Subscribe
Share
Share
Embed
Work Bench for raw audio interviews
Noah, welcome to Tech Nation.
Noah Giansiracusa:Thanks for having me.
Moira:Now the subtitle to your book is take control of the algorithms that run your life. What are algorithms, and how do they relate to what we load on our phones and tablets and laptops and using the Internet, and we can go on from there? How do they relate to all this technology in our lives?
Noah Giansiracusa:Sure. Well, very broadly, an algorithm is just some process often automated for making a decision or performing an action. But that could be so many things. That could be an algorithm for baking banana bread, be an algorithm for navigating to my grocery store in town, could be an algorithm for what videos I see on TikTok, could be an algorithm for whether I get a loan approved or not from a bank. So it's just anything that makes decisions or does actions.
Moira:And those actions run our lives. Now we're paying attention.
Noah Giansiracusa:More and more they do. That's right, and that's the point is there's always been algorithms in various forms, but they're just getting to be more and more and more of them, and so much of what we do is just filtered through algorithms. Just you wake up in the morning, you grab your phone, if you touch an app, I guarantee there's an algorithm involved. But you go out in the world and more and more there's algorithms, ordering at restaurants, buying at stores, driving cars, flying on airplanes, a lot of your work at your job, just more and more are algorithms. Part of that is because we have more technology in our life.
Noah Giansiracusa:Part of it is that technology is becoming more advanced. Things like AI just kind of turns everything into an algorithm.
Moira:And now computers run on numbers, you know, and this brings us to something you call numerification. What's that?
Noah Giansiracusa:Well, that's the idea that let's put it in a little context first. It's that algorithms, as you said, they're happening on computers usually, and computers love numbers. They just work a lot better with zeros and ones, but all kinds of numbers. So if you think about it, we're in a world full of people and nature and plants and animals, things that have nothing to do with math or numbers or algorithms, and yet this world is kind of being shoehorned into an algorithmic world where everything is binary and numbers and math and formulas. So, numerification is that process of turning us, people, individuals, our actions, turning us into numbers, into data.
Noah Giansiracusa:And that's been happening for a while, and it's funny because if you go back about a decade, people loved big data. It was the most exciting thing in the world that we have all this data, they can see what people are doing and where society's going and learn so much. And now after about ten years of that, it's kind of turning our stomachs because we realized we used to be humans, now we are just data points in these big data collections. And that's kind of unappealing in many ways.
Moira:The worm turned. Exactly. So let's get to Shakespeare. Yeah. Now let's talk about Robin Hood, the legend and the spirit of Robin Hood in this sea of data and algorithms.
Noah Giansiracusa:Right. So the legend is this ancient English figure who was stealing money from these corrupt kings and bringing that money back to the people. You can imagine, don't know the details, I don't know if the details are even known, but you can imagine that taxes are being levied on things and people feel like, Hey, that's our money that's being stolen. And this heroic figure who's sort of a thief, but sort of a hero comes in and takes that money back from rich and gives it to the people who feel like it's rightfully theirs. And Robin Hood, the spirit or the allegory, the metaphor, however you want to think about it, I think today we certainly have a lot of economic power, economic inequality.
Noah Giansiracusa:Money is still driving a lot of things. But I do think a lot of power nowadays is boiling down to algorithms, is boiling down to tech. If you look at the wealthiest five or 10 people on the planet, a huge percentage of them are tech moguls. If you look at the biggest, most successful companies in the world, most of them, if not all of them, are tech companies. So we're getting to this place in society where tech is power, where the wealthy, the top, the people leading the world in society are not the kings.
Noah Giansiracusa:They are the tech moguls. They are the Elon Musks, for it to be very concrete, and others like him. Well, Elon Musk and others like him, they love efficiency, they love optimization, they love squeezing out profit, they love reducing headcounts on staff, and they use algorithms to do this, to look for patterns and data, to look for redundancies that they can exploit, and in many cases recently, to just replace humans with artificial intelligence. So, people at the top have become very wealthy, because they've ridden this wave of economic growth in the tech sector. They've been the ones using formulas and algorithms and statistics and data.
