutes with those? Yes. You can make it happen, but not in seconds. So like I said, if you are communicating this is a huge accelerant for you.
Darin: The other word I love that you said was productive. We've all heard the terminology, insane is doing the same thing and expecting different results. Well, there was an example that you shared that person was able to use the same language that they're used to using, but they got different results because it softened it up.
Lior: That's right.
Darin: The receiver viewed it differently.
Lior: That's right because, again, you're in your head. And as soon as you're talking to somebody else, you need to figure out how to talk to them starting with actual language, literally. Then what style because you talk differently to your friends and family and old friends from school and different to a cashier and different to a business partner, different to a lawyer. The context is also if you're talking in a keynote at a big conference, very different than talking to a bunch of friends at home. So you change your style and tone all the time based on the context.
This being an assistive technology in that language is incredible. Then the only missing pieces are more context to make it available to you. A lot of demos right now have these wearables in your glasses or on your chest, whatever it is that has audio. They ingest the situation in real time. They give you recommendations in real time like having an assist in your ear and it tells you how to say what to say at that point in time.
Why Having AI in Real Estate Make Things a Lot More Easier
Lior: So these are all engineering problems. These are all just how do you take the core technology that already works and make it more accessible at the right cost. With the right, battery life and all these problems to make it relevant, you change it all the time. And because it's a master at any language and any language style and format and tone and content, it will truly affect any single thing you do. You want it to affect every single thing you do because it's going to make it better. It's not going to make it worse.
Darin: So a couple of other use cases. So I've heard people say with education, college kids, they have to write a paper. Well, they've had the ability to write a paper. Go to people that are a year or two above them, and go to a workshop and have them edit it and they go back and massage it and make it better. Now, "Hey, I want to write a paper on quantum physics," and it pops out right away. So some people think that "Well, that's bad because the kid could write a paper right away."
I'm of the mindset of, well, look, it used to be where, to your point of computing, there were people that computed, there were people that didn't want kids to have calculators. Now, there are these calculators that are really intensive, but you still have to know which buttons to press to get the answer. So how do you view that type of scenario?
AI as the Next Gen of Tutors
Lior: So education is going to change profoundly. Khan Academy just had a major announcement last week showcasing their AI. I forgot what they're called, not Khan AI, Magi AI, or something. Incredible. So they've basically had early access to GPT-4 since last summer. And they've honed in an AI tutor that cannot only understand what your gaps are based on your answers but explain the concepts based on your level, which you see the demo, it's mind-blowing.
Darin: Oh, man, that's huge. Think somebody's not good at math.
Lior: They're not good at anything.
Darin: Yes, but whatever their subject is, maybe they're really strong in four subjects and one subject, I'm just going to pick math. Maybe they're not their hot topic, but rather than have to wait or go ask a friend or somebody that's more senior or a TA or whatever.
Lior: A tutor.
Darin: Tutor or whatever, they could actually say, "I don't understand this question."
Lior: Well, it's more than that. This is a working product that works. You can sign up right now and use it. It asks you a question to do something. You can certainly be proactive and ready. If you know how to articulate your problem, that's a big difference. So if you know, "Here's what I don't understand and what I know." It's going to do a great job, but it can do something even better, which is it can ask you a question and can see your answer. And based on your answer, it can deduce what concept are you missing, and then teach you that concept or a sequence of concepts. Incredible.
From Memorization to Critical Thinking With AI in Real Estate
Lior: Education is a good example because most of modern education is based on this linear instruction of saying, "Here's a curriculum. I have kids in front of me without knowledge. I have teacher on the other side with knowledge, and I'm going to translate the thing that I know to their brains." Generally, it's linear in the sense that you have everything in sequence. These are the types of lessons we're going to have every single week. This is how I'm going to teach it.
I'm going to teach it by articles or by quizzes or by whatever method you chose, and however, the teacher connects to his particular class, with their particular class. It's incredibly antiquated. It's been antiquated for many years. Primarily, that's more of a personal opinion. The biggest challenge with this education system, as it is, is that it's not focused on helping you think. It's focused on helping you memorize, so growth memorization, which goes to the issue of, generally, well, if somebody's doing the work for you, do you know how to think?
