Artificial Intelligence (AI) Panel - 361Firm's NY Tech Summit Feb. 25, 2025

Artificial Intelligence (AI) Panel - 361Firm's NY Tech Summit Feb. 25, 2025 SUMMARY KEYWORDS Artificial Intelligence, generative AI, venture capital, seed funding, Hippocratic, LLM, job displacement, AI revolution, energy solutions, food security, humanoid robots, quantum computing, stakeholder model, economic impact, technological advancement. SPEAKERS Speaker 1, Alex Zhuk, Rashmi Joshi, Ben Narasin, Speaker 2, Maisy Ng, Mark Sanor, Zoe Cruz   Mark Sanor  00:00 Um, introduce yourself again, share an insight and what scares and excites you about AI, there you   Ben Narasin  00:06 go. What is this the one? All right, Hi, I'm Ben Naris, and I run tenacity venture capital. I spent of any a large venture firm about three years ago. I focus on seed. I've been doing that for about 18 years. Last year, I saw 2000 companies. I funded three. They were all generative. Ai related. It it is not because I have an explicit focus on AI. By the way, focused funds of under perform generalist funds for 40 years, find that data out there and think about how you invest. It's that the best and brightest always go to the shiniest, most exciting thing, and that is certainly generative AI right now, I even have I paid them personally, the most I've ever paid for company, $500 million for a company called Hippocratic, which is creating LLM based nurses. And what's fascinating about AI, I guess there's so many things, but one, we don't know how it works and how it thinks. These machines are thinking. And people that are in the business will acknowledge they don't actually understand how, two, it totally changes the value and reality of time. So let's use the example of Hippocratic they have an LLM that is trained on the nursing notes from major medical facilities. It calls in audio every person that leaves a hospital or doctor's office and checks in on them to make sure they are staying in tune with the things they need to do to get better. When in the past, would that ever have been possible? You know, 1000 people leaves a hospital in a day, there is zero chance you can afford to get the people to do it. But AI can spend infinite amounts of time and spin up infinite instances, and it will totally change things that we are able to do. I make one more example of that. I was listening on calls that the AI made to different patients. And it called a woman that had diabetes, and it, you know, did its check in. And then she said, Hey, can I eat, you know, beans? Yeah, beans are fine. Can I eat bread? Well, bread is bad at spikes. And then she listed off one by 156, foods to see if they were okay to eat. And the AI, very patiently, said, yes, no, yes, no, that would never happen. But not only can the AI allow infinite time to be utilized to do things in parallel, but the people on the other side can take advantage of it in ways they never would have with a traditional nurse in this instance. So I think there are going to be so many things that happen that we are not expecting. I am not worried about I am a little worried about the single purpose tool labor, the person that is not able to be retrained well, because that's not the culture they grew up in. They didn't value education there, you know? But hey, I walk down New York City streets today, a lot of people swing and sledge hammers and dig in dirt. There'll be plenty of things to be done. It's just if you have a single if you're that high or gal in the call center in Bangladesh. Woo. I hope you can find someone else.   Mark Sanor  02:45 Okay, so Maisie, also introduce yourself. And again, what scares, excites, insights.   Maisy Ng  02:53 Hi. My name is May Z. I'm founder, managing partner of the light capital. We're VC head quarter in Singapore. We're now doing our second fund, the first invest in Southeast Asia tech companies that celebrate the UN SDGs. The second fund will invest in AI companies across the AI tech stack. We're really excited about this opportunity because, I mean, AI is like a tech East dream, right? So, because it's so revolutionary, the same O, same o doesn't work anymore. We need a whole new class of semi conductors, data center technologies, new software that will empower new applications. So this is, we think it's like, I think the aircraft guy, this is once in 100 years as well. So, and what excites me, I think, well, the sort of paradigm shift that AI brings, it enables us humans to do things we never thought was possible. And initially, for example, when deep mine was started, it started by trying to play chess. And initially it basically took all the grand master strategies and train the software to play like a grand master. And so it played against grand masters, and they win some and they lose some, and then they decide, okay, fine, we just tell the computer, these are the rules, and you just go play. And because computers can basically, you know, like, work really fast, they could play, like, a million games overnight, and very soon