Noah Giansiracusa:That's what drove them to the top. And now that they're at the top, they're using that power to kind of amplify it, to make our world even more algorithmic, even more optimized. Where does that leave the rest of us? Kind of in the dust. I just feel like so much of my life has been optimized for someone else's profit.
Noah Giansiracusa:And it does feel a lot like this medieval Robin Hood story, where someone's life is designed around paying taxes to this corrupt king. And I'm just thinking, we've had enough. How do we take power back? Where's the Robin Hood in our day? Who's gonna take power from these tech moguls and give it back to society?
Noah Giansiracusa:And I thought a long while about this, and there hasn't been a lot of legislation passed, especially in The US, but elsewhere, that tries to level the playing field. And I just thought, you know, what's giving these algorithms and these tech people and tech companies so much power, what enables them to optimize and do all the things they're doing? It's math. And heck, I'm a math professor. Maybe I can take some of the math that they're using, and I can give it to the rest of society and help everyone get some of that same benefit.
Moira:And I have to tell everybody, because this is audio, he's not wearing a Robin Hood costume. You just have to take my my word for it.
Noah Giansiracusa:But I am wearing green today.
Moira:Oh, there you go. There you go. And you're not saying we have to stop it. You're saying we have to sort of mitigate it for ourselves. We have to navigate it for ourselves.
Moira:So the question is, how do we do that? Where where should we start?
Noah Giansiracusa:Well, I think where we start is understanding the tools that these Empower, these companies are using. You could say using to exploit us, or you could just say using to make money to succeed the way companies do. But let's give a really concrete example. Meta is very, very successful. TikTok is very successful.
Noah Giansiracusa:They use algorithms to decide what we see. If I'm swiping on social media, whether it's Facebook or Instagram, X, TikTok, YouTube, all these things, they're all using algorithms to determine what we see. And they're trying to maximize the amount of time that we spend on there, the number of ads that we see, the number of ads that we click, the profit that they generate from us. The consequence is that people like you and me and everyone else who are just using these apps, who aren't Mark Zuckerberg and others, we're just kind of getting pushed around. We're seeing a lot of videos and a lot of posts that aren't necessarily the ones we want to see, that we were hoping to see online.
Noah Giansiracusa:We're just seeing what the algorithm decided. But remember, the algorithm is not representing your best interests. The algorithm is trying to outsmart you and get you addicted and spending money on the platform, through watching ads and clicking ads, or just whatever else, wherever they send you. So, I think understanding how these algorithms work is the main start, because they're the ones that are pushing us around and taking advantage of us. And the second step, once you know how they work, is maybe we can do what you're saying, mitigate a little bit.
Noah Giansiracusa:Can we resist that algorithmic pull? Can I use social media in a way where I'll see what I want to see, not what Mark Zuckerberg and others want me to see? Can I do that? I think the key is understanding how these algorithms work. Otherwise, we're just kind of shooting in the dark.
Moira:You're listening to Tech Nation. I'm Mori Raghunn, and my guest today is Doctor. Noah Chancirakisa. He's an associate professor of mathematics at Bentley University and a visiting scholar at Harvard. He focuses on algorithms and society, and he's here today with Robin Hood Math.
Moira:Take control of the algorithms that run your life. Well, certainly in the past, and if we don't think about it, we'll go forward. We're always trying to find the one person, the one expert, the one doesn't have to be expert, but someone who understands and, you know, they'll I'll just talk to this person and they'll get me straightened out.
Noah Giansiracusa:The guru.
Moira:The guru. Well, it could be a financial adviser. It could be one doctor. It could be anything. Now there could be information out there, an analysis.
Moira:And that looking for one person, you say, you've got to change that.