Now, a think calculator is a great example. I remember when I was taking math in high school, I had this book of all the tests in Israel that people had done for graduation tests over the decades. I saw the tests in the '60s and '70s that my parents took, and I'm like, "Whoa, this is way harder than what I have," and they had to do it without a calculator at the time. So I'm like, "Oh, that's super impressive," but you can't argue that using calculators took math backward. It certainly took math forwards and made it more accessible to more people.
The Future of Information With AI in Real Estate
Lior: Now, for the people that are really good at it, they can do it without a calculator. It's not a big deal because they're really good at it, they're passionate about it. I think assistive tools just like Google, meaning if you had to find something. When I was in school, I had to go to a library, find stuff to find information. And use a copy scanner, whatever, to get the copies of this page back home and highlight it and so on because even Yahoo and AltaVista at the time weren't great search engines.
Most of their results were junk. Then Google came and suddenly it was, oh, a relevant experience. Also took the world by storm, and I'm saying this in quotes for people that listen because it took a few years to get adoption. It was a rapid adoption, but not a million in five days. Now, everybody has the Google muscle. Everybody knows that if they try to find a reference for something, some piece of knowledge, they go to Google or their favorite search engine.
And they put in a keyword and these search engines use their massive indexes to match this word. These words that you put in the search bar to something they have or something close to that. With these sets of technologies, because they can generate, can create on the fly, they have a massive advantage over search engines. The search engines are limited by their index.
Darin: So will ChatGPT or something like it replace search engines? So instead of saying, "Google it," you'll say, "ChatGPT it," or whatever?
How AI in Real Estate Will Improve the Availability of Information
Lior: So Google just yesterday for the timing of this interview, Google yesterday in their IO conference announced their entire AI integration suite and they've revamped it. They had a product they've launched many years ago called Conversational Google, which was never good. So people are not aware. I think it launched in maybe 2013, something like that, so about a decade ago. Did you hear about it about Conversational?
Lior: It's been around though, but it was junk. This is a trend that's been happening for a few years. So by far, you have this trend of search engines becoming answer engines. So instead of you putting a query and then you go into clicking through a link and try to basically get a resolution for your question. Google already years ago, if you search for a local business, you just get the local business listing. If you search for a product to buy, if it's in the index, you get the product and the price and the place and all the things.
Same for flights, you search for a flight, you just get the flight information. So you get the answer instead of clicking through them. What is enabled right now, and that's what they demoed, is just getting the digested answer within the search result page without you needing to go in. Bing had an incredible bet a few years ago being the largest corporate investor in OpenAI. Bing Chat is the first one that launched that experience and it's an incredible experience. It's way better than going into specific links.
AI in Real Estate Takes Center Stage
Lior: Now, some parts of search are still very relevant, and I don't think they're going away anytime soon. But I think because the experience is so better in giving you the actual final detailed answer, I don't think the winner is ChatGPT. Meaning to your question, the product winner I don't think it's going to be ChatGPT. I think there are going to be other products that are going to win, but the experience that now everybody's expecting is going to win.
That's why you see it in Bing, you see it in Google, which is, "Just give me the answer. Don't make me go through these click hoops and try to figure it out myself." So I think for education, the profound change is going to happen and it's going to happen very quickly, very rapidly. If the system right now is based on we're just going to give you the information in a linear way in sequence. Then the way we're going to make sure you know it, we're going to make you produce a knowledge product, an article, homework, whatever it is.
Answer a test, questions, and answers, that is going to have to get replaced very quickly, and I don't mean very quickly in a decade. I mean very quickly within a year or two because it's simply not relevant. When you have access to these tools and you can produce a work product in an asynchronous way. Meaning you're in the classroom, and somebody's giving you homework. And then you have a few days to come back with the result or certainly, when you can do it in a few seconds, it's completely irrelevant.
