$22B Co-Founder: What the Market Predicts for Your Job in 2026 | Luana Lopes Lara

0:01This episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome, that's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50-page restoration block, or finally break down that long article you've had open for weeks. Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses set up required, compatibility and availability varies 18+. When you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications and more.

0:38Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a$75 sponsored job credit at Indeed.com slash podcast. That's Indeed.com slash podcast. Terms and conditions apply. Need a hiring hero? This is a job for Indeed Sponsored Jobs. 73 % chance that AI will be number one reason for job cuts. Like how much should I be relying on public opinion versus reality? Even with 5 ,000 in volume, you already see convergence to a very calibrated number. So that number should be trusted for sure. This is Luana Lopez-Lara. Forbes calls her the youngest self-made woman billionaire in the world.

1:15She built Kalshi, a$22 billion company where anyone can look up the odds on things that haven't happened yet.

1:29Can we make some predictions? Let's do it. What actually changes for a normal person by the end of this year? I still think that by the end of the year, we're going to see work change the most. Do you think 2026 is going to feel lighter or heavier for the majority of people? That is a tricky question. Do you see any jobs suddenly becoming safer? For now, I would actually claim that. Juana, welcome. Oh, thank you. Thank you so much for doing this. Of course. I'm really happy when I have women on my podcast because my podcast is AI and business. And mostly, most of the time, it's guys building and like guys watching as well.

2:05I think we're 70 % male. But it makes me really happy to interview one of the youngest self-made. Are you the youngest self-made billionaire? I think so. I hate the title, but yeah, I think so. This is very, very impressive. And you're an immigrant. I would love to talk to you about the future. Let's do it. Where you're building at CalSheet, you're basically making a ton of predictions about different markets. And I want to talk to you about what you were saying. Is there something that you see at CalSheet that we're not talking enough about? One example that actually my co-founder loves giving that is the C-Trini scenario.

2:42I don't know how to say this, C-Trini or C-Trini scenario, which is kind of like a little bit of a doomsday scenario for AI. And I think there's five conditions on unemployment levels and all of that. And actually, the odds are around, I think, 26 % or 30%, which is extremely high, if you think about it. For the doomsday scenario. So there are five conditions, and I don't know them all by heart. But the market is if three out of five of them hit, the market will pay out to yes. And the odds are a lot higher than what people think. And it's very liquid. The market's traded millions of dollars. And that's a market we look at a lot because obviously impacts our life so much.

3:18I think it was a very big report that came out a couple of months ago that just got so much attention. So we have a lot of markets on the AI side, obviously, on sports. I mean, we're in New York, so the Knicks, I think it's 37 % chance they're going to win the finals. There's a lot of very interesting markets. And I think a lot of our job is figuring out what are the big questions out there in the world that people want to know, forecast for, and what they want to know, you know, have data for and try to frame the right market that gets to that question. Because not every question is a very simple yes, no.

3:49You have to actually figure out what do people mean by saying AI did this or, you know, the economy is in this position and really define it. But a lot of our job is doing that. So it's very fun. Some of the requests, I can actually see them in my app. They're already public. But there are a lot of requests that you are seeing privately, right, of what people are asking for. And then you decide what goes on the platform. Is there a trend in anything related to AI that you're seeing? A year or two years ago, most of the markets proposed were about AI capabilities. People were interested in like, will AI be able to do this?

4:20Will AI be able to do that? Which model will be better than which model? Gemini Quad is still trending. That was kind of a big thing. Nowadays, actually, a lot more of the requests that we get are more on the impact of AI. So like tech layoffs and like just unemployment in general and kind of like how how that side would pan out. And I think it's like it's interesting because we see a lot of what people request of markets kind of show a shift also. And like I think there was a lot of excitement for AI at the start. It wasn't like mainstream that everyone knew what AI was. Nowadays they do. And I think you can see that kind of like vibe shifting to a more conservative, more skeptical vibe.

4:58And we see that in the market requests that we get. You have it monthly where you ask about tech layoffs. And for May, will AI be the number one reason for job cuts in May? And that's, well, it's 30 ,000 volume. It's one of the smaller markets, but still, like, we've done a lot of research. We have an arm of the company called Couchy Research that looks at the markets. And even with, like, I think 5 ,000 in volume, you already see kind of convergence to like a very calibrated number. So that number can be trusted for sure. 73 % chance that AI will be number one reason for job cuts. And that was the truth for April and March.