they learn how to play. And then they did this go, which is a far more complex game than chess, and just by playing against itself, they found new strategies that Grand Masters would not think of like in the chess game. They be sacrificing pieces, left, right and Sanor, and then they win. And people just can't understand how they did that. And a couple of days ago, I read this article about scientists using AI to design basically micro wave circuits, and they said that the design that comes out looks really weird. It's not something that an engineer would design, because it's not something you've been taught in school. But so it looks really weird. Doesn't look like a circuit board, but apparently it worked better than any other circuit. So I think that is opportunity that we can have with AI. What? What scares me a bit to what Zoe said. I mean, someone once said that basically, software would eat the world. So guess what? Ai. Eat the software. And to Ben's point, people will lose jobs, and this is a major program shift. Some of the jobs aren't ever coming back, and so you gonna have, like, massive layoffs, and what people are gonna do so the consumption will drop, because people just don't have jobs they can spend. So I think the governments and the companies need to know and try to plan ahead, because the core, I guess, social compound we have capitalism is that if you make money as a capitalist, you are supposed to invest the money to create more jobs, build factories. But what we saw in the past decades is that people who made money from outsourcing globalization, they didn't build more factories. They did hire more workers. What did they do? They bought Yach, they bought art. And so all this rent seeking behavior didn't help the economy, and that is a problem. So if you take AI, that's going to be like compounded a train in times exponentially. So I think companies need to be aware of that. Governments need to be aware of that. It may be that we have to do either tax on robots or UBI just to what people picking up. Pitch Fox,   Mark Sanor  06:00 okay, let Alex go next.   Alex Zhuk  06:05 Thank you, Mark. Thank you for having me. Great to see all of you. I'm going to give you a very short introduction by myself, because I haven't met many of you. I'm a founder of an AI company that uses satellites to map the environmental footprint of every farm on the planet to help ensure food security through resilience, but also decarbonize agriculture, which is the second largest emitting sector in the world. I'm also on the side involved in critical mission asset development, primarily energy solutions and data centers, starting with building a digital twin of the electrical grid, because it's becoming very hard to connect to it, as many of now, in terms of an insight that I think hasn't been shared by these experts near me, I think we are under appreciating, or at least I did for a very long time, the way in which industries that have been established as part of humanity's operation for 1000s of years will be disrupted. So I work in agriculture. We've been farming in a mechanized, or at least structured manner for centuries, but you could argue 1000s of years we are actually for writing of climate reasons, but also just the way we've been farming since the 19th century, are on track to erode the size of arable land. It's about the size of Latin America, which puts in tricky position, especially with a growing global population, right? What do you mean by a road? So the way we farm, we've been farming for past 100 years is we've been blank to chemicals non stop on the soil, mechanically turning it over same crops. And what we found recently is that process over time kills us well. Now the question is, how do you deal with that? One way is to improve how we've been farming before. So precision agriculture, but you know, there's a completely different paradigm on hand, right? So, much like a century ago, in order to get a diamond, you would go down a mine shaft, you would dig it up, you would clean it, you process it, you ship it over. Now you can start with a kernel of carbon and grow it, right? Similarly, for example, with meat, we're getting to a point where we can grow real patties that are juicy, feel more or less the same taste and a real meat in a lab, what the consumption and the water and the energy needed to raise through animals at scale. So I think it's an opportunity in that AI can provide real resource abundance and a quality of life for each and one of us in terms of volume, that is fundamentally different from how we've been approaching it as humanity for hundreds of years. The question comes back to actually something you mentioned and several other panelists, which is, how do we tackle the social question, and how do we deal with the tension if the haves, if the gap between the haves and the have nots increases far greater than we've seen before.   