The Investor's Edge Lives in Their Head — Anekha Sokhal, Moshi

Hello everyone. Hello everyone. Thank you very much for coming to Thank you very much for coming to Thank you very much for coming to my speech. I am Annika. my speech. I am Annika. my speech. I am Annika. I'm the founder and I'm the founder and I'm the founder and CEO CEO CEO of Moshi, and I'm of Moshi, and I'm of Moshi, and I'm going to talk about the going to talk about the going to talk about the context financial context financial context financial investors investors investors hold in their hold in their hold in their heads and how we can heads and how we can heads and how we can use that use that use that to to to build machines today. I build machines today. I build machines today. I want to do a quick want to do a quick want to do a quick survey. How many survey. How many survey. How many people in this room people in this room people in this room feel feel feel overwhelmed by overwhelmed by overwhelmed by the amount of data they use to the amount of data they use to the amount of data they use to make decisions? make decisions? OK. And who OK. And who feels that now, feels that now, feels that now, due to the huge due to the huge due to the huge amount of data, your amount of data, your amount of data, your attention has become less attention has become less attention has become less focused? This is precisely the focused? This is precisely the paradox that financial investors face every day. financial investors face every day. So, in 2025, the world So, in 2025, the world So, in 2025, the world generated about 175 generated about 175 generated about 175 zettabytes of data. This is zettabytes of data. This is zettabytes of data. This is exponentially exponentially exponentially more than what we more than what we more than what we had two decades had two decades had two decades ago. We used to be ago. We used to be ago. We used to be lucky if we lucky if we lucky if we had an internet had an internet connection, newspapers, and connection, newspapers, and connection, newspapers, and books, but today books, but today books, but today we have all of that, plus we have all of that, plus we have all of that, plus context from PDFs, context from PDFs, context from PDFs, subscriptions, tons of subscriptions, tons of subscriptions, tons of podcasts, and an podcasts, and an podcasts, and an endless stream of endless stream of endless stream of information from X and LinkedIn information from X and LinkedIn . What does this mean for . What does this mean for . What does this mean for financial financial financial investors? Because investors? Because investors? Because more data does not more data does not more data does not equal more equal more equal more information information . In reality, this just . In reality, this just . In reality, this just creates the creates the creates the same flow of same flow of same flow of information for everyone, but your information for everyone, but your information for everyone, but your attention remains attention remains attention remains the same. So for the same. So for the same. So for financial professionals financial professionals financial professionals who are paid to who are paid to who are paid to synthesize data, the gap synthesize data, the gap synthesize data, the gap between attention and data is between attention and data is between attention and data is getting wider and getting wider and getting wider and wider. Who am I?

wider. Who am I? wider. Who am I? I've been on both I've been on both I've been on both sides of the barricades. I've sides of the barricades. I've sides of the barricades. I've spent my career spent my career spent my career fascinated by fascinated by fascinated by data, and I've data, and I've data, and I've worked with it my entire life worked with it my entire life . I was a derivatives trader for almost ten years . I was a derivatives trader for almost ten years , and then , and then moved into the field of moved into the field of moved into the field of data science and data science and data science and machine learning. I machine learning. I machine learning. I received a full received a full received a full scholarship to scholarship to scholarship to study machine study machine study machine learning and learning and learning and published a paper published a paper published a paper at ICLR while working on at ICLR while working on at ICLR while working on machine machine machine learning systems with NASA. learning systems with NASA. learning systems with NASA. And now I have combined these And now I have combined these And now I have combined these two worlds. These two two worlds. These two two worlds. These two worlds became the basis worlds became the basis worlds became the basis for Moshi. Moshi is your for Moshi. Moshi is your for Moshi. Moshi is your 24/7 24/7 24/7 quantum analyst, quantum analyst, quantum analyst, monitoring the monitoring the monitoring the markets 24/7, understanding markets 24/7, understanding what is happening and what is what is happening and what is what is happening and what is important to the important to the important to the financial financial financial investor. Having been investor. Having been investor. Having been on both sides, I can on both sides, I can on both sides, I can understand the strengths of understand the strengths of understand the strengths of each each . And this is exactly what I want to . And this is exactly what I want to talk about in detail today. I want talk about in detail today. I want to talk about that to talk about that to talk about that car that everyone is car that everyone is car that everyone is trying to trying to trying to outdo. outdo. outdo. The human brain, for The human brain, for The human brain, for context, context, context, operates on a power of operates on a power of operates on a power of about 20 watts, which is about 20 watts, which is about 20 watts, which is equal to the power of a equal to the power of a equal to the power of a dim light bulb. That's dim light bulb. That's dim light bulb. That's 86 billion neurons 86 billion neurons 86 billion neurons that create connections that create connections that we still can't that we still can't that we still can't fathom. And fathom. And fathom. And all this works on the all this works on the all this works on the power of a regular power of a regular power of a regular light bulb. And the systems light bulb. And the systems light bulb. And the systems we're trying we're trying we're trying to emulate, even for a to emulate, even for a to emulate, even for a fraction of their fraction of their fraction of their power, power, power, require an entire require an entire require an entire power plant to power plant to power plant to operate. So, in terms operate. So, in terms operate. So, in terms of of of efficiency, the efficiency, the efficiency, the human brain is human brain is human brain is millions of times millions of times millions of times more powerful than more powerful than any systems we any systems we any systems we try to create try to create . And what this really