Noah Giansiracusa:Yeah, so the context for what you're asking, this is an example of taking a mathematical idea or tool that everyone who's been kind of trained in quantitative and mathematical methods, or at least a lot of people use freely. And yet, most of society is left behind and no one explains. There's actually this really simple thing that you can do that'll help you understand what's happening and make better decisions. And nobody bothered to tell you about this. So this is an example of a Robin Hood example, Robin Hood math, where I think we're not stealing a mathematical idea from those in power, but we're just saying, Hey, they're doing it.
Noah Giansiracusa:Why can't we too? So what's the story? Well, one of the interesting examples you mentioned, you had a great list of all these situations where I rely on experts and I want to know what's going on. One of them, my favorite one you mentioned, is a doctor. And we all know after you go to a doctor, what's a reasonable thing to do?
Noah Giansiracusa:You get one opinion, you get another opinion. You go to one doctor, you get diagnosed with something. You go to another doctor, you get another diagnosis. So, okay, that seems pretty easy, right? But what do you do with that information?
Noah Giansiracusa:Let's say I go to one doctor, they say I have disease X. The second doctor says I have disease Y. What do you do? You could just think, Well, the second doctor was the most recent, so I have that second disease. Or The first one is the doctor I trusted, I didn't really like.
Noah Giansiracusa:You get kind of lost. Now you have two different opinions and you don't feel any better off. In fact, you feel less confident than before. You could see a third doctor and then where do you stand? Well, maybe you see the third doctor and the third doctor says, Yeah, you have disease X.
Noah Giansiracusa:Now I have two votes for X and only one for Y. Now I'm starting to feel some clarity. I think, Okay, I think I see what's going on. It's not unanimous because doctors don't know for sure. The world is probabilistic.
Noah Giansiracusa:There's a lot of randomness. They're guessing, but it's a very, very educated guess compared to what I would do on my own. So an educated guess, I can't wait too much on it, but if I have multiple educated guesses, I can look for a pattern, a trend. And here I'm seeing two against one, I'd be pretty confident saying, You know what? I'm going to go with that too.
Noah Giansiracusa:And that's a really reasonable thing to do. There's even math theory and formula saying that's a very reasonable thing. But there's nothing special about doctors, or diagnoses, or even the number three. You can always do this. If you're trying to figure out what the mortgage rates, or inflation rates, or unemployment rates might be a year or two from now, you might find a liberal think tank that says one thing, a conservative think tank that says another, libertarian one that says something.
Noah Giansiracusa:You can get lots of different opinions. This happens all the time. What's wrong with just taking all those guesses or those predictions, whatever you want to call it, and averaging them? Turns out nothing. That works great.
Noah Giansiracusa:In fact, that almost always is better than just relying on a single one. I might think I'm a Democrat, I'm going to follow Democratic think tanks and see what they see. They might be blinded by their own ideology and prejudice, and they might miss something that the Republicans are talking about. But I don't want to fully trust everything I read in their think tanks. Just take their estimates and average them.
Noah Giansiracusa:Take their predictions and average them. Even silly things like the rain. You do your own numerification.
Moira:Yeah. You do your own numerification.
Noah Giansiracusa:That's exactly right. They're sort of doing it, and each person is predicting, and I want to combine these predictions. And you're hitting the nail on the head, which is how do you combine predictions? You kind of have to turn them into numbers, because numbers we know how to combine. We can average them.
Noah Giansiracusa:So that's actually key is when you do want to leverage multiple opinions, often it's helpful to translate them into numbers and that could be a probability. So let's say you're interested in the weather. Maybe I think there's a 50% chance it'll rain tomorrow, some weather app says 70% chance, another weather app says 30% chance. I can just take those numbers, they've already numerified the weather, and I can average them. So averaging turns out to be a very, very important tool, and yet it's kind of amazing how widely it's used.
