Why Nice People Struggle with Money, with Dr. Sandra Matz, Professor at Columbia Business School

24:19And we have a limited amount of time. We have a limited amount of cognitive resources. So whatever is the default, whatever is the baseline, is what most people are going to stick with. And there's fantastic research in different domains. The one that resonates with people and that sticks is organ donations. So we know that there's countries, for example, where you automatically opted in to an organ donation registry. So you opted in. Nobody's forcing you. You can always opt out, but you're kind of enrolled automatically. And then there's other countries like the US, for example, where you have to go and it's hard.

24:51You have to go to the DMV. You have to actively say you want to be part of that. And what's interesting is that most people across the board, across different countries, agree that being part of an organ donation registry is actually the right thing to do. It's a good thing to do. So intentions and motivation is there. Now, if you had an educated guess of like, what's the percentage of people enrolled in these different countries? You can imagine there's a huge gap. So for the countries where you actively have to enroll to be an organ donor, it hovers around 10 % of the population that actually turns their intention into action and signs up.

25:28For the countries where you have to opt out, that number is hovering around 97, 98 % of the population. So almost everybody is enrolled. And the only difference there is what was the default? You were enrolled by default or you were not enrolled by default. And you can translate the same principle to the world of data. So in the world of data, again, most people would probably say, yeah, in a world where I could have it all, I would probably want my data to be protected somewhat. somewhat. But now the current default is, well, you have to opt out of all of the tracking that is happening. And nobody has the time to do that, right?

26:06Nobody has the time to really go through all of the kind of specific terms and conditions, understand the legalese and then actively manage. So there is a world where kind of coming back to this idea of paternalism, we just try to change the default such that if you don't do anything, your data is protected. And now companies actually have to convince you that by using your data, they're able to offer much better services and much better products. Right now, it's lip service most of the time, right? So it's like this comment, like, well, we need your data to make the product better. But there's no way for you to actually know if that's true or not, because most of the time you can't even use the product without consenting to all of your data being sampled and being extracted.

26:45The idea that changing the default benefits consumers directly because it protects their privacy, but it also in a way should lead to more innovation and better products. Because now companies will only get the data if they actually live up to their promises and make their products better. Otherwise, you're not going to say yes to granting them access. Right, so create an incentive for why I should opt in. We see the same opt in and opt out with 401k enrollment. When previously 401k enrollment was opt in and far fewer people were enrolled in 401ks. And then there was a big cultural shift. And now in many major workplaces, The default is opt-in and you have to actively opt out.

27:26And that has dramatically increased 401k participation. And it's such a great example for helping you accomplish what you want to do anyway. I think most people would say, no, this actually makes sense. This is something that is helping me. I just didn't have the bandwidth with all the push and pulls of my daily life, whether that's kids running around at home or whether that's stress at work. That was just not top of mind and I never made the time to do it. And there's a really nice analogy that a friend of mine mentioned at some point that I like in this context. So like if you want to change behavior, it's like launching a rocket to space.

27:59One thing that you need is thrust. So essentially that's the motivation part. So you need to be motivated. You need to have an incentive to change. Like in the 401k, it's like most people would say, yeah, it's a good thing. But you also need to reduce friction. And I think that's where the defaults are coming in. So even with the strongest thrust and the strongest motivation, if the process is really, really difficult, you're still never going to accomplish that. And the same is true in the data world. Even when most people say, no, I would love to protect my data somehow, the process is currently set up in a way that makes it so, so hard that you never get that space off the ground because there's so much friction in the process.

28:38So let's go back then to the previous question, the one that you asked a few minutes ago, which is how do we manage this as individuals? How do we manage all of this? I think this is actually now building on your previous question. If we assume that we can't do it alone, if we assume that transparency and control are somewhat necessary, but they're not sufficient to make sure that you manage your data properly. For me, the one thing that I've been thinking a lot about is how do we actually get support people with this competent crew and community that in a way has similar interests in how they use data, but they're also now a collective that can make much better decisions and maybe hire experts.

29:22To give you just one example, expecting moms. When you're expecting a baby and you're pregnant, it's a terrifying time because there's so many unknowns and you kind of see your doctor maybe every two weeks, every four weeks, and you get like a quick thumbs up, thumbs down, like the baby's doing well or not based on a quick ultrasound. But what you really want is almost like a day-to-day based on my genetics, based on my medical history, based on my lifestyle, based on everything that you know about me, right? Take biometrics, take whatever you want. I want advice that's personalized to here's what you should be doing in terms of nutrition, in terms of exercise, whatever it is, and to make sure that you are healthy and your baby is healthy.