5:34So it looks like... It looks likely that it would be again. Yeah. And what you mentioned is very interesting. A year ago, people were still trying to figure out what AI is. And now with all the headlines, they're like, oh, okay, interesting. And now it's actually having some impact on my job, not for everyone, but for a lot of tech workers. Is there anything else you see in terms of like how AI impacts the day-to-day decisions. Are more people asking about a stable job or a business? I don't know. What kind of bets can you make? Yeah, on the AI front, I think that I would still quote, like divide the world of the AI markets between the impact that they have in jobs and, you know, government, even in elections.

6:13I think there's a lot of people asking, how can we define a market of the AI impact on electoral thinking around AI there? And the other side is just really like capabilities and all of that. But we have a lot of markets and for other things as well. For example, like big, you know, math problems being solved. There are a lot of things that we're doing more on the kind of like FDA drug approval trials and timings for those. Those markets we're getting a lot of interest in now. It's interesting because if you look at the history of prediction markets, right, a lot of the most important things that prediction markets do is try to price these kind of unknown innovation and tech things that we look at, like future of AI or the future of, you know, a lot of different drugs or the future of crypto or quantum computing.

6:55So we really try to have as many markets as we can for those. And now that we give interest on positions and dollar that you have in the account, you can actually, it makes sense for you to invest in something that's like five years down the road because you actually get paid on that, the interest. So we see more activity on those. Those are some of our favorite markets. This episode is brought to you by Google Chrome. You think you know a browser, but Gemini and Chrome, that's new. It can help you with practically anything on the web, like restoring a vintage motorcycle from a 50-page restoration block, or finally break down that long article you've had open for weeks.

7:30Gemini and Chrome is here for it. Ready to make anything online make sense? There's no place like Chrome. Check responses set up required, compatibility and availability varies 18+. When you need to build up your team to handle the growing chaos at work, use Indeed Sponsored Jobs. It gives your job post the boost it needs to be seen and helps reach people with the right skills, certifications and more. Spend less time searching and more time actually interviewing candidates who check all your boxes. Listeners of this show will get a$75 sponsored job credit at Indeed.com slash podcast. That's Indeed.com slash podcast.

8:05Terms and conditions apply. Need a hiring hero? This is a job for Indeed sponsored jobs. I was listening to some of your podcasts that some people are hedging their risks with AI. The example that I heard was floods. But now that I'm thinking, if you're fearing that AI is going to take your job, and it takes your job, you can bet against that on CalShe. So you can have some insurance payment. Exactly. We actually just had yesterday, not on the AI side, but on the sports side, a bar, I think in the Upper East Side here in New York, that was going to run a promotion that basically was, whoever comes in, we're going to pay for the entire tab if the Knicks win.

8:41And they were very concerned because they were like, we might be down like$10 ,000,$20 ,000. So then they bought a hedge that way. And I think that one of the kind of like the prediction market adoption curve, I think a lot of what we're going to see, that's my forecast there, is like at the beginning, everyone was also like not sure what was going on, what are prediction markets, all of that. Then there was a lot of skepticism. And now that people are starting to really understand what they are, you're going to see them starting to understand the other use cases like hedging and all of that, that we really see growing on small business side, but also beginning of hurricane season now in Florida, the amount of people coming and saying, like, can we have a hurricane market for this specific part of Florida I live in?

9:18Because I want to, like, you know, be able to hedge my deductibles or this or that. Because insurance wouldn't work if something happened. Exactly. For a person like me, I'm not into betting. I don't have time for that. I know some people do it professionally. What do you think is the use case for me as a user of CalShe? 70 % of our users actually don't trade on anything. They're just coming to ingest, like to just look at almost like the news. They're just coming to see what is the forecast of different things. So basically what you just did to look at their 70 % chance that AI would be the main reason for job cuts in May.

9:50They're going to come and kind of digest all that information in the morning from sports to culture to, you know, who's going to win Love Island and all of that. And that's the vast majority of the use case. Obviously, like, look, we make money on transaction fees. So we make money when people come in and trade. But at the end of the day, what prediction markets are really good at and how we get to impact the like billions of people really is with the data that we're bringing. And I think that that's kind of the best use case. That and also like, obviously, if you want a forecast for something, if you want the data for something that we don't have the market for, you can suggest, we can add it.