. And what this really . And what this really means is that it's important means is that it's important means is that it's important to understand where exactly to understand where exactly to understand where exactly you're taking these you're taking these you're taking these technologies. So, at technologies. So, at technologies. So, at what stage of the what stage of the what stage of the decision-making cycle decision-making cycle decision-making cycle should should should the technology, i.e. the AI the technology, i.e. the AI the technology, i.e. the AI we are currently building, be applied we are currently building, be applied ? Because if ? Because if ? Because if the most efficient the most efficient the most efficient thinking mechanism thinking mechanism thinking mechanism is between is between is between your ears and your ears and your ears and runs on the power of a runs on the power of a runs on the power of a 20-watt light bulb, then 20-watt light bulb, then 20-watt light bulb, then where is the place for where is the place for where is the place for machine learning? machine learning? Are you trying Are you trying to replace it—which is what I think to replace it—which is what I think to replace it—which is what I think AI is AI is AI is trying to do trying to do trying to do every day—or are you every day—or are you every day—or are you actually actually actually helping it? So let helping it? So let helping it? So let 's take a 's take a 's take a few steps back few steps back few steps back to understand how the to understand how the to understand how the brain of a trader brain of a trader brain of a trader or financial or financial or financial investor thinks. So, what do investor thinks. So, what do people really want? Any people really want? Any investor you investor you investor you talk to is looking for talk to is looking for talk to is looking for one thing—advantages. What one thing—advantages. What one thing—advantages. What is an advantage? In is an advantage? In is an advantage? In practice, the advantage practice, the advantage is only three things. I is only three things. I is only three things. I want to know what's want to know what's want to know what's happening before anyone happening before anyone happening before anyone else. I want else. I want else. I want to know my "blind to know my "blind to know my "blind spots." So, spots." So, spots." So, tell me about the tell me about the tell me about the second and second and second and third order consequences of what's third order consequences of what's third order consequences of what's happening happening . And I also want to have . And I also want to have . And I also want to have confidence in my confidence in my confidence in my actions so that I can actions so that I can actions so that I can increase the volume of the increase the volume of the increase the volume of the deal I believe in. All deal I believe in. All deal I believe in. All three of these things together three of these things together three of these things together give you an advantage.

give you an advantage. give you an advantage. Now here's the Now here's the Now here's the problem. If you're problem. If you're problem. If you're building something in this building something in this building something in this area and everyone has area and everyone has area and everyone has access to the access to the access to the same models and same models and same models and systems, then what's the systems, then what's the systems, then what's the advantage? Is this a advantage? Is this a advantage? Is this a skills file? Is that a skills file? Is that a skills file? Is that a bunch of MCP connectors? Is bunch of MCP connectors? Is bunch of MCP connectors? Is it just a better it just a better it just a better prompt? Is this a better prompt? Is this a better prompt? Is this a better model? Everyone has model? Everyone has model? Everyone has access to all of this access to all of this access to all of this . So, if everyone . So, if everyone . So, if everyone has access to these has access to these has access to these things, what is the things, what is the things, what is the advantage then? The advantage advantage then? The advantage is actually something that is actually something that is actually something that cannot be cannot be cannot be downloaded. This is what downloaded. This is what downloaded. This is what is in the is in the is in the investor's mind. And how do they investor's mind. And how do they investor's mind. And how do they get it? That's get it? That's get it? That's a decade of working a decade of working a decade of working and learning from the and learning from the and learning from the best traders in the best traders in the best traders in the industry, sitting next to industry, sitting next to industry, sitting next to them. And here's what's them. And here's what's worrying. What worrying. What happens when happens when happens when you have an investor you have an investor you have an investor who has everything in their head who has everything in their head who has everything in their head and they and they and they move on to another move on to another move on to another firm or firm or firm or retire? This retire? This retire? This intellectual intellectual intellectual property goes with property goes with property goes with it. So, I think it. So, I think it. So, I think before we start before we start before we start thinking about how AI thinking about how AI thinking about how AI fits into all of this, fits into all of this, fits into all of this, we have to understand we have to understand we have to understand what a what a what a financial investor does.