30:04Maybe track data along the way to see if you're on track or not. Now, to get to this point, we need to pull data, right? Even if I get access to all of my genetics and my medical history, it doesn't tell me anything because I can't map it against like healthy pregnancies or unhealthy pregnancies. So if we wanted to get these insights, what we would need is to have a large pool of data where women come together and share their data. Now, I don't want a pharma company to have all of their data. I don't want any central entity, like a third party have access to their data because again, it's like super intimate.

30:38If you think about genetic data, if you think about medical histories, so it really needs to be someone that you trust. And there's this idea of data trusts or data co-ops, which are member-owned entities that actually come together and they say, well, we're kind of all expecting moms and I want to be part of that member-owned entity, which means I actually have a say over how that company is run. And similar in how we do this in the financial world, so banks have these fiduciary responsibilities to their customers. The co-op now has fiduciary responsibilities, so they're legally obligated to act in the best interest of their members.

31:14and now because you have all of these moms coming together, first of all, the insights that you can generate are far superior than anything you could do by yourself because now you can, again, see here's how this plays out for maybe different social demographics, different lifestyles, different genetic histories, but you can also benefit immediately, right? So it's not that you hand the data to a company and maybe you can then purchase their product in 10 years time to see what you should be doing or should have done 10 years ago. So you immediately benefit from these data. And for me, it's this idea that, yeah, I by myself, first of all, my data itself is not as valuable, but I also don't know exactly how to manage it.

31:51I don't know how I should be storing and securing my genetic data in a way that doesn't leak or isn't the victim of a security breach. But if we come together, now we actually have the resources to say, let's hire experts. Let's hire experts whose sole responsibility is to try and figure out how do we both protect your data, but also maximize the value that you can get from that. And for me, that's a totally different way of thinking about data, right? Because it's not saying I'm just going to put you in charge. It's like, no, there's so many people who have the same incentives, the same ideas of what could be done with their data, and we just need to get them together so that they can benefit from the strength of the collective, whether that's, again, just the insights or the fact that you can now have someone manage and read through all of the terms and conditions and think through the legalese on your behalf.

32:38It's quite onerous to put that responsibility on any one given individual. It is nice to have the right to do that, but also with rights comes responsibilities. And so it is also, both are simultaneously true, nice to have the right, burdensome to have the responsibility. But if you have support in doing that, then it really becomes a right. So you now have someone who's been trained to do that, who has the time, that's their full-time job, and they're doing it with your best interest in mind. So they kind of figure out what is the collective one. And there's already examples of these data coops.

33:09So my favorite one is in Switzerland. It's called MyData. And they're in the medical space. So they're thinking about these rare diseases that we still don't understand really well. So multiple sclerosis is one of those diseases that it's just so complicated because it's determined by your genetics, by the environment, by your medical history. and what they do is they say, well, the way it currently works is if you suffer from MS, the best outcome you can hope for is that a pharma company takes your data, does some R &D, kind of tries to understand the disease better. And now in the best case scenario, maybe in 15 years time, you can buy the drugs that they develop based on your data for millions of dollars, right?

33:51Worst case, you never benefit from that data being out there at all. And what my data does is, again, similar to the expecting mom example, it says, let's get people together who suffer from MS. Also, let's get people who don't so that we can have a comparison. And we, as an entity that's, again, member-owned, we analyze the data. So we try to extract the most insights. And it's not just that we try to understand the disease at the meta level. We also understand you specifically. So we can get insights and send them to your doctor and say, here's something that we learned based on the tracking of symptoms, maybe mapping against what the symptoms of other people look like.

34:29And we make recommendations to the doctor of like, here's how they should be treating you. And now you have like almost this full feedback loop where the doctor can then go back to the data corp and say, hey, here's something that worked for that patient. Here's something that didn't work. And it's a completely different setup. So the patient benefits immediately without having to spend those millions on the drugs. If anything, oftentimes these data co-ops allow you to monetize some of the data, right? So if there is a pharma company that the data co-op trusts and they think could be helpful in understanding and developing better treatments, well, then they can have deals with their company.

35:05Because now you have suddenly 300 patients, 500 patients, and your bargaining power with that pharma company for the data grows exponentially, right? Because the data of 500 people is just so much more valuable than the data of one individual. What I'm hearing from you is how people collectively can come together to use data in a way that better advantages groups of individuals. But how can any one given individual who's listening to this right now, who wants to save more money and you're extroverted, you're agreeable, you're male, you're between the ages of 40 to 50. So you occupy all of these different cohorts.

35:44You work in manufacturing, you live in Michigan, you have all of these disparate cohorts that you belong to, and all of them have some certain attributes, some of which, broadly speaking, tend to often be overlapping and some of which are not, you know. Maybe you're also super into ballet, in addition to everything I've just named, right? So given the mosaic of who a person is? How can any individual who's listening to this figure out a way to make it all work for them? I mean, that's really where these large language models, the likes of Chet, GPT, Cloud3, Gemini, really shine. And they've entirely transformed this space because it used to be that the psychological insights required, first of all, a lot of data to build these models.