Mark Sanor  09:17 So thank you, Alex, somewhat hopeful, maybe, maybe. Zoe, you're now on an AI panel. If you stick around, you could be on a health tech panel. What are your thoughts on on AI specifically scary and exciting. I   Zoe Cruz  09:35 mean, to me, this young man is Exhibit A why it all is going to be very good again. My concern is the transition. And right now, the way we allocate capital to wonderful things like AI is in at the traditional paradigm, which is, you know, stocks and bonds go up if x. Why, you see, there's a paradigm. I went and re read actually, and that's where AI is helpful. There is a book that was written in 1955 and it's basically the Structure of Scientific Revolutions. And it was the first time they talked about paradigm shifts. And in that they said scientists do a lot of work in a particular paradigm, and then Copernicus says, no, no, no, the sun doesn't go around the Earth. It's the other way around. That's a paradigm shift. So you do something different. So for me right now, as my 29 year old son says, technology exists to take the carbon out of the air to even get these meteorites to go off. The technology exists. How do you deal with the existing capitalist model, where you have existing capital allocated to things that will go to zero? So I do believe this is something spectacular and exciting, but I can't put the two and two together. That comes up with four in terms of regular transition. And you know, one of the things I said to my son, because the world is now run by HEPA gene octogenarians, never mind heptogenarians, why don't you guys get more involved? I mean, he's a brilliant young man. He started evolutionary biology. He plays the classical piano. He should get involved. And you know what he said? Talk about socio economic issues. What's the point? We have to wait until you guys die off. Now he didn't mean me, but hopefully, but   Mark Sanor  11:42 so we were in Germany at a round table, and apparently there's, there's legislation afoot to reduce or incentivize you to reduce voting at later ages. So you've sort of heard the panel, if you guys want to make some comments. But otherwise, I started opening   Ben Narasin  12:04 it up to comment on something I very much disagree with. Maisie. I don't think the evidence is that people are greedy, venal yacht buyers. I think it's quite the opposite look at so I look spend a lot of time in trucking space. Trucking is the number one job in the world by head count, although nobody wants to do it anymore, and there's an issue with aging out, etc, etc. But I was very concerned for a long time, because I was also looking at autonomous trucks and the massive displacement number one job in the world by head count, it should be done by machines. Okay, these people are out of work over time. So I started looking backwards. And one of the great example. See what the very one of the very first commercialized robots was the card scanner at the gas station. Now, if you're unfortunate enough to live in New Jersey and drive a car, you are in one of two states that unions which, by the way, I could not despise an entity more than I despise union. So please, no union leaders here have insisted that a human being pump your gas, an incredibly inefficient experience that drives me insane whenever I'm forced to deal with it. By the way, yesterday I was in a apartment. We were looking at buying an apartment here, and they have a man who pushes the button in the elevator like talk about it doesn't matter how much we do, the unions will make sure people have ridiculously stupid jobs and get paid. So anyway, what happened with all that great wealth that was created because now they didn't have to employ people to pump the gas in 48 states the United States, did people just stick with what they were doing? Absolutely not. They created what is now known as the convenience store when you go to a gas station, instead of just having gas pumps, which back then was all there was, maybe a counter with gum and candy, full fledged stores with all kinds of food and drinks and slushies, those stores ended up employing more people than the gas station attendant jobs represented same thing with the ATM everybody said, Oh my God, all these banks, the tellers ought to work more banks today than there Were with ATMs. I   Mark Sanor  13:59 think maisie's Point was different, because and Esther again, Esther Dyson asked, What billionaires, you know, have become better people in the last 1020, years, some have, but we, of   Ben Narasin  14:13 course, remind me, exaggerates you. It doesn't change you.   Rashmi Joshi  14:18 Hi. Thanks for that. I have actually three questions, so you might have to come back to me in a bit, but I'm curious, as an AI founder myself, what industries or new verticals Do you feel like are going to be established as a by factor or a consequence of us getting rid of all of these mundane tasks and grunt work type of jobs?   