financial investor does. They monitor the They monitor the markets, which are now markets, which are now markets, which are now operating 24/7. operating 24/7. operating 24/7. They are trying They are trying They are trying to understand what is to understand what is to understand what is happening and how happening and how happening and how it all connects. it all connects. it all connects. So, those billions of So, those billions of So, those billions of different different different data connections. They data connections. They data connections. They are trying are trying are trying to compare and to compare and to compare and understand the historical understand the historical understand the historical significance of an event that is significance of an event that is significance of an event that is happening right happening right happening right now. What happened in now. What happened in now. What happened in the past, why it the past, why it the past, why it happened, and how happened, and how happened, and how the market reacted the market reacted the market reacted when it happened. when it happened. when it happened. And they're trying to And they're trying to And they're trying to put it all together put it all together and then run it and then run it and then run it through the through the through the models they have. And the last models they have. And the last models they have. And the last part is actually part is actually part is actually making a decision. making a decision. making a decision. So a lot happens before a So a lot happens before a So a lot happens before a financial decision is made financial decision is made . I want . I want to stop with you at the to stop with you at the to stop with you at the last stage, last stage, last stage, namely namely namely decision-making. So, let's decision-making. So, let's decision-making. So, let's consider what consider what consider what AI has actually changed. AI has actually changed. In the past, if you In the past, if you worked in finance, worked in finance, worked in finance, you had a team of you had a team of you had a team of quants, quants, quants, data scientists, data scientists, data scientists, data engineers, and data engineers, and data engineers, and traders working together to traders working together to build a build a decision-making engine. This is a decision-making engine. This is a decision-making engine. This is a whole staff of very whole staff of very whole staff of very expensive and expensive and expensive and highly qualified highly qualified highly qualified specialists. However, with specialists. However, with specialists. However, with the advent of AI, these the advent of AI, these the advent of AI, these tools have made tools have made tools have made you believe that you believe that you believe that you don't need these people anymore you don't need these people anymore you don't need these people anymore . You . You . You only need one only need one only need one person and a few API person and a few API requests to replace requests to replace requests to replace this entire staff of this entire staff of this entire staff of specialists. But here's a specialists. But here's a specialists. But here's a gap that no gap that no gap that no one really one really one really talks about. When you've talks about. When you've talks about. When you've spent a weekend spent a weekend spent a weekend building a dashboard, the building a dashboard, the building a dashboard, the moment you moment you moment you want to launch want to launch want to launch it into production and it into production and it into production and share it with someone share it with someone share it with someone else, you need to else, you need to else, you need to think about security, think about security, think about security, database, storage, database, storage, database, storage, reliability, and reliability, and reliability, and most importantly, most importantly, most importantly, credits. How do you actually

credits. How do you actually credits. How do you actually manage these manage these manage these systems? We see systems? We see systems? We see that pricing that pricing that pricing has changed. People has changed. People has changed. People no longer pay for no longer pay for no longer pay for tools, they tools, they tools, they pay for each pay for each pay for each individual API request. So individual API request. So , if everyone is now , if everyone is now , if everyone is now creating the same creating the same creating the same tools, what tools, what tools, what makes them different? makes them different? makes them different? There are two types of There are two types of There are two types of instruments in finance. The former instruments in finance. The former are are are data extraction tools, and the data extraction tools, and the data extraction tools, and the latter are latter are data transfer tools. And data transfer tools. And if you query a if you query a if you query a model, you yourself model, you yourself model, you yourself become the context. You become the context. You become the context. You explain to her, you explain to her, you explain to her, you argue with a argue with a argue with a model frozen model frozen model frozen in time, who tells you, in time, who tells you, in time, who tells you, "Hey, go to bed at 2 "Hey, go to bed at 2 "Hey, go to bed at 2 a.m., it's late." Although it's a.m., it's late." Although it's a.m., it's late." Although it's actually 3:00 PM on actually 3:00 PM on actually 3:00 PM on Wednesday. And all of this, when Wednesday. And