36:36Then they needed someone to create a model, right? It says, I can predict something about your psychology based on the data that I find about you. Large language models make these insights accessible for anyone, right? So you can ask OpenAI or ChatGPT and say, look, here's what I want you to know about me. And in a way that also gives you control because you control what data you want to use as an input, right? You can say, here's my social demographics, here's my preferences, here's something else that you should know. I'm a little bit neurotic. So in your advice, please take that into account and factor it in somewhat heavily.

37:09Or I really care about my dependent. So make sure that as you're giving me financial advice or you try to help me change my life in whatever way you're hoping to, and make sure that you keep my loved ones in mind. And you can just ask it and say, well, what would you recommend? And what's remarkable about these large language models is that they're incredibly good at simulating different personas, right? They've read the entire internet. They've essentially digested all of the human narratives that we've created over the years in terms of stories, in terms of news articles, in terms of like blogs and really anything that's out there.

37:45And they're really good at putting themselves in the shoes of someone of specific traits. I, for example, use it all the time when I travel. It used to be so time consuming to say, okay, I'm going to go to Barcelona and I have certain preferences. I want to kind of live in a hip neighborhood. It has like lots of good coffee shops. I also want a lot of playgrounds so I can take my son there. And it was really hard to find. Now I can just tell OpenAI and say, okay, here's all of the factors. Think of me as like mid-30s, having a kid, husband is traveling with me. I'm into food. I like these hip neighborhoods, but make sure that it's safe for the kid.

38:22And it comes up with remarkable recommendations. It's really just playing around with these language models that can just pretend to be a certain type of person and think through the world in this way, it's just a remarkable opportunity that we now have at our fingertips that we'd never had before. And I think of it, the analogy that I always use is, it's a really, really competent intern with unlimited time, resources, and knowledge. You need to give them a good task. So you need to be very specific of what are you asking for. So whatever context you give it, whatever the prompt, whatever information You give it about yourself.

38:59That's what it's going to run with. And you also probably want to check the output. Most of the time it's doing a great job. But once in a while, it's just like off by a little bit. Same way with an intern. You probably want to supervise the work. But then it's like just the most efficient intern that you can imagine because it's seen so much. And it's just really good at simulating these different personas. You know, earlier this year, I went to Panama. I had Claude act as essentially my travel agent, travel planner. You know, I said, here's what I'm interested in. Here's what I'm not interested in.

39:30Here are some of my parameters. Here are some of the responsibilities that I'm going to have to tend to while I'm traveling that are going to be kind of limitations on what I can do on a day-to-day level. Based on all of this hyper-personalized information about me, can you design an itinerary? It gave me an output. And I said, all right, what I dislike about this output is X and Y and Z. And then we went through five or six iterations. and by the final iteration, I was like, all right, we've nailed it. And then once the trip began, of course, things came up. I had a pet that was sick back home.

40:06And so I spent a day FaceTiming with the veterinarian. So then I just went back to Claude and said, hey, you know what? Our original Tuesday plans got totally waylaid because I spent all of Tuesday just FaceTiming with a variety of vets. How do we remake this? And so then we responded in real time once again. And it probably provided emotional support as well. Claude is really good at that. But you're totally right. What is so remarkable about these large language models, as opposed to traditional advertising and persuasion, is that it's a conversation. So it's far closer to this, like you have a good friend that you can also push back on and say, well, here's something that I don't like.

40:44Just kind of try again. And it's almost like you're co-creating the perfect outcome as opposed to me just giving it to you. And the one thing that I've been thinking a lot about in this context is on some level, in most cases, it's just incredibly useful, right? Because it's like it's very good in giving this advice. The one thing that I'm slightly concerned about is essentially if it's going to make us far less complex and interesting over time. So the idea that if we just outsource all of our decisions to these large language models, what they're kind of trained to do is predict the most likely outcome, right?

41:24Even the way that like a large language model operates at a very basic level is it predicts the next word in a sentence. What's the most likely next word in a sentence? So if you ask it, well, what color should I paint my wall? It would probably say something like beige or white or something that's very common. It's probably not going to say turquoise and pink and stripes with blue dots because nobody's, it's never seen that answer in all of the internet, right? And it's very unlikely that most people are going to like that. So even when you train a model based on your preferences, you can imagine that it's going to go for the most likely preference that you have.

42:03It's in a way, if you think of yourself as a distribution of preferences, most of them were somewhere in the middle, but then you sometimes have like these obscure preferences on the edges that come from serendipity and they come from discovery and like these fun moments where you suddenly explore something new that you've never seen before and you realize, no, actually, I like that too. if we're not careful and we're just overly outsourcing all of our decisions to these large language models, we kind of lose out on the discovery part that requires still a lot of human intentionality and saying like, for one afternoon, I want to do something that's completely different.