AI in Real Estate: Leveraging AI for Marketing
Lior: So there's going to be a new philosophy or a new and old. Everything's probably going to go back to Socratic times. The Aristotelian methods of talking and actually making sure you know how to think and how to create as opposed to us using your ability to write an article as the yardstick.
Darin: No, that's huge. So let's take what we've been talking about with this AI. If I'm a multifamily investor or I'm a syndicator, how do I leverage this technology?
Lior: I think endless, and certainly for some activities, so most syndicators have some form of marketing. So the marketing might be in the form of creating a landing page location with a deal, ads, or trying to get investors if you're doing cold marketing. A lot of them do webinars. You can go to ChatGPT right now and say, "Hey, I'm a syndicator. I'm promoting this multifamily deal. It's three complexes in Texas and so-and-so units, and here's the cap rate, and here's the maintenance cost, and here's how we found about the deal."
Just do it in natural language. Don't be sophisticated about it, just use your raw language, however poor or advance it is. Doesn't matter. Just use your own language. You can even just go on the website, on your phone, just use the voice-to-text, just talk. Don't worry about being long. Don't worry about being not nuanced. Just vomit everything from your head into the system and say, "Can you create a 60-minute webinar for me containing all of this information?" and it'll do it in four seconds.
Confidentiality Considerations in AI for Real Estate
Lior: Same for anything else you have, "I need to write an investor update. Here are the top line numbers for this update or here's the under contract. We're negotiating with banks," again, raw information. Don't overthink it and say, "I need to write an investor letter. Please make sure it's not too long. Make it a little bit informal. You know what? Can you add a couple of tweets with it so I can use and post it on my Twitter or our Facebook group?" or whatever it is, and then it's going to add shorter summaries of this and add icons.
Darin: That's huge. Now, what about the confidentiality? So if you've got a business and you're actually like you mentioned earlier, you can talk to like it's your chief marketing officer or chief technology officer, whatever. But if you're giving real financial information into it, what are the confidentiality concerns that that information can get out someplace, your private information?
Exploring the Privacy Measures of AI in Real Estate
Lior: Just last week or maybe two weeks ago, in the user interface, they actually added something in the setting that you can just opt out of that. So you still own the content and nothing is used by them for training. On the API side, it's the reverse by default. So by default, if you use the API, if you have a more advanced use case, if you have some developers helping you build. It's default opt-out from training and you have to opt-in if you want to help, basically, improve the ecosystem.
Without getting too technical, that's the thing that turns something into something else, the transformers. It's not like Google where it scrapes the information and then piecemeals back a response. That's not how the technology works. So when you get a response from ChatGPT, it's not copy/pasting from somewhere. It's important to understand that. If you and I right now go to ChatGPT and put in the exact same prompt, we are not going to get the same answer and we can run that multiple times.
Darin: I'm really curious as to what the answer how it works because I'd still think that that's what it's got to do, right?
Understanding the Uniqueness of AI Responses
Darin: I've heard that two people put in the same exact thing, they're going to get back a different response. Why?
Lior: We can go technical if you want to.
Darin: I don't want to go too technical, just if you can put it into everyday language.
Lior: What's interesting is that artificial neural networks work. It's quite a surprise to a lot of people that the original hypothesis of saying, "Here's a neuron, here's a cell in our brain." A neuron basically fires or not. It's one or zero just like a computer bit, it's like a computer works. Then you have synapses and dendrites, things that connect neurons and make them fire together or fire in a certain sequence. When artificial neural networks started working like we saw that we can put something in a network as an input and get something as an output, it was like, "Oh, shit. This actually works like a brain."
Obviously, our brain is way better in so many different ways, but the effect, the thing that we think is thinking is like thinking actually exhibits something we like. So when you put in a prompt, and a prompt, I don't necessarily love that frame because a lot of people conflate prompt and question, they think about Google, they think about a query like a question, but it's not. It's really a before and after. You put in before, some an address, and it puts in an after.