all of this, when Wednesday. And all of this, when compared to compared to compared to human reasoning, human reasoning, human reasoning, simply does not stand up to simply does not stand up to simply does not stand up to any comparison. And any comparison. And any comparison. And then you have a then you have a tool that tool that sends you data itself. This is a sends you data itself. This is a sends you data itself. This is a tool that tool that tool that floods you with floods you with floods you with information, and that is already a information, and that is already a data engineering problem. So when data engineering problem. So when you have these two you have these two you have these two tools, you're tools, you're tools, you're solving different solving different solving different problems: one is problems: one is problems: one is context, and the other is context, and the other is data volume. You have to data volume. You have to data volume. You have to think about what think about what traders actually use and what they traders actually use and what they think about each of think about each of think about each of these tools. And these tools. And these tools. And one thing that we one thing that we one thing that we see imperceptibly see imperceptibly see imperceptibly breaking down in breaking down in breaking down in finance is when you finance is when you finance is when you go into a bank or go into a bank or go into a bank or fund, sit next fund, sit next fund, sit next to a trader or a to a trader or a to a trader or a team, and absorb team, and absorb team, and absorb everything that's everything that's everything that's going on. You going on. You going on. You learn from them and learn from them and learn from them and just listen to everything around you just listen to everything around you just listen to everything around you . You keep . You keep . You keep one ear on the pulse of your one ear on the pulse of your one ear on the pulse of your work, and the other work, and the other listens to listens to listens to everything that is everything that is everything that is happening around you happening around you . And it is . And it is . And it is consistency that consistency that consistency that brings results brings results . And if everyone . And if everyone . And if everyone manages their own

manages their own manages their own team as a “ team as a “ team as a “ one-man team,” one-man team,” one-man team,” what will happen to the what will happen to the what will happen to the future generation future generation that has to learn that has to learn that has to learn this context? And this is the this context? And this is the this context? And this is the real price that no real price that no real price that no one is one is one is taking into account right now. Therefore, in taking into account right now. Therefore, in taking into account right now. Therefore, in finance, you cannot finance, you cannot finance, you cannot simply hand over the decision-making cycle simply hand over the decision-making cycle to a team of agents. Why ? Because there ? Because there ? Because there is a client on the other side of the deal. If is a client on the other side of the deal. If is a client on the other side of the deal. If it is it is it is raw material trade, then it is raw material trade, then it is raw material trade, then it is cargo, physical cargo, physical cargo, physical cargo. If it's a cargo. If it's a cargo. If it's a company, then it's a company, then it's a company, then it's a team of people. And so team of people. And so team of people. And so when something when something when something goes wrong financially, you goes wrong financially, you goes wrong financially, you need someone to need someone to need someone to come out and say, "I come out and say, "I come out and say, "I made this decision for made this decision for made this decision for such and such reasons." such and such reasons." such and such reasons." If the agent grinds If the agent grinds If the agent grinds your API key, they your API key, they your API key, they simply say, “Oh, simply say, “Oh, simply say, “Oh, sorry. I shouldn't have sorry. I shouldn't have sorry. I shouldn't have done that. done that. done that. Please change Please change Please change it." But in finance, it." But in finance, it." But in finance, this is not possible, because the agent this is not possible, because the agent this is not possible, because the agent tries tries tries to correct itself, and to correct itself, and to correct itself, and one small mistake one small mistake one small mistake begins begins to grow exponentially. So, this is the to grow exponentially. So, this is the question that question that question that concerns me: if the human concerns me: if the human concerns me: if the human brain is the most powerful brain is the most powerful brain is the most powerful tool, why do we tool, why do we tool, why do we spend so much spend so much spend so much effort trying to effort trying to effort trying to push it out of the push it out of the push it out of the process? Why don't we process? Why don't we process? Why don't we change our approach and change our approach and change our approach and instead of replacing it, instead of replacing it, instead of replacing it, start start start fueling it? If fueling it? If fueling it? If you think about what's in a you think about what's in a you think about what's in a trader's head, trader's head, trader's head, they hold everything that they hold everything that they hold everything that leads up to the trade itself.