42:39Maybe even ask it, right? And say, hey, what would you recommend to someone who's totally different? Maybe it's like a different age, doesn't have kids. What would that world look like? Because Otherwise, I think we're just losing this ability to explore, which is really how humans and the human species learns. We kind of learn by exploring new things. And sometimes it's a mess, right? Sometimes we try a new restaurant and it's like a total failure and we would have been better going to the one that we already know. But sometimes it's also the biggest hit. And sometimes we find something new that we didn't know was coming and we like even more than our typical go-to restaurant.

43:15I like the idea of prompting experimentation. asking Claude, what would you recommend to somebody who had the opposite characteristics of what I just named? You can prompt a large language model to give you something else, almost as an opportunity to say, well, I don't know what the reality of someone who looks completely different to me, has a completely different experience in their life, has grown up in a totally different way, lives in a completely different part of the world. I don't know what their experience is like and what they would want to do when they're in Barcelona or which movie they would want to watch on a Saturday night.

43:49But the large language model does, right? Because it knows all of the personas. So it's like this almost unique opportunity that we have with technology to say, hey, help me sample more of the world. Help me understand what the world looks like from different perspectives. But you have to be intentional about it because that's not the way that it typically works. The echo chamber swap. That's an interesting concept. You know, you also wrote about that in terms of give me the news that this other person, this other very different profile would be getting on this day. Right. And to just have a window into what are the headlines that they're seeing?

44:26What are the sources that they're reading or listening to or watching? That echo chamber swap can sometimes pull you out of your bubble. I've talked to many friends who are like, I didn't realize how much of an echo chamber I was living in until I just grabbed somebody else's device, you know, and like looked at it and was like, whoa. And it's so hard because right now you almost have no way of breaking out of that echo chamber. There's no way that you can have easily like a different Facebook news feed or like your Google page one that's customized for someone else. But there's all of these opportunities to actually do that.

45:00And the really interesting part is, could you even take it a step further and say, instead of just showing you the news that other people see, could I also help you relate to those news to some extent by making them more personal? Because you can imagine, actually, the Wall Street Journal had this blue feed, red feed website at some point where for the same news items, they just said, like, here's what this looks like for Democrats. Here's what it looks like for Republicans. seconds, I think the potential danger that you have there is you see what it looks like from the other side and you find it so appalling that you just dig your heels in even deeper, right?

45:34So there's always a risk of reactants and saying like, this is so far removed from anything that I believe in that I don't even want to look at it. And if anything, you're kind of even more alien to me now than you were before. But I think if we thought about this echo chamber swap really as a personalized experience of saying like, look, here's the entire experience of that person. Here's not just the news that they see, but here's what their day looks like. Here's why they think about immigration in a different way, because they've grown up the way that they've been educated, the way that their family is set up, whatever it is, probably makes it much more likely and that they think about it in that way.

46:11And for me, that would be really like a perspective exchange. It's not just taking a glimpse into what your reality looks like, but kind of immersing yourself a little bit more in a way that makes it more human and somewhat more relatable. That's a tough order though. You know, another variation of the echo chamber swap is when you go on Bloomberg, the headlines that you get if it's Bloomberg US versus Bloomberg Europe versus Bloomberg Asia. And the economic headlines that you get if you're looking at Bloomberg Asia or Bloomberg Europe is totally different than what you read when you're looking at Bloomberg US.

46:48It's just a way to get a much more global perspective on the world economy. And that is even still visible to you. Right. So you can still go to Bloomberg, Europe, Bloomberg, Asia and see what they're talking about. In most instances on the internet, it's not even visible because I can't easily see what you see. There's no way for me to publicly access that. So we all live in these kind of small realities that I don't even have an opportunity to hop into yours the same way that I could at least manage my news if I were to look at all of these different Bloomberg outlets.

47:26So imagine it's midnight, you're on the couch, you're scrolling through this new website, hitting the add to cart button, and you decide to check out, but you remember that your wallet or your credit card is in the other room and you don't want to get off the couch. That's okay. You don't need to because there's that big iconic purple shop button. You know the one. With that shop button, that's Shopify. And there's a reason that so many businesses sell with it because it makes everything easier from checkout to creating your own storefront. Shopify is the commerce platform behind 10 % of all e-commerce in the US, from household names like Mattel and Gymshark to brands that are just getting started.

48:04Shopify will give you a leg up from the beginning with beautiful ready-to-go templates to express your brand. And you don't handle everything in one place from inventory to payments to analytics. And they've got built-in marketing and email tools and that iconic purple ShopPay button used by millions of businesses around the world. If you want to see less parts being abandoned, it's time for you to head over to Shopify. Sign up for your one month,$1 per month trial period and start selling today at shopify.com slash paula. Go to shopify.com slash paula. Shopify.com slash paula. This Friday, the NBA on Prime tips off with their debut doubleheader.