10:21And then you can kind of like get answers on the spot as well. But I would say that that's the, almost the main use case for people. So what are you looking at every morning? I look a lot on the economy stuff and I look a lot of the election stuff. I'm very, I love, you know, American politics. I love politics in general. I'm from Brazil. So I like Brazilian politics too. And in an election year, we've been looking a lot at that, especially we launched on. And that's something that, for example, driven by the use case of the forecasting and kind of getting information. Right. We have thousands and thousands of election markets for the midterms, all the primaries, all the House raises, Senate raises, all those things.

10:56But it's actually pretty complicated to digest all of this into like one number of like, how is the country leaning? right because you can look at the senate and it's like the senate is moving this way but it's always like there's one seed here how do you think about that the house is another way what we really wanted to create was a number that you can look at that will kind of track well like an ai assistant now that i'm thinking if i just ask what's the sentiment about ai today right and it runs all the exactly exactly and that's a lot of what we're working on now which are these like indices of how do we aggregate a lot of data about the world but also all of our market forecasts and try to create kind of like one number that is the sentiment or the index for something so we released the kawshi power american power index uh which is we call k-pow which is basically tracking is the country more republican more democrat based on current state of the world and our forecast what does it say last i checked was like point two like plus two for the republicans yesterday and we want to do more and more of that because i think it really helps and adds on the on the forecasting side.

11:59And we want to build more and more on the kind of new side. So now, before I used to look at race by race, there's some key Senate races that you look at like Maine, you can look a lot of the California races are very interesting, but now you can look at one number that do it. So like now the past couple of days, I just opened, you know, couch.com research. It's on our research tab and then, and then see the number there. But I try to look at, I'm looking at the markets the whole day. That's kind of my job. So it's fascinating. It's another way you consume news, but it's not from a particular news outlet.

12:27It's basically what people are trading on and betting on. Can we make some predictions? Let's do it. AI, what actually changes for a normal person by the end of this year? Work is one of the angles that's going to change the most. But I still, I don't think, when I look at like Kaoshi, for example, I don't think there's any role that we've completely just switched. We don't need this role, we have AI. All the roles have been like augmented by AI. So like an engineer now has like 20 cloud agents and all those things. And I think we will see more and more changes in other things. So for example, we are investing a lot in having kind of this couchy AI agent that's kind of like everyone has their agent.

13:02But anyway, we're investing a lot in like how to solve a lot of classic company problems with AI, like as the company grows, communication and contacts are a big deal, right? Like someone that just joined doesn't have the contacts to make decisions. They don't really know what they have to do. So how can we use AI to solve that? And I still think that by the end of the year, we're going to see work change the most. For example, for me, one thing that changed the most also with AI is like travel planning. I was traveling for a weekend and before I used to have to be like oh where should I stay whatever now I'm just like plan this whole thing for me for two days what are you using for that um I just use JGPT for that so that maybe so you just give it whatever you're thinking of and it gives you suggestions yeah but then you still go and book yourself I still go and book myself maybe maybe that's yeah well you know I did that yesterday and I was talking to my husband I'm like how is it possible 2020 since I'm still booking every hotel myself I'm clicking all the buttons right Exactly.

13:53And I think that it's like a lot of these are menial tasks that people don't actually like doing that. I think that that would be. But I still think that the biggest impact would be work. And I think that the concerns that people have with the impact of in their work is valid. I just feel like I'm more of an optimist than an optimist. What about money in general? Do you think 2026 is going to feel lighter or heavier for majority of people? That is a tricky question. I think it depends a lot on the direction of the war, I would be honest, because I think that like most people would think about gas prices as kind of like a big dependent on that.

14:24But I like how you think about that. So if somebody has a concern about money, they can go to call sheets and see what people are betting on. And we have markets on all these things like recession and inflation and all of that. I would probably say it's neutral. That would be my forecast if you want to go with that. But well, with the summer travel, I already feel like I'm, I don't know, 30 % poorer because of the ticket prices. That's crazy. That's fair, right? Because it's crazy how many things are impacted by gas prices or oil prices at the end of the day. It's kind of, we even forget about that.