leads up to the trade itself. But before that, there is a huge amount of huge amount of other work to be done. They other work to be done. They other work to be done. They read, gather read, gather read, gather information, information, information, check facts, check facts, check facts, they monitor the they monitor the they monitor the market. Because market. Because market. Because we have so we have so we have so much information now, much information now, much information now, the most important question that the most important question that people ask is: where is the truth? What people ask is: where is the truth? What do do do different sources actually say? What different sources actually say? What different sources actually say? What is the discrepancy is the discrepancy is the discrepancy in this information? How in this information? How in this information? How can I trust what can I trust what can I trust what is around me? And when I is around me? And when I is around me? And when I have done all this, only have done all this, only have done all this, only then do I have a solution. then do I have a solution. So, if we So, if we divide this into two divide this into two divide this into two parts, we parts, we parts, we will understand that will understand that will understand that part of the work is part of the work is part of the work is routine, reading, collecting, routine, reading, collecting, routine, reading, collecting, cleaning data, and cleaning data, and cleaning data, and the other part is the other part is the other part is judgment, that is, judgment, that is, judgment, that is, making a decision. And making a decision. And making a decision. And machines are incredible at machines are incredible at machines are incredible at this routine. They are this routine. They are this routine. They are incredibly skilled at incredibly skilled at incredibly skilled at cleaning data. They are cleaning data. They are cleaning data. They are incredible at collecting incredible at collecting incredible at collecting data. And the man is data. And the man is data. And the man is incredible at incredible at incredible at making decisions. So making decisions. So making decisions. So why not just why not just why not just split it into two split it into two split it into two separate parts? Which separate parts? Which separate parts? Which brings us to brings us to brings us to Moshi, which we created. Moshi, which we created. Moshi, which we created. This is a 24/7 This is a 24/7 This is a 24/7 quantum analyst quantum analyst quantum analyst for institutional for institutional for institutional traders. This means traders. This means that we run a that we run a that we run a team of six team of six team of six different agents, different agents, different agents, built in a built in a built in a deterministic deterministic deterministic way, that way, that way, that monitor the monitor the monitor the market 24/7, analyzing market 24/7, analyzing market 24/7, analyzing more data than a more data than a more data than a human could, human could, human could, plus investor data plus investor data plus investor data contained in contained in PDFs, PDFs, PDFs, emails, various emails, various emails, various subscriptions, understanding it subscriptions, understanding it subscriptions, understanding it all in a way that all in a way that all in a way that helps the trader helps the trader helps the trader make better make better make better decisions. So, Moshi decisions. So, Moshi decisions. So, Moshi read 2.4 billion read 2.4 billion read 2.4 billion data points. And in the 4

data points. And in the 4 data points. And in the 4 seconds it took you to seconds it took you to seconds it took you to read that sentence, Moshi read that sentence, Moshi read that sentence, Moshi learned 300 new learned 300 new learned 300 new things. And these 300 new things. And these 300 new things. And these 300 new things she things she things she learned are a new learned are a new learned are a new context for the context for the context for the model. The problem is model. The problem is model. The problem is that new that new that new context is context is context is worthless if you don't worthless if you don't worthless if you don't know: can it be know: can it be know: can it be trusted? Where does it trusted? Where does it trusted? Where does it come from? Do come from? Do come from? Do I even value this I even value this I even value this analyst's opinion? Because in analyst's opinion? Because in analyst's opinion? Because in finance, even if finance, even if finance, even if it's a very confident it's a very confident it's a very confident answer, if it's answer, if it's answer, if it's wrong, it's wrong, it's wrong, it's worthless. So what do worthless. So what do worthless. So what do we do? We we do? We we do? We measure it. We measure it. We measure it. We measure measure measure hallucinations. We hallucinations. We hallucinations. We verify every verify every verify every statement. We statement. We statement. We are looking at how far are looking at how far are looking at how far this can be traced. And this can be traced. And this can be traced. And we also provide we also provide we also provide the user with the the user with the the user with the assumptions we assumptions we assumptions we had and explain how had and explain how had and explain how we came to that we came to that we came to that decision. After all, if decision. After all, if decision. After all, if an investor asks for an investor asks for an investor asks for your opinion, he wants your opinion, he