Alex Zhuk  14:42 Sure, happy to so the near term industry that has gone from, I would say, sort of in the shadows, a little bit boring, to very exciting. That was obviously energy. So we're realizing that if we're in a race at international level, we. Can't afford to lose, to concentrate now, as to how do we power these machines, both to train the models, but also humanoids, once automation is commercialized, which we're seeing happening very rapidly, that's exciting. How that will be solved, whether it's nuclear, whether it's other source of energy, is a guessing game, but that's a very exciting space. We haven't seen this growth infrastructure in decades. Personally. You know, I mentioned example of how we can similar to how we can synthesize proteins for medicine, create new foods, right? So, there is a company that was able to create cow free milk, and they tasked an AI to come up with ingredients that would when combined, taste, smell and feel like milk. And when you know, you might wonder what those ingredients were. Those were pineapple and strawberries, right? So ingredients are completely unintuitive to the human mind, that when combined, we're able to synthesize something that we want to consume. And I think we'll see that across food, I think we'll see that across health care.   Mark Sanor  16:03 But those are interesting vectors. But I think your question was the people, sort of your earlier point about job, you know, people who are going to be out of out of jobs, was your question like, Where will they be going? Where should they be where's the puck going for people? Is that it   Alex Zhuk  16:20 very difficult question for me to ask Mark, I would say the best bet would be for the verticals that are growing the fastest,   Mark Sanor  16:29 or maybe this goes back to Steven SPI about education. Anyone else want to answer that skill set   Speaker 1  16:38 would be, oh, I will cycle into something different, maybe more productive, just like, Well,   Ben Narasin  16:41 that was a great example. One of my one of my founders, made the point we brought a YPO group in, and he said, you know, you were talking about farming before the Civil War, 90% of the US population farmed. So we have seen a massive wipe out of an entire population of workers before it was all of America, but then they moved to cities. And guess what? When you're on a farm, you don't cut you cut your own hair. So all kinds of jobs were created that didn't exist when we had a mono culture of farming as the primary job, hair cutter, barber being one, and there were infinitely more. I think, by the way, if we could answer your question, we wouldn't tell you, because we'd be investing in at least two of us would be investing in it right now to get ahead of it. Yeah, well, yeah, I'm you must not have met many VCs, because we're very greedy in the first round to get all the ownership we can. That's the only chance we get. But it's, I think it's unpredictable, but I'm not worried that it won't happen. I think that, look, we have been through this before. The difference is that this is the first time software ever attacked the labor force instead of just process. But the labor force has been attacked many, many times. I mean, the Luddites are obviously the most commonly quoted example. But you know, it's like labor is lake water. It flows to the place it's needed. I do have material concern about, I'll just say, because I'm not gonna go too deep and dark here certain populations that might not have the historic advantage of or desire to reinvest in their own education. And I think that sometimes it's unrealistic for highly educated people to believe that everybody can be re educated, and that they'll even want to be and so where does that end up going? But here you want hope there's 100,000 unfilled jobs the United States right now in construction that are paying over $100,000 it's a good place to start. There's many places where jobs are unfilled. And lastly, a lot of the AI will augment people's ability and take over jobs that aren't filled, that are wanted and needed. As someone once said, You're not at threat of a of being your job being taken by AI. You're at threat of somebody that's better of using AI, taking your job.   Maisy Ng  19:01 I think I might have mis understood your question. So if you allow me, I'll give you a misunderstood answer. So I think there's, I mean, AI could be used also for, like robotics. So for example, I think, you know, we have really seen from Boston Dynamics that like dancing robots, but that isn't too useful for most of us. You don't buy a dancing robot. But a couple days ago, I saw this really interesting video. I think it's a US company that has basically built robots that can be used for domestic work. So can you imagine a robot that cleans your house? And this one was cool. So there's like two humanoid robots and standing side by side, and basically the owner comes in and gives them a bag of groceries, and the robot just look at them, and they sort it out. And if they took up a ketchup and they know its ketchup, they put at the top shelf of the fridge, they open it and they see there's eggs. And one robot picks up the eggs very gently, hands it to the other robot, who then puts it in the fridge. I mean, that's pretty cool, because you need computer vision. You also need an LL. Am, and you know, you can train a domestic robot for all scenarios, right? So the robot has to know that if it's an egg, you handle carefully, and this may be a quills egg, so he would know to the LLM that is a quills egg, it's an egg, so I handle it gently as well, so that, I think would bring tremendous, I mean, advantage for us, because nobody wants, you know, to do housework these days. Can I   Rashmi Joshi  20:20 just piggyback off that for a   Mark Sanor  20:24 second? One second, piggyback on the mic.   