Noah Giansiracusa:So to just give you one example, Nate Silver became very, very famous for making these incredible political predictions, predicting not just which person would win the presidential election December, 02/2012, and to some extent in '16. But he predicted which states they would win, and it was very, very accurate. So what did he do? He took a whole bunch of previous polls and predictions, and he averaged them together. And it turned out that is very, very helpful.
Moira:But it's not just averaging. It's also what you happen to know.
Noah Giansiracusa:It's what you happen to know, and I did sweep one thing under the rug, but I'm glad you gave me a window to bring it back.
Moira:Bring it back out?
Noah Giansiracusa:Which works best is the weighted average. So what's a weighted average? You can actually mathematically put more weight or credence or importance on some predictions than others. So when Nate Silver did this, he didn't just use all polls and treat them equally, he looked at the ones that had the best prior track record, and he gave them the most weight. You might think, well, wait a minute, why not just use the one with the best track record?
Noah Giansiracusa:Why are you bothering averaging the good ones with a bunch of bad ones? It's because having a track record means there's a good chance you're going to do well in the next election or whatever you're predicting, but it's not a guarantee. What if that one goes wrong and a different one with the worst track record has a really good year? So averaging allows you to kind of exploit the best of all of them, rather than just limit yourself to a single one. And it's just like we said with doctors.
Noah Giansiracusa:I don't want to just find the one doctor that I fully trust and just go with it. I'd like to have a whole bunch of opinions and see that, yeah, they are converging to one view that may not be consensus, but it's better than just randomly picking one doctor and going with it.
Moira:Now let's talk about risk. In fact, let's talk about managing risk versus managing your happiness.
Noah Giansiracusa:All right, great transition, because we were just talking about Nate Silver trying to predict elections, And a lot of risk management is trying to predict what's going to happen in your life. This could be predicting whether you're in a car crash. It could be predicting maybe I don't want to go on a drive because the roads might be icy. That depends on the weather. And that's a prediction.
Noah Giansiracusa:It could be predicting how a stock is going to do. Maybe I'll put all my life savings in a particular really promising stock, but I don't know how that stock's going to go. And if that stock tanks, there go my life savings. So a lot of risk management ends up being about making good predictions. So the very first thing I'd say with risk management is use the technique we just talked about average different predictions to get better predictions.
Noah Giansiracusa:But it goes more than that. One of the biggest concepts that's it's not even a formula, it's just a way of thinking is we tend to focus on a specific outcome, and it's actually very helpful to think of the whole distribution. So for instance, I might think, oh, this stock is projected to be worth $100 a share in a year. Okay, that sounds you know, whatever I'm just making up these numbers that sounds good, I might go ahead and invest. But there might be one stock that's predicted to end up around $100 a share, and the range might be something like $80 to $120 So it's going to go up to about $100 all the experts agree.
Noah Giansiracusa:They're all admitting they might be off by 10 or 20, but not really more than that. There might be another stock that people think will go up to 100 a share, but it might range from 0 to 200. So, oh, maybe we're wrong and the stock is worth twice as much as we think, or maybe it's worth half as much. And this is a stock where we believe there's a bigger distribution of potential values, meaning we can kind of figure out where we think it might end up, but there's a big range around that. And the other stock is much narrower.
Noah Giansiracusa:And it's a pretty I don't want to say simple, but a very elementary, important concept. This idea that anytime you're dealing with any kind of uncertainty trying to guess what's happening you have to really be careful to not just focus on what people think will happen, or even what you think will happen, but what are the range of outcomes. Because if a stock goes up, you think it's going up to $100 and it ends up at $200 great! You did twice as good as you thought! If it ends up at $0 you lost all your money.
Noah Giansiracusa:If it goes up to $100 and then it maybe wasn't quite as good as $100 it only reached $80 maybe that's good enough. So this is just to say, risk is a lot about considering the range of possible things that might happen, and trying to come up with probabilities or how likely these different things are, rather than just saying, what's the most likely thing? I got to consider the cases that may not be the most likely, but if they could still happen, I better still consider them.