wants your opinion, he wants to know not only what to know not only what to know not only what you think, but also how you think, but also how you think, but also how you came to such you came to such you came to such conclusions. Therefore, we conclusions. Therefore, we conclusions. Therefore, we leave leave leave decision-making to the person, decision-making to the person, decision-making to the person, as we believe that this is what as we believe that this is what as we believe that this is what people do people do people do best. There is no best. There is no best. There is no machine more powerful than the machine more powerful than the machine more powerful than the human brain. But human brain. But we provide everything else you need for work. We believe we provide everything else you need for work. We believe that every firm that every firm that every firm needs entire needs entire needs entire departments of quants, departments of quants, departments of quants, engineers, and engineers, and engineers, and data analysts, and that's data analysts, and that's data analysts, and that's exactly what we exactly what we exactly what we provide you. Because it is impossible provide you. Because it is impossible provide you. Because it is impossible to provide this for to provide this for to provide this for each employee each employee each employee individually. It is impossible individually. It is impossible individually. It is impossible to maintain your own to maintain your own to maintain your own team of quants and team of quants and team of quants and data scientists, and data scientists, and data scientists, and it makes no sense. From an it makes no sense. From an it makes no sense. From an economic point of economic point of economic point of view, this is too view, this is too view, this is too expensive for each expensive for each expensive for each individual specialist. And individual specialist. And individual specialist. And so, this is what we so, this is what we so, this is what we do. We do. We do. We apply apply institutional-grade analytics , like those used by large

, like those used by large , like those used by large financial institutions financial institutions , to give you , to give you , to give you the confidence to make the confidence to make the confidence to make better deals. So what does better deals. So what does better deals. So what does this actually give you? this actually give you? this actually give you? This gives you up This gives you up This gives you up to 35% to 35% more efficiency, so you can more efficiency, so you can more efficiency, so you can spend your time and have spend your time and have spend your time and have the space to think, the space to think, the space to think, to more confidently to more confidently to more confidently execute the execute the execute the deals you need. If you deals you need. If you deals you need. If you talk to talk to talk to a lot of traders, a lot of traders, a lot of traders, they’ll say, “Hey, I they’ll say, “Hey, I they’ll say, “Hey, I don’t want to don’t want to don’t want to talk to anyone all day. talk to anyone all day. talk to anyone all day. I just want to I just want to I just want to trade." So why trade." So why trade." So why give these people just an give these people just an give these people just an LLM so they can LLM so they can LLM so they can discuss, "Hey, is the discuss, "Hey, is the FOMC meeting today or tomorrow?" And then FOMC meeting today or tomorrow?" And then LLM says tomorrow, LLM says tomorrow, LLM says tomorrow, even though it's today. It's even though it's today. It's even though it's today. It's just a complete waste of just a complete waste of just a complete waste of time. And this is exactly the time. And this is exactly the time. And this is exactly the opposite of what opposite of what investors want. They investors want. They need something that will need something that will need something that will give them the power to give them the power to give them the power to think more and think more and think more and use use use their own intelligence. their own intelligence. their own intelligence. So, Mosaic is not you. So, Mosaic is not you. So, Mosaic is not you. Mosaic is not a model. It's Mosaic is not a model. It's Mosaic is not a model. It's you "on steroids", with the you "on steroids", with the you "on steroids", with the same same same instincts, but instincts, but instincts, but fueled by all the fueled by all the fueled by all the data data data scattered in scattered in scattered in different places. So you different places. So you different places. So you can spend your can spend your can spend your time making time making time making decisions, not decisions, not decisions, not arguing about arguing about arguing about when the FOMC meeting is. In when the FOMC meeting is. In when the FOMC meeting is. In terms of scale, terms of scale, terms of scale, we track over we track over we track over 2.4 billion 2.4 billion 2.4 billion data points, and it's data points, and it's data points, and it's growing growing growing every millisecond. This is every millisecond. This is every millisecond. This is a decade of knowledge a decade of knowledge a decade of knowledge about derivatives trading about derivatives trading 24/7—something no 24/7—something no one person could one person could one person could build on their own build on their own . Next is sending . Next is sending . Next is sending data. When you data. When you data. When you pull data from LLM, pull data from LLM, pull data from LLM, you give it you give it you give it context based on the context based on the context based on the old model. When old model. When old model. When you are sent new you are sent new you are sent new data, it forces your data, it forces your data, it forces your brain to think about new things brain to think about new things brain to think about new things and make connections that