Navigating Fact and Creativity Using AI in Real Estate
Lior: It's like, "Well, given this before, what is the likely after?" That's really what it does. The basic technology is to guess the next best word, statistically speaking. The next best word becomes the next best phrase, the next best paragraph, and so on and so forth. If I ask you right now a question that's a creative question, you'll also, probably every time we talk, give me a different response because it's not about fact retrieval. This is important to understand. These are reasoning machines. They're not fact machines.
Actually, it's a big problem right now that a lot of companies including OpenAI try to solve when it gives you a response that is fact-based and not creative-based. Creating meaning there's no standard of truth. It's just a creative response. It could be any response and it's going to be okay. Chocolate or vanilla, that doesn't matter. Choose one of them. There's no right or wrong. There's only relevant or not for you. So it's a relevancy question, not a fact question.
They're trying to solve it. They're trying to figure out when it gives you a response, how can they give you a reference as to how I came up with this response. Because if I told you right now, if you asked me, "Hey, give me the top 10 highest rental markets right now that are 30 miles away from big cities, at least a population of 300,000." You're asking for a fact. One, if I ask you this question right now, it'll give you an answer, a very confident one, and it'll order it in one through 10. But you actually have no idea if that answer is true or not.
Risk Assessment and Customized Analysis of AI in Real Estate
Lior: So the fact that it gives you an answer doesn't mean it's true. That's a limitation of the system right now, which is getting solved really every week with plugins and extensions and making the models better and so on. Now, OpenAI, if you go to it, if you had plugins already enabled and it works, you have Zillow. So now when you put in a query and it understands this answer might be in Zillow, it actually goes to Zillow and gives you facts, and it tells you, "This came from Zillow,".
Darin: Oh, man. I love that example that you brought up because you could just go into AI and say, "What are the riskiest real estate markets valuations right now?" You don't know whether it's right or wrong, but it could provide you a starting point to say, "Hey, I need to look at these. Do I have investments in my portfolio?"
Lior: Here's what you can do. You can ask it that, which is an abstract. You use the word risk, but the risk for you is not a risk for me, not a risk for BlackRock, not a risk for a first-time investor. But it's going to give you an answer, and then you can literally ask it, "Can you explain how you came up with this answer? Can you explain how you decided on risk for this answer?" Then you can redefine risk and you can tell them, "Can you run it again and make the properties under this and under this value?
Unlocking Creativity With AI in Real Estate
Lior: "I want to see properties that require less value add or more value add in the properties," and so on because there are so many forms of risk. "I don't want to see section eight," whatever. "I don't want to see coastal options in Florida," whatever it is. There's so many forms of risks.
Darin: That's huge.
Lior: That's the thing, and you clued in right. It's a great writer's block. It's a great kindling for the thought process. And this is a good example of where the system is so valuable in a process that a human needs to synthesize the responses to the next step of action, whatever it is. As opposed to replacing the human for all the actions involved. Normally, if you wanted to do that research. One, I've been writing copy for decades, but even now when my team asks me to write something, a blog post or an email, I just immediately get stuck, deer in headlights.
Now, I'm trained and I can get myself over it, and what most writers do, is just force themselves to write, but the fact that I can go in and do exactly that, I just put it in four words. I'm like, "I need to write an email," and just that, and it'll write an email, "Dear subject," "Dear person," subject name, place. So it'll just do something. As soon as it does something, for me to edit, I'm way faster at editing than I am at creating. So having this quick iterative process on research, on creative work, on whatever, writing anything, again, communicating anything. Writing anything just unlocks the writer's block. It eliminates it.
Key Performance Indicators in Real Estate With AI
Lior: If it's writer's block for creative purposes like marketing, whatever, or just research purposes because otherwise. The amount of intellectual work you needed to articulate the best result and the frame of the project is just so much you're going to abandon it because you'll become hungry or whatever.
Darin: That's massive. You talked to me about a product because I was asking you like, "Well, how are you using this technology, and have you created any companies around it and any products around it?" You had mentioned this joinpulse.ai and that you were developing KPIs for different industries. I just started thinking, "Wow." In our industry, there's syndicators that might have 10 different properties, and what are the key performance indicators that they have and how could they get access to that quickly? So can you talk about that?