48:48First up, Boston brings the action to the Garden as the Celtics face the New York Knicks. Then out west, it's a battle of the superpowers as the Lakers try to poster the relentless pace of the Minnesota Timberwolves. The Celtics and Knicks, who has the most hustle and defense? This game will come down to who wants it more. and the Timberwolves and the Lakers. It's two Western Conference heavyweights going toe-to-toe. Buckle up, folks. This one's going to be electric. Celtics-Knicks, T-Wolves and Lakers. Coverage starts Friday, 7 o 'clock Eastern, only on Prime. This episode is brought to you by Progressive Insurance.

49:31Do you ever think about switching insurance companies to see if you could save some cash? Progressive makes it easy. Just drop in some details about yourself and see if you're eligible to save money when you bundle your home and auto policies. The process only takes minutes, and it could mean hundreds more in your pocket. Visit Progressive.com after this episode to see if you could save. Progressive Casualty Insurance Company and Affiliates. Potential savings will vary. Not available in all states.

50:05You mentioned you teach negotiation at Columbia Business School. How has your background in big data fed into your practice and teaching of the field of negotiation? It's an interesting question because in most cases, it's actually relatively far removed. So I would think of the negotiations world, still the world is like the messy human world, right? Not necessarily including any of these algorithms, data and AI. At least that was the case when I started. So when I started eight years ago, it was very much focused on just like face-to-face, giving you the tools that you need to become better.

50:43Now, I would say over the last two years or so, that's changed a bit because the same way that we think about AI assistant in different parts, helping you pick the next vacation, helping you maybe come up with an exercise plan. We can also think about them in the context of negotiations. And sometimes that could just take the form of, can we build a tool that allows you to practice, right? So much of negotiation is, can you go through a couple of practice rounds where you interact with different counterparts? Like you play through different scenarios. And if this happens and I'm up against a counterpart who's just trying to eat my lunch and is trying to screw me over, how would that conversation unfold?

51:20How should I be acting? How should I be preparing? If I know that I'm interacting with a partner that is going to be a long-term relationship, how does that play out? So what we now can do is we can develop these tools where an AI not only plays the part of your counterpart, right? So it's not only that it's the person who negotiates on the other side, but we can also build these coaches that are based on AI and say, look, here's everything that we know from science, from behavioral science, from years of experience that are helpful strategies in how you get more at the table, right? Here's the questions that maybe you should be asking.

51:58Here's how you can make offers to the other side. Here's how ambitious you should be when making a first offer. All of these pieces of advice that we typically teach the students in class, we can teach an AI to act as a supervisor and as a coach. So as you go through your conversation with your AI counterpart, the coach can say, hey, well, actually just take a step back, take a pause here. Here's something that you did, which might have actually not been super beneficial to you. Want to try this again by changing the way that you approach this negotiation. So like AI in a way acts both as a sounding board, as a tool to practice, but it also on some level gives the students, me as the instructor on a very much personalized basis, right?

52:44Ideally, in my negotiations class, I would listen into all of the negotiations the students go through and say, hey, you know what, Tim, here's something that I think you could have done better based on the conversations that I've seen. Here's something that I would want you to try differently next time. Now, I can't do this for the 50 students in the classroom, but an AI can do it and it can do it pretty much as well as most of us instructors. Wow. It's interesting that you mentioned that. So I teach an online course on negotiation and we put a huge emphasis on peer-to-peer practice because sure, I can teach concepts and tactics, but really the practice, the peer-to-peer practice is where the true learning happens.

53:23The notion of getting AI involved, I guess that is the next inevitable iteration of this. Yeah, if you don't have a peer, it's certainly a good fallback option, right? For us, I would never replace the peer-to-peer because there's also something from peer-to-peer that you get that is still somewhat more difficult with AI. And those are the perceptions, right? One of the big, I think one of the valuable parts of the class is you go through negotiations with your classmates and you can stop after negotiation and say, first of all, let's turn over the cards, right? Let's see how much you got from the negotiation as opposed to your counterpart.

53:57part. How much of this offer did you get? You never get that in the real world, right? You kind of negotiate and maybe you thought you got a great deal, but actually the other side totally screwed you over. You're never going to find out. Or you felt like, man, this is like really tough and I was totally screwed over, but you actually really got a good deal. First of all, the students see that, but they also get feedback from their classmates and saying like, well, I actually think you are not assertive enough. I think you could have been much more assertive and I didn't even mind. Or the opposite is like, well, you came in way too strong here.

54:25I can to some and replicate that and kind of give you the same perceptions and the same impressions. But I still think there's something unique about hearing it from another person as opposed to an AI. Right. Yeah. There is a very innate need that we have at a very visceral level to be around other humans rather than robots. And I don't think that will ever go away. Human interaction can never be fully replaced by automated interaction. The way that I think about these AI assistants, it's really a compliment, right? You can imagine this in many different contexts, like negotiations is one. Think about it in the context of coaching.