Rashmi Joshi  20:27 So as humanoid robots become more and more similar to us, let's say I can build you a robot that would be your perfect husband or partner, right? And it's indistinguishable from the real thing.   Maisy Ng  20:44 I think I can distinguish that   Rashmi Joshi  20:47 today, sure, but maybe five years from now, maybe not, right? So my question is, then, what is the value in being human?   Maisy Ng  20:57 I think we still have a soul, which I don't think that. I mean, we could probably train the robots at some point, but I don't know, it's a tough question to answer. So I think, I mean, that's something that we had discussed internally as well. I mean, so do we teach robots about, you know, like life after life and so forth? I mean, do do when you   Mark Sanor  21:17 say So internally? I mean, your fund internally discuss this friends   Maisy Ng  21:20 and within the partners and so what it means to be human, and basically, what do we need to teach, you know, the robots and so forth. So I don't know. I mean, it's an honest answer. I really don't know good to see how it goes,   Zoe Cruz  21:33 because I'm gonna leave after this. Are you gonna drop the mic and just go? What an amazing question, in the sense that, first of all, the idea that I'm going to have this made in my home, this robot that I can't control, that somebody else actually can control, I don't know that I'm going to get to that dysfunction. To me, we're not again, we don't need we can take off the table. How amazing AI is going to be. Let's take it. It's not. You don't need to argue it. It's going to be amazing. Okay, the land of plenty. This thing about human beings, my experience at Morgan Stanley was, if you in the ability of human beings to do amazing things if you inspired them, is mind boggling. If you inspired them, that's what humanity is. And so this idea that we're going to replace human beings, you're going to build me the perfect partner. No, thank you. What I want to ask again, of all of us, why is it that we talk in terms of the stakeholder? We're talking about is the shareholder of a company that's going to make a lot of money because they're going to fire employees, and therefore productivity is going to go up, and therefore you're going to be rich. That's basically the discussion. Yes. Now the old capitalist system that I started growing up in as a young, you know, graduate of a business school was you had three stakeholders as a company, shareholders at the head of the que, clearly, your employees and your community, those were the stakeholders. And I think how we got to the only stakeholder in any kind of for profit organization is your equity holders. Is what stops us from doing inspiring things. I'm not inspired to be rich or they say the shroud has no pockets, so when you're six feet under, it doesn't matter whether you are multi billionaire or sent a millionaire. Did your life make a difference? So with that, sorry.   23:53 Thank you.   Ben Narasin  23:54 Just one comment on humanoid robots. I mean, Japan has been trying to do humanoid robots for decades. It is not clear that human beings want them, and I'm looking think about your eggs, example. So what's better a humanoid with two hands and two feet, or an octopod, pod like creature that has eight you know, building for functionality will ultimately so you'll go back to one thing. You wanna know, it really scares me. So I was a writer for 10 years. I got a lot of freelance of freelance writing. I want to write a science fiction book on the following. Jump forward 10 years. Quantum works. Okay? I don't know how many of you spent time looking at Quantum. We have no flipping clue what it can do, right? It changes everything. And the only thing we worry about is end point, security. Well, how about literally everything else? It's things differently than human beings find ways to do things that we would never consider okay. So now we're 10 years forward. We're at chat GPT 10. Now someone express some optimism that China and the United States would get together for some positive Oh, hallelujah moment, which, yeah, good luck with. That I'll take 10 to one odds against it happening. China wants to replace us, not to be our buddy. So now you take chi and you take Putin. They, you know, probably two of the richest people on the planet, considering certainly how Putin has raped