Lior: Yes, and thank you for mentioning that. So joinpulse.ai, started with a very simple product that we started building a few weeks after ChatGPT launched. The original premise was like, "Wow, so many business owners have this, what we call, implicit anxiety about how their business is doing." It started with just, "How much cash they have?" Then how are the sales going or is anything going wrong?"
I think so many entrepreneurs share that feeling if they're running 100 million dollar revenue businesses or 10,000 revenue businesses. You're just constantly thinking about it. So we started off creating a simple product where you can connect all your business systems and just get it as a text message everyday and on your calendar everyday. So you don't have to log into anything. Then we said, "Well if we can get the data, let's give you insights on top of the data."
Enhancing Business Insights With AI-Powered KPIs in Real Estate
Lior: So we be that watchdog. We can say, "Hey, we noticed this recurring subscription," or, "We noticed that rent is late," something that should probably we put a spotlight on, and then it's up to you to figure it out. Once we created that, people asked, "Hey, can you add Shopify for e-commerce people or Salesforce or my Facebook marketing spend?" So we ended up calling it Pulse, like keep a finger on the pulse of your business. The core feature is we say, "Now, your KPIs have AI."
So not only you can get those KPIs everyday or every week or every month, whatever cadence you want, and if you're the kind of person that wants to see one number every week, you can. If you're the kind of person like me that wants to see 50 numbers everyday, you can. So you decide on the level. My investors, for example, they get a monthly calendar feed with how much cash we have and what runway we have. Just two numbers. They don't need to see the payroll transaction. What we enabled actually just last week is now you can have a chat with those KPIs.
So in the product, if you connect your QuickBooks, you can have a chat with your financials as if it's a CFO. Same questions you would ask your CFO, "Most profitable month? How's the balance sheet looking? Any big transaction coming up? Any age receivable that I should look at?" Just same questions, natural language, and you could answer. If you connect your marketing systems, you can ask questions like a CMO. If you connect HR systems, you can ask marketing like a CHRO and so on and so forth.
Contextual Understanding Of AI in Real Estate
Darin: So this is just a thought that just came from what you were saying. Look, in the last number of years, I've heard many and many business people talk about leveraging, say, virtual assistants, and how can you do that. There's people who just, swear by it, it's just improved their business dramatically. But then what you're saying right here is you could actually have highly intelligent conversations in different aspects of the business with a reasoning machine that comes back with instant feedback.
Lior: This would be like an unlimited context because, unlike a human, humans are amazing at reasoning. We're incredible, but we can't keep context like computers can because computers can always remember everything and keep it in the frame. So the fact that you can have access to this incredible reasoning machine that has endless context and instant is incredible. Like anything in life, any communication, the quality of the communication is to the extent that you know how to articulate what you're looking for.
You would never go and talk to a copywriter or a CFO and say, "How's it going? Give me my books. Tell me about my financial." It's too abstract. You are going to immediately get some clarifying question, "Well, what do you mean? What date range? What do you want to do? When do you want to do it? Do you want a one-liner answer? Do you want a page?" You need to clarify. So if you get good at articulating your thoughts and your intent and understanding what is the thing that you're looking for, be very good at asking. Again, not just questions.
Exploring Opportunities and Partnerships in AI for Real Estate
Lior: Maybe that's one last topic we can talk about on how to think about AI in those forms. Then you're going to get really, really, really good responses. I'm working with a pilot right now with a manufacturing company on helping them figure out how to better have access to their data in real-time with a couple of medical practices. Enabling the doctors and the nurses to do their job better, again, because they're very good if they get the right information at the right time. But if they're not, they're not very good, they're not very productive, and we can talk about there's a concept I'd love to share maybe if we have a couple more minutes.
Darin: I know you have a timeframe, so I would say, one, if you can share, what do you look for? I know that you're very curious and you like to partner on new different opportunities. So what do you look for in that type of opportunity? What are the types of people and opportunities you're looking for?