55:02Think about it in the context of therapy. There's always the fear that, well, is AI going to replace human therapists? Well, probably not. for anyone who can afford a therapist in blood and flesh, they're probably still going to go through that therapist, but there's still so many other people who either can't afford a therapist at all, or who might actually benefit from also having an AI that they can talk to in between sessions, right? If you think about like a therapy session, you probably see a therapist maybe once a week, maybe twice a week. And there's so much that happens in between, and you can only retrospectively talk about it in the session.

55:37Whereas if you had an AI that says, well, I'm in In the middle of a fight with my sister, I just came out of it and it's fresh in my mind. Emotions are still kind of running high and I can capture that moment using an interaction with a chatbot. And now I can take that interaction to my therapist next week. Now that I'm cooled off again, we can discuss it in a more rational way. That would be hugely beneficial. And I think the same is true in context like negotiations. Yeah, you kind of practice with your counterparts, your human counterparts. But to make the most of that, you might actually have some practice in the middle so you can iron out some of the things that are difficult for you.

56:14Try different things and say, well, I really feel terrible being too assertive, but I want to give it a shot one time. Or I want to try and see how it feels to lie to my counterpart when they ask me a question. You probably don't want to do this with another human being. They're your classmates. They're going to find out once you turn over those cards. And even though it's just role plays, people take it very personally. But doing some of that with an AI and just seeing how does it feel, how do they react, I think can be hugely beneficial as a complement. Right, right. I could see also when you're using AI as a negotiating counterpart, if you could upload a PDF into your AI, into Claude or ChatGPT, that gives them the role of the other person.

57:01But if you yourself have not read that PDF, so you can create that information asymmetry, right, then it's possible that the AI can reflect that information asymmetry in that they have information that you don't, and then you can have that negotiation. I think where it becomes challenging is, even though it would be counterproductive, it would be so tempting to read that PDF, right? That's true in a classroom too. Yeah. So the one thing that we tell them in the beginning is if you read the materials of the other side, that negotiation is going to be useless for you. And the same is true if you have an assistant, like an AI assistant, right?

57:37If you read the materials of your AI assistant, it's not going to be useful for you in any way. Yeah, exactly. Preserving the information asymmetry from the two parties is fundamental to having good peer-to-peer practice. How do you manage it, though? So this is one thing that we do in our class on negotiation. Towards the end of the class, we start with single-party, single-issue. Then we move to single-party, multi-issue. So you've got two individuals who are negotiating multi-issue. But then we go to multi-party, multi-issue. And that's where coalitions come in. People form factions. Those are the fun ones.

58:14Yeah, exactly. They're the hardest to manage because you've got to collect like nine people into the same Zoom room at the same time. So logistically, it's kind of the hardest to schedule. But they're the most fun and I think most beneficial practice sessions that we do inside of our class. But how do you do that or do you do that with an AI or with data? Yeah, that's much harder to do. So you can still technically set it up that way because you have different AI agents interacting with you, interacting with one another. so you could technically still set it up. I think in terms of the benefits for students, it would probably be lower.

58:47I 100 % agree. So I think AI is oftentimes really amazing for like these teaching people the basic skills so that you can then spend more time in the classroom dealing with the big negotiations, right? So I can save time on the single issue negotiations that I go through and the kind of single party multi-issue negotiations and dive in more quickly into the more complex ones because I can just outsource some of the easy ones to AI practices outside of the classroom. Right, right. Well, this is a fun conversation. I didn't expect to be talking so much about negotiation, but it's one of my favorite topics.

59:23Me neither. And I think it's like this one skill that I would love everybody to learn more about. When I joined Columbia Business School, I did negotiate my contract. I absolutely hated negotiations. I always felt like I'm just going to ruin all of my relationships. I don't feel comfortable asking for more. Why would anyone want to go through this? And then two weeks later, I think I was told that I was going to teach negotiations. And I was like, you're going to be kidding me. I'm the worst person to teach negotiations. It was the best thing that happened to me because I learned so much. If you think about it from a psychological point of view, from like reading other people, figuring out interpersonal dynamics and so on, I think I've learned so much and I think about it completely differently now.