Moira:And to throw in an old saying, don't put your eggs all in one basket.
Noah Giansiracusa:Yeah. Or another way is, Jane.
Moira:So if you're talking about all the money you have, you might not have anything by the end of this.
Noah Giansiracusa:You know what's funny?
Moira:And you're gonna be very unhappy. No doubt about it.
Noah Giansiracusa:That's right. And we'll come back to happiness because you brought it up. But just to finish fun there, the eggs in one basket thing, one mathy way of saying that is to diversify. You're suggesting you should diversify your investments or your actions, and that's absolutely true. And what's funny to me is diversifying really means averaging.
Noah Giansiracusa:Rather than going with one investment, I can invest in five different things, and that is a form mathematically of averaging my investments across the five. So it's no different than we talked about with predicting elections by averaging polls, going to doctors and averaging their predictions in a sense, diversifying just like you said, don't put all your eggs in one basket. It's all the same message. Just spread your actions and your guesses by averaging across the range of things you're interested in.
Moira:Now there's happiness. Oh,
Noah Giansiracusa:happiness, right. So let's not forget happiness. So, the other thing I think helps to understand with risk is we tend to fixate on numbers a little too much, and specifically if we're talking about money, tend to think about financial things, we tend to think about dollar amounts. And I think that can be misleading, in a dangerous way. Not that dollars are not true, you know, they are what they are, But I think if we think of how much getting a big promotion would benefit our life, we kind of overestimate the impact of a dollar amount.
Noah Giansiracusa:If my salary goes up to this amount, or this amount, or that amount, We just focus on those dollar amounts. What these psychology studies have found is what matters more than the number of dollars when you're talking about your net worth, your wage, your income, any of these things what matters more than the number of dollars is the number of digits in these things. So, is your salary going to go up by a whole digit, or is it going to come close to that? That matters for your happiness. But just imagine, I can tell you, I don't earn a million dollars a year.
Noah Giansiracusa:But if someone said they earn $2,000,000 versus $3,000,000 is that a big difference? From where I sit, I would say no. That's going from wealthy to wealthier, but it's not a big difference. Mathematically you might say, Oh, those are a million dollars apart. So this is just to say, don't focus on the amount.
Noah Giansiracusa:When you're trying to figure out anything involving risks, means should I quit my job to take a different job? Should I put all my money in something? Should I invest in this? Just remember, it's going to take more money to get that extra bit of happiness than you think. And that's just because digits seem to matter more.
Noah Giansiracusa:Like people always say, a 6 figure salary, a 7 figure salary, an 8 figure. People who use that terminology, I think that's the right terminology. Psychology studies have shown figures are what matter.
Moira:So if I make $100,000 I'm really not gonna get happier until I make a million. Okay. Well Send it over, Noah. Send it over. Venmo it to me.
Noah Giansiracusa:All of nothing. I wouldn't give up a lot to go from 100 to $1.50. Let's even I don't know if you're joking a bit, but let's take it even more extreme. Warren Buffett. Imagine the first time he became a billionaire.
Noah Giansiracusa:That must have been a really big deal. Now imagine Warren Buffett when he went from $100,000,000,000 net worth to $101,000,000,000 net worth. There's no way that billion dollars mattered to him at all. So, is just a reminder: we can't fixate on these exact numerical amounts, we just have to think, how much impact is it having on my overall financial well-being. And that's that's kind of where the happiness comes in.
Moira:Now we have to talk about the Reverend Thomas Bayes, and he hasn't been with us for a couple of centuries now. But he talks about, even then, revising the strength of your convictions in light of new evidence. And boy, today, new evidence is appearing all the time.