and make connections that and make connections that you hadn't you hadn't you hadn't thought of before. This means thought of before. This means thought of before. This means you think about the “ you think about the “ blind spots” blind spots” blind spots” you’ve been told about and you’ve been told about and you’ve been told about and try try try to figure out, “Okay, I would to figure out, “Okay, I would to figure out, “Okay, I would n’t have noticed that, so n’t have noticed that, so n’t have noticed that, so what do I do with this what do I do with this what do I do with this new information?” And new information?” And new information?” And then—deep then—deep then—deep analysis, the ability to analysis, the ability to analysis, the ability to dive dive dive into one thing in detail with a into one thing in detail with a into one thing in detail with a 360° overview of information that 360° overview of information that 360° overview of information that you haven't you haven't you haven't had before. We do this had before. We do this had before. We do this with big with big with big data and all the data and all the data and all the specific data. specific data. How is that different How is that different from an instrument from an instrument from an instrument you build on the you build on the you build on the weekend? Because a weekend? Because a weekend? Because a tool that has to tool that has to tool that has to work 24/7 in work 24/7 in work 24/7 in every every every big data market, big data market, big data market, working in a way that you working in a way that you working in a way that you understand context, understand context, understand context, being reliable and being reliable and being reliable and trustworthy, becomes trustworthy, becomes trustworthy, becomes infrastructure. It's infrastructure. It's infrastructure. It's not a weekend tool anymore not a weekend tool anymore not a weekend tool anymore , because you , because you , because you have to be sure have to be sure have to be sure that other people can that other people can that other people can trust it and you trust it and you trust it and you can track the can track the can track the result, because that's what result, because that's what result, because that's what you have you have you have with a person: someone can with a person: someone can with a person: someone can stand up and say, stand up and say, stand up and say, "Hey, I made this "Hey, I made this "Hey, I made this decision because of XYZ." And decision because of XYZ." And decision because of XYZ." And not call an agent not call an agent not call an agent who never...who who never...who who never...who will just will just will just apologize to apologize to apologize to you. So, one you. So, one you. So, one question that I do question that I do question that I do n't think we've n't think we've n't think we've answered yet is: if the answered yet is: if the answered yet is: if the real advantage real advantage real advantage resides in the human resides in the human resides in the human brain, how much of brain, how much of brain, how much of it can a it can a it can a machine actually grasp? And I do machine actually grasp? And I do machine actually grasp? And I do n't think we've n't think we've n't think we've answered that yet, um, and answered that yet, um, and answered that yet, um, and I think whoever I think whoever I think whoever answers that is answers that is answers that is not creating a product, not creating a product, not creating a product, they are creating a new they are creating a new they are creating a new way of thinking in way of thinking in way of thinking in finance, and that's what we're finance, and that's what we're finance, and that's what we're trying to trying to trying to explore. So, we've

explore. So, we've explore. So, we've been working very quietly with been working very quietly with been working very quietly with some of the biggest some of the biggest some of the biggest players in the trading players in the trading players in the trading industry in finance, industry in finance, industry in finance, and so this is the first time we're and so this is the first time we're and so this is the first time we're openly communicating openly communicating openly communicating with the public, because with the public, because with the public, because we believe that we believe that we believe that everyone should have everyone should have everyone should have access to access to access to institutional-grade analysis institutional-grade analysis without having their own without having their own data science or quantitative data science or quantitative analysis team. Um, so if analysis team. Um, so if analysis team. Um, so if you're interested, you're interested, you're interested, come talk come talk come talk to me. I love to me. I love to me. I love talking about everything talking about everything talking about everything related to data, related to data, related to data, finance, and finance, and finance, and financial financial financial engineering. Um, and yes, engineering. Um, and yes, engineering. Um, and yes, before we start before we start before we start trading, we want trading, we want trading, we want people to talk to people to talk to people to talk to Moshi. Thank you.