1:00:10I don't think of it as like this tug of war anymore, right? Where I'm just kind of trying to screw you and you're trying to screw me, but much more of like this creative problem solving. Like there's all of these puzzle pieces on the table and how do we make them work such that you get a good deal and I get a good deal. And for me, that's really changed entirely how I think about negotiations. And I wish there were more people learning the art and science of negotiation. Yeah. Actually, you know what? I do have a follow-up question. I'm curious what your thoughts are when it comes to that creative problem solving and puzzle solving when there's a power imbalance, when it's a negotiation between an employee and a boss, the employee wants to get a raise, the boss has certain budgetary constraints, but there's this power dynamic, right which makes it so inherently different from let's say selling a car on facebook marketplace where you you're lateral to one another no totally and we could now dive in into the topic i mean the one thing because you mentioned salary negotiations i think the one thing that we talk about with students a lot is just how do you put yourself in a position where you level out some of these dynamics just by you having alternatives right if your life depends on that job you probably have a harder time asking for more.

1:01:27If you know that even if your boss, worst case says, I'm not going to give you a raise and now I also don't like you and you're not going to get promoted, but you have something else lined up, even if you don't mention that in the negotiation itself, we know that just psychologically, it makes you a lot more confident. So you might be going in more confident and asking for more. So that is one thing. And the other one is, if you think about this, it's like a tug of war as opposed to creative problem solving and trying to figure out really kind of interests rather than positions. So like, why is it that you care about this stuff?

1:02:03What really matters to you as opposed to what are you asking for? In a way, if you're lower in power, that becomes even more important because in a way that's the only thing that you can do, right? You don't have the formal power to impose a solution on your boss because still your boss, but you can try and figure out, okay, how do I actually suggest something that gets me what I want, but also make sure that the boss is happy and gets what they want, right? Because they oftentimes have to save face and justify what they give to you, to other employees, to maybe their superiors. So how can you be creative in terms of finding a solution that still works for you, but also for your superior?

1:02:40So the creative problem solving, I think almost becomes more important if you don't hold a formal power. Right. How do you manage that when so much communication now happens online? You know, when you're having real time face to face synchronous communication, these creative problem solving conversations become a lot easier. But when this is done over email, it becomes a lot tougher, you know, because this asynchronous communication on a Slack thread or in email or text message is a very challenging way to hammer out any issue. and yet so much of the time the counterparty will insist on having asynchronous communication.

1:03:20I was going to say, if you are someone who thrives in face-to-face, try to insist, right? You can't be overly pushy, totally understood. And sometimes your counterpart insists on you doing it over email, but to the extent that you can, you might want to change that context and say, hey, I would really love to meet face-to-face, any chance that we could catch up, even if it's like half an hour, because I just want to see what this looks like from your side. and tell you a little bit more about why I'm thinking about it the way that I'm thinking. Again, if that counterpart completely walls and says like, no, we have to do it over email, then you're stuck with that.

1:03:52And I think there it's oftentimes still helpful to both signal flexibility, but also ask for the things that you want, right? So you want to be very clear, and this is true for face-to-face, but it's even more necessary for these formalized email conversations of here's the things that really matter to you. A lot of people are afraid of communicating that because they think, well, if I tell them what really matters to me, aren't they just going to exploit me? And you know, like in the grand scheme of things, you're probably going to be exploited once in a while if you follow that strategy. But more often than not, just by you signaling, hey, here's what's important to me.

1:04:29Now let me hear a little bit about what's important to you in terms of where do you see my career trajectory? What is it that I could actually contribute to the company that I'm not currently doing that I could be doing in my next role. So the more you make it about this, again, like two-sided, I'm going to tell you a little bit about what I care about and here's why. So kind of providing this rationale and saying, here's why I care about getting a title that reflects my level of seniority, because I think it's going to allow me to better interact with both internally, the people that I'm supervising, but also maybe external suppliers and so on.

1:05:01And so I think the more that you justify and engage in this, okay, but also what can I do for you? That I think is is always helpful, but you just need to make it very explicit when you communicate online. Well, thank you for spending this time with us. Where can people find you if they would like to learn more? I recently published a book that's called Mind Masters, which has a lot of the topics that we discussed in it. And then I also have a website, sandramatz.com. Thank you to Dr. Sandra Matz, a professor at Columbia Business School and the author of mind masters. What are three key takeaways that we got from today's conversation?

1:05:38Key takeaway number one, nice people struggle with money, but there is a fix. People who are agreeable, those who are trusting, caring, high in empathy, people who are agreeable consistently have worse financial outcomes than their less agreeable counterparts. That's what the data shows. There is a solution, and the solution is not to become meaner, but rather to reframe your financial goals around what you actually value, like protecting others or helping others. So the research shows that if you have a more competitive personality, then you respond better to framing saving and investing as a way of getting ahead in life.

1:06:20But by contrast, if you have a more agreeable personality, then you respond better to framing savings and investing as a way to support your loved ones, to make a positive impact in the world. So if you want the motivation to continue saving, to continue investing, and motivation, it's like showering. It's something that we need every day or at least every other day, right? It's like motivation doesn't last. That's why we need to continually re-up it. That's why we need to continually engage with it. That motivation comes from reframing your financial goals in a way that fits you and your personality.