14:55Absolutely. Do you see any jobs suddenly becoming safer? I was at the gym the other day and I was actually thinking that, for example, trainers are going to continue. I think everything that's more physical in nature are going to continue. we'll see how all the robots kind of like the optimists and all those things develop. But I think it's like for now, I would actually claim that the engineering roles and all of those roles that people thought were safer before. I think it's kind of clear that they're going to be less safe. So yeah, I would say anything that's more like craft and physical, probably.

15:25When it comes to trusting those numbers that you see on CalSheet, how many of the bets? So for example, if people are betting for something like 70 % agree, and then the reality is completely different and it's flipped, how often do you see that happen? Like how much should I be relying on public's opinion versus reality? I think the most important thing to think about is that Kausha, what we give off are probabilities, right? There's not an answer. So even if it's 99, it's still like, if you think about probability as like the frequency is like, you still have one in 100 that it's not going to happen, right?

15:57So for example, the Pope, the American Pope, He was around 1 % at Kaoshi the whole time. And the news were all like, oh, the Kaoshi markets were wrong. The Kaoshi markets were wrong. And I mean, one is not zero, right? You still have 1 % chance of something happening. And I think that that's kind of like the best example of a completely closed information system, which is a conclave. And there's no information that gets out how hard it is to forecast from the outside. But we've done a lot of analysis and research on our calibration. So basically, like if a market says a 70 % chance, is it actually 70 % chance?

16:26So we can actually plot, like do some calibration math. And the calibration is actually very, very, very good. And I think that even like there's a Fed paper that came out about prediction markets and how it's much better than any other forecast. But I think the core of it is understanding that 70 % is not 100%. Is there a number where predictions are right? Like an average percentage? That really depends on the time to expiration and the type of market. So, for example, for an entertainment market is actually different than from a politics market. And even in a politics market, if you see like one week before, I think you need like maybe one or three thousand dollars for it to be extremely accurate if it's one week before.

17:06But if it's six months before an election, then I think you need a lot like more on the like tens of thousands, maybe ten thousand. I'm not exactly sure on the numbers there, but I think it depends on a lot of things. I still think, though, that it's like the whole point of a prediction market is that people are putting money where their mouth is. It's a system that's from the start designed to incentivize truth in information, like good information, because people are incentivized to do their research because if they're right, they make money. Because if they're putting their money, that means they put some thinking behind.

17:35Exactly. So that's kind of how we really see it as kind of like directionally from the start is a better system. It doesn't mean that from the start you're going to have like if there's one dollar traded, you're not going to get a better answer than an alternative. But we've actually way less than what people expect to start getting there. So we touched upon some agents, and I really like that topic. Can you talk to me about how agents have transformed your life as the founder? I think it transformed a lot of every single part of the company. As of yesterday, we won 70 people at the company, and I think that we're able to do everything that we do a lot because we kind of just build AI systems from the bottoms up of how we were thinking about engineering, how we're thinking about market operations, how we're thinking about all of those things.

18:19And I think what it's helped me the most is like able to get context on things a lot faster and I'm able to know what's going on a lot faster. So I'm able to manage a lot more threads and a lot more people in a way more effective way. We are very like metrics driven in the company kind of everywhere. Obviously, for example, a great example is market operations, right? Like the way that we think about market operations is almost the same way as you think about a factory. We think about, you know, number of mistakes, but also like listing latency, determination, latency, coverage, all of those things that we kind of you think about it in a factory.

18:50and kind of how to define these metrics, how to get these metrics in real time and all of that is kind of all built on top of AI because it's very complicated to think about a lot of these things in the context of like markets. So yeah, I think it's like on the metric side and how like communication flows and is aggregated in the company. It's kind of all like that. And it becomes a lot simpler for me to do my job because I can just have my cloud agents kind of like do everything I need them to do. Can you talk to me about a couple of agents that you build for yourself? Something that anyone who's a knowledge worker could deploy for themselves as well.

19:20Well, one thing that I think is useful for a lot more people maybe is on kind of like weekly planning and kind of like state of things that I think it's like, how do we get updates from the entire company, track from what the updates were from the week before, flag, what hasn't been done. What hasn't been done. How do you collect all the data? Do you use any tools? Does every employee have their agent? How do you collect it all inside one database? Yeah, that is a great question. And I think that we should ask our engineers who know better because I'm very lucky that they can build a lot of the things for me.