Lior: So my best businesses have been joint ventures with other entrepreneurs. My special skillset is integrating this world of CTO, CMO, and being at the forefront of whatever technology it is. Knowing what to take advantage of and what to ignore because a lot of things people have a tendency to accelerate in their minds because they have very good imaginations. But they think the future is closer than what it is. In some cases, they don't appreciate how fast things are moving in other cases. I have developed a pretty good objective way of taking action on these types of bleeding-edge technologies.
Key Qualities to Look for in Collaborative Ventures
Lior: What I'm generally looking for is people that are really good at whatever they do. They have a unique insight into pain in their industry, and the pain doesn't have to be super sophisticated, but it has to be unmet, and unanswered. The pain could be novel, meaning nobody had answered it or the pain could be because it was answered, but it's extremely expensive and it's cost-prohibitive. And because it's cost-prohibitive, it's not as accessible to as many clients.
So then they share the pain. Maybe they have an idea for a product, but I love to think about those solutions as well. So as long as they have an insight into the pain and they can be a good first client because something painful that they actually spend money on, ideally, then we partner up. We build a company together and we grow it together.
My frame is I'm not looking for another job. I'm an entrepreneur. I love building companies. So we build a company together and then like I said, we take it off the ground, but then we staff it. Somebody's going to CEO it and maybe the partner will CEO it because that's what their next thing is going to be or maybe we just find somebody to grow the company. So I love that. I'm very industry agnostic. I have projects in pretty much any industry you can think of, medical, legal, FinTech, and real estate.
From Business to Adventure
Lior: We can talk for two hours about small nuclear manufacturing if you want, nuclear reactive manufacturing if you want. So I'm very indifferent to what it is. I'm more clued in on, who's the partner? Are they a great person to work with? Are they kind? I have a big no asshole rule in how I do business. Is the opportunity interesting?
It's not very interesting if it makes another $500,000 a year. I'm looking for the big one, but I don't need a 500 million dollars a year either. Is it impactful? Is it truly when people in the market hear about it they're like, "I want it. I need it," and if you asked, then I'd love to hear from you. Find me on LinkedIn. Very easy to find me on LinkedIn.
Darin: So do they just look up your name on LinkedIn? Is that the best way for people to reach out to you?
Lior: That's it, linkedin.com/in/liorweinstein, and you'll find me.
Darin: There you go. So what do you like to do outside of work for fun? Now, you've got a family and you've got three little ones.
Lior: I've got three kids that I love. We travel a ton, a ton. And we have a big RV, We love RV travel. We fly together all the time.
Darin: That was one of the common denominators we both love the RV lifestyle.
Lior: We have a similar size RV as well, a class A pusher. So we do RVing. We fly to usual places like Disney and we fly to cool places like Hawaii or Africa, whatever.
The Value of Conversation
Lior: So we travel a lot and love traveling with them and sharing that experience with them. Three kids with the car seats and all the luggage, so travel days are not great, but then you get somewhere cool, so it's worth it. Then different activities, I love spending time with my wife. We love trying new food places, new movies. We're big movie fans. When I'm in Hawaii, I love snorkeling and scuba diving, and surfing, those types of water activities. Not so much those in Atlanta, Georgia.
Darin: Hawaii, Costa Rica, whatever. Well, I appreciate you coming on and sharing this. This AI world is fast-growing and it's going to impact all of us one way or another. I think if I was to wrap up what you said is that look, you're going to be looking for people that know how to leverage this technology to do their jobs better. It's all facets of the business, it's all levels of the business that can improve their performance and be more productive because they're able to use this technology.
Lior: The last tip I would give people, use it like a conversation. Use it like you would talk to another person. Go at it. Don't edit. Don't try to be too short. It'll respond to whatever you give it, but the easiest way is to actually just have a conversation with it and see what it does back, and personal or professional doesn't matter.
Darin: That's huge. Listeners, I hope that you enjoyed that one. Until next week, signing off.