1:06:58And that makes a much bigger difference than trying out some new budgeting app or trying out stricter rules. What we found somewhat counterintuitively was that people who score high on agreeableness, so those are the nice guys and people who are caring, they're trusting, they're empathetic. Those are usually the people that you want to have as friends and that hold society together in a way, but they also seem to have a harder time managing their finances. That is the first key takeaway. Key takeaway number two, if you want to predict your financial future, you can find evidence of that in your digital footprint because there are companies that are already using your social media activity, your spending patterns, your smartphone data.

1:07:41They're using all of that to predict whether or not you're going to default on loans, to predict how much you spend, or to predict what kind of products you're most likely to buy. When you understand this, you have some power because you can use the same AI tools as your personal financial advisor by feeding those AI tools your personality traits and your money goals and seeing how they reflect that back to you. That doesn't mean that they're a replacement for a human financial advisor, but they're a supplement. A research study that was done almost 10 years ago, just looking at the Facebook pages that people follow, we call our spouses our other halves, right?

1:08:23And they, similar to family members, go through life with us in many different situations, different moments. They see a lot of us, both in terms of how we want to be seen in public, but also sometimes these more intimate scenes where we might not be our ideal selves. And yet still an algorithm with just access to 300 of your Facebook pages can make more accurate predictions of how you think of yourself in terms of personality than those people who know us intimately. Technology is already being used to influence your spending decisions, but you can flip the script. You can use it to improve your own money management.

1:08:58So that's that second key takeaway. Finally, key takeaway number three, you can use AI to break out of your financial echo chamber. So the thing is, when we talk about digital echo chambers, I think we all know that social media, the algorithm feeds us the same echo chamber political content, for example, over and over and over again. In that same way, Your financial habits, your financial mindset, that can also get echo chambery and it can get stuck in a rut. You can prompt AI to show you how somebody with completely different characteristics would approach their money. For example, ask AI what a super risk-taking entrepreneur would do.

1:09:41Or ask AI how somebody from a totally different culture might handle retirement planning. Ask AI about different personalities, different phases of life, different life circumstances, and then prompt it like, what would a person in this circumstance, with this mindset, with this approach, and then with this type of portfolio, what would they do? Use it to sort of try on different hats, right? To see life through different perspectives, to break out of your echo chamber and to widen those horizons. because it's a way that you can discover financial strategies that you would never consider on your own.

1:10:18You can prompt a large language model to give you something else almost as an opportunity to say, well, I don't know what the reality of like someone who looks completely different to me, has like a completely different experience in their life, has grown up in a totally different way, lives in a completely different part of the world. I don't know what their experience is like and what they would want to do when they're in Barcelona or which movie they would want to watch on a Saturday night. But the large language model does. Those are three key takeaways from this conversation with Dr. Sandra Mads.

1:10:50Thank you so much for being part of the Afford Anything community. If you enjoyed today's episode, please do three things. First, share this with all of the people in your life. Share this with that agreeable person who puts everyone first and share this with that competitive person who's super motivated about getting ahead. Share this with the person who's got 300 Facebook likes and share it with the person who asks Google all kinds of questions that they would never actually ask their friends. And share it with the person who likes curly fries because you know that they're smart because smart people like curly fries.

1:11:22Share it with the person who talks about themselves using first-person pronouns when they're stressed and share it with the person who insists on doing salary negotiations by email. Share this with all of those people and more, because that's the single most important way that you can spread the message of F-I-I-R-E. Number two, open up your favorite podcast playing app, Apple Podcasts, Spotify, Pandora. Whatever it is you use to listen to this, open that app, hit the follow button to make sure you don't miss any of our amazing episodes, and while you're there, leave up to a five-star review. We absolutely appreciate it, and the more reviews we get, positive ones, the better of a chance that we have of bringing on big, awesome, thoughtful, insightful guests who can share knowledge with you.

1:12:09Also, head to our YouTube page, youtube.com slash afford anything. Hit the follow button there because the more subscribers we have, the better guests we can bring on. We have a course on how to negotiate. It's called Your Next Raise. It's been in beta for the past many, many months. And in August, we are going to be releasing it in its full non-beta version for the first time. If you would like more information about it, make sure you're subscribed to our newsletter, affordanything.com slash newsletter, because that's where we'll send out all of the info when we launch in August. Again, that's affordanything.com slash newsletter, where we will send out more info.

1:12:48Also, if you remember from the last First Friday episode, I promised that I would write a piece about student loans, about how to graduate from college debt-free. I'm going to publish that on the last day of July. Make sure that you're subscribed to our newsletter so that you can read that piece at the end of July. Again, that's affordanything.com slash newsletter. Totally free. Thank you again for being an afforder. This is the Afford Anything Podcast. I'm Paula Pant, and I'll meet you in the next episode.