Noah Giansiracusa:Well, that's right. So he's an interesting character. He was a Presbyterian minister in the eighteenth century. He wrote an entirety of two math papers. One had to do with calculus, and one was this kind of obscure thing that never got published.
Noah Giansiracusa:He passed away, and his friend was going through his papers and found that other discarded paper, and thought, This proves the existence of God. The friend was very enthusiastic and talked about it, I would say overzealously. History has kind of lost interest in that direction of the paper, but what people have extracted from the paper is there's this really important formula that we now call Bayes formula, which goes back to exactly what you're saying. The idea is probability often we think of as measuring how often something happens. So if I flip a coin, there's a 50% chance it's heads or tails.
Noah Giansiracusa:If I roll a die, there's a one in six chance that I get each specific number. That's how we tend to think of probability, but probability can also express a conviction. I really believe that there's a 90% chance that I'm going to meet the love of my life in the next few years. That's a statement I can make. It's not a mathematical statement, but it's using numbers and probability to express a conviction.
Noah Giansiracusa:I could talk about my conviction that there's intelligent alien life out there somewhere anything, that there's bodies buried under my house, it could be far fetched or close anything. I can put numbers to express my belief in something. Why would you do that? Because it's so vague to just say, I think something is likely, or very likely, or not likely, or unlikely. When you use words like that, nobody has any idea what you're talking about.
Noah Giansiracusa:You don't know, like, how likely is likely, or unlikely, or
Moira:Is likely 51%, or likely 98%.
Noah Giansiracusa:Exactly, it could be anything. So people have kind of realized, even if you don't know the thing you're trying to guess, you know, are there aliens? Nobody knows. It still matters to say, I think there's a 90% chance there's aliens, and you might think there's a 50% chance. It's still helpful to use numbers.
Noah Giansiracusa:So with that in mind, we can go back to Bayes formula, and it's a formula that basically says, if you have some conviction, you have some probability in your head of how likely something is of happening falling in love or meeting aliens, or maybe both, who knows then some new thing happens, some new event. There's an earthquake and you discover something, or you see some UFO flying through the sky. Whatever it is, something happens. That might be rare, but the probability of that other thing happening should impact your first probability. Just to say more clearly, when I see evidence, I should adjust my beliefs.
Noah Giansiracusa:And this is very relevant, not just in these made up alien types of examples, but things like presidential election, let's say. I might think that so and so has an 80% chance of winning the next election, and then some scandal breaks, or the stock market crashes, lots of things can happen. Lots of new events can happen. And Bayes' formula helps us use those new events to update your probability. And we love having that formula because without it, let's say I think there's an 80% chance that someone's going to win a presidential election and then the stock market crashes.
Noah Giansiracusa:What's the chance now? I don't know, 90%, 50%, you could just kind of guess. And Bayes formula, it helps guide you to the right numbers to adjust your old belief based on the new evidence. And that is very, very useful. And that is used not just in political settings, just for aliens, but tech companies that use that all the time in their algorithms to look for patterns and data.
Noah Giansiracusa:So it's everywhere. It's an amazing formula.
Moira:And you don't have to have any numbers to know that it's so comforting as a human to say, I got this down. I understand it. I know. And that we have to remind ourselves, things change all the time, and we have to reset how we're thinking and what we think is the right way to go. So, I really appreciate it.
Moira:Well, Noah, thank you so much. I hope you come back and see us again.
Noah Giansiracusa:Absolutely. Thank you.
Moira:My guest today is Bentley University professor, Noah John Sirocooza. His book is Robin Hood math. Take control of the algorithms that run your life. It's published by Riverhead Books. I'm Moira Gunn.
Moira:You're listening to Tech Nation. It's published by Riverhead Books. For Tech Nation, I'm Moira Gunn. Hey. You're going on air now.
Noah Giansiracusa:Alright. This is good. That was good. Right? It felt
Moira:it felt a lot better. That was good. Yeah. Yeah. Yeah.
Moira:It did. Okay. Great. So now