19:52In terms of that, like it's connected to everything that we do. Emails, docs, Slack, everything. All of that. We actually have an AI team now that is actually building. We obviously have a very, very good like engineering side of the AI equation is very good. But we're trying to build kind of like every new employee should get an agent. That's kind of like the biggest problem we have there that we're trying to figure out is how to figure out like permissions in the right way. we need to make sure that for example we have a lot of legal work or like surveillance and all of that that it has to be very you know just some people have it and how do we think about it uh that way but i would say that like planning organizing and collecting information sundays are very like heavy days for me because it's like when i stop and i look at the entire week everyone what they was done what we need to do the next week look at all the metrics and all that's all i do on sunday and now i'm actually able to like have brunch on sunday because i'm like i i have a lot more time to think about things, but because a lot of it is kind of done in the way that I expect.

20:45But I would say is like, like really looking at like, for the past X number of weeks, this person has over-promised, under-delivered, these are the things, like these metrics are not. I'm trying to build something for my, like that for myself, but what I realized, we need to hire someone. So we try to build internally and my team is like creative producers and now we hired someone with an engineering background to do that. Yeah. And that's the thing is also, it's like, we were kind of putting engineers in every single part of the company to kind of figured this problem out because obviously market operation is a great one but for example design um design is something we didn't use a lot of ai for and now we're kind of cloud design or cloud design yes but we're trying to also figure out a lot better on like how do we also empower almost everyone to be a designer in a better way uh obviously we have a design system and all of those things but if an engineer just wants to ship something like how do we actually build something that it's not we're still defining it but it's not just like you can right now you get a design system you kind of can ship an experiment very quickly but how do we actually like do it in a great way from the start because i feel like that's the point of design right you can you can just like built-in reviews or something yeah but also like engineers are very good like if you want to just test a new module on a page right you can very easily put it out but what we have at the companies they put it out and then we get like we test and we're like okay this was good direction to go let's ship it and then when we ship it we actually go back to design and then the design team actually makes it good because before i was just like it didn't look awful, but it wasn't great.

22:09And we're trying to figure out how can we actually not need that loop anymore by making, yeah. A lot of things we're thinking about. Interesting. That's interesting. So you have that agent running, giving you all the information, something that I'm trying to build, I can relate to a lot because also like information is all over the place and you need to collect it. And you want to make sure the agent knows what's a priority, what's not, because otherwise it's a very long email of all the things you need to do. One thing we struggle a lot with is that on Slack, we have these two, like this, basically this channel is like product feedback, right?

22:40That we put a tweet that user is complaining or a user from discord or internal, like everyone puts stuff there. And it's very, very tough to prioritize, track what's done or not done. And we have all those like linear integrations and all those things. But what we run, it's almost like a, we were talking about this yesterday. It's like a bug, something to, for that extent to like, we'll track everything immediately, prioritize them, see what's live and not live without adding engineering burden that I think would be very helpful. Obviously, like on the QA side as well, we're trying to figure something out so that it's a lot better.

23:11There are a lot of companies we tested on the QA side. Nothing was great. So we're trying to figure it out. Anyway, we're trying to put a lot of time in. Burning a lot of tokens from what I'm hearing. Is there anything that you've built yourself, for yourself? Not really, to be honest. I think that both Tarek and I, we're kind of very lucky to have a great team that does a lot of these things for me. Because a lot of like what we think about is like, obviously Tariq and I, we should be always trying to be as productive as we can and be more and more productive. But I think it's more about, for us, our biggest question is how do we build the most efficient company that we can that will keep, I think the big differentiator for CalShade, that people used to think it was a regulatory piece.

23:51I really think the actual real differentiator is how fast we've moved and like how good our product has been by how fast we're moving. And for us, that's the biggest question. A lot of people are talking how a founder can be a solo founder now because he can deploy, he or she can deploy so many agents. What I'm hearing from you is completely different. You still, you're hiring more engineers to build those things for you. And this is what I'm experiencing myself. Yes, we tried to build something and something's working, but if you want to do really something sophisticated that doesn't make mistakes, well, or at least the mistake rate is like 3%, then you have to hire someone.