this country of its capital. And they each put a half a trillion dollars in a bucket, and they build out the largest data farm in the world that runs entirely quantum computing. And they bring in all the best people who, by the way, if they don't perform, get a bullet in their head and get buried in the back yard. And they get them to run the newest issues of chat, GPT, and they ask that system, that trillion dollar system, do just one thing, figure out how to destroy the United States. That's what I worry about. I hope we can stay strong enough that we have a really good chance. And while I'm not a political person and we, you know, the pendulum is a nightmare, we will spend well on defense. We will allow AI to flourish. And if we're not a leader, we have a very good chance of being a distant 12th 10 years from   Mark Sanor  25:57 now. And what's your last thought, Alex, actually, you're going to stick around because you do AG, so the panel you originally on, you'll stay, you'll stay for and rash me is going to come up along with Chris, and we're at two. This is why there's an AI for that. There's not an AI for my glasses. 229, so last, any last questions or thoughts for AI? Yes, sorry, David,   Speaker 2  26:27 so Alex, love what you doing. The thought is, you know, 50 years ago, there's probably people in a room, and they were talking about how spectacular we'd gotten at crop farming and the use of these fertilizers and this mechanization, all the stuff that's now proven problematic at that point seen ground breaking. What are your views on how we've grown in terms of thinking about the how of technology and being able to mitigate for all of because everything has trade off, so everything has unforeseen circumstances. Are we just plowing ahead, same as we did 6070, years ago, expecting perfect results, when actually we've seen that. That doesn't often happen. That's   Alex Zhuk  27:06 a fantastic point. And to give context to that comment, you know that process, which is the HP process, which allowed us to manufacture these chemicals for farm and very cheap and scalable, did prevent famines, and, you know, solved a lot of issues at the time were post World War Two, especially, really pressing. I think today, there's a component of that, which is, there are problems we can see in the near term, and it's extremely appealing to solve those at the expense of, you know, something we will have to figure out later down the line. And I will also compound on the comment I've heard earlier I can remember who mentioned it, which is that both great powers Today, China and America, realize that in particular, the AI race is the new nuclear race, and it's a race neither one of them can totally afford to lose, and the importance of which supersedes profits. So you combine that dynamic with where we today, and I don't see not only any one of us stopping, but how we could, even in the in national interest, slow down our progress given the dynamic internationally. Hope that answers your question, do you   Mark Sanor  28:22 want to hit that or you good. There's one other thing that I think you all should know. You all know open ai, llms, just give, give 3060, seconds on, on pricing model,   Maisy Ng  28:34 right? Yeah, we are investing in a new company that does the world's first large pricing model. So basically, there's lot of content in the world, but there's no price on it. So this company has figured out a way how to price different content. So just like you train an LLM with text input, I mean, with lots of text, so that you can figure out, using transformer model, what's the probability of the next word, and therefore, in doing so, be a performance sentence and reply to a query. So basically, LM has been trained on copious amounts of text to give you an answer when you input a tax query. So what these guys have done is, again, they've trained the large pricing model on a huge amount of content. And instead of figuring out a tax output, what it does is, when then confronted with a content input, it can then spit out the monetary value of that content. And so the use is immense. Because right now, if imagine, if I go to farmers market that was sharing this angle, we don't have time for the farmers market, but they can price any content.   Mark Sanor  29:31 But the point is this, this is, yeah, this is another new frontier that I think, is talk to talk. We'll be having round tables very soon. So thank you to this panel. We appreciate it. Alex, stay I'm joined our 361 firm community of investors and thought leaders. We have a lot of events created by the community as we collaborate on investments and philanthropic interests. Join us. You. You can subscribe to various 361 events and content at https://361firm.com/subs. For reference: Web: www.361firm.com/homeOnboard as Investor: https://361.pub/shortdiagOnboard Deals 361: www.361firm.com/onbOnboard as Banker: www.361firm.com/bankersEvents: www.361firm.com/eventsContent: www.youtube.com/361firmWeekly Digests: www.361firm.com/digest