AI Agents Are Already Changing the Price of Goods | Odd Lots

Vídeo em Inglês, publicado pelo canal Bloomberg Podcasts.

About this content: The summary, key points and topics below are generated automatically from the video transcript. They reflect the content, opinions, claims and predictions presented in the video and do not represent a position, validation or fact-check by IA Dots.

Video summary

This video discusses modern pricing strategies, including algorithmic pricing, dynamic pricing, and surveillance pricing, and explores how companies and AI agents interact in the e-commerce space.

Key points

• Pricing has evolved from a simple cost-plus model into a highly engineered science utilizing AI algorithms and dynamic pricing.
• Surveillance pricing allows companies to set personalized prices based on consumer data, potentially leading to widely different prices for the exact same item.
• Experiments with AI shopping agents have shown that agents may sometimes route users toward sponsored or more expensive options rather than the absolute cheapest ones.
• The concept of fixed, transparent price tags emerged historically to eliminate haggling and provide predictability for consumers.
• Companies are increasingly using alternative profit streams, such as data monetization and algorithmic pricing optimization, to test the limits of consumer willingness to pay.

Topics

Algorithmic PricingDynamic PricingSurveillance PricingAI Shopping AgentsRetail EconomicsConsumer BehaviorInflation

Video transcript

Transcript from available captions; it may contain errors.

I love the idea of folks having help getting out of subscriptions. I love the idea that folks would have help with comparison shopping. Comparison shopping is a total bear, in this digital and e-commerce age, precisely because of the pricing technologies that I've talked about. The problem in my mind is the due diligence. Right? Like you're using an agent, presumably because you're already time strapped and don't have time to go look at all of this yourself. And then you're going to review the output of the agent and be like, well, was this really the cheapest price? How do I know that the agent is actually finding me the best bargain? Which Lindsay, by the way, is why you are proposing fiduciary duties for AI agents, right? Yeah. I mean, I don't think this problem is new, right? We often I think with new technologies, I think we need a whole new set of considerations or a whole new set of policies. But this is just the classic agent question who does your agent work for? Yeah, if it's calling itself an agent, and particularly if it's calling itself your personal agent, it should probably work for you and not muse or whoever is offering sponsored content. And the highest bidder, it should probably not be rooting around in your emails to work against you. Hello and welcome to another episode of the Odd Lots podcast. I'm Tracy Alloway and I'm Joe Weisenthal. Joe, have you tried muse yet? Muse. You actually did. I'm sure you couldn't help, but I did use it quickly and I found it to be very impressive on day one, and it actually saved me several thousand dollars because it identified sending an email that I had missed. But then I haven't used it since, and I've actually seen the stats of like a bunch of people installed it and usage has tailed off, but it was like, whoa, this is very powerful. What did it identify it? This was I'm dealing with something with my house or something, and someone sent me a long email, and it was so long that I didn't bother to read the last paragraph. And it turned out there were some important details in the paragraph. Okay, so it it read for you. Yeah, read for me. All right. I thought you were going to say something about the actual wallet and shopping aspect, because one of the things that's being pitched when it comes to muse is this idea. Well, now you have a personal shopper kind of at your fingertips. And so there's a democratization of shopping where you can tell the agent to go out and find deals for you and scan thousands of prices and come up with, I guess, the best opportunity for you. No, it's totally true. You know, obviously, look, we've had like price comparison engines before, whether it comes to like airlines, you know, Booking.com compare prices and so forth. But obviously like this idea of like this complete end to end ness, it may get to the point where they hope to get it to the point where it's like it orders the, trip before, you know, you want to take a trip or whatever, like possibilities like that. I still think, and this I think this will be very temporary. Like, it doesn't quite work in the sense that for many things, if you buy them online through an AI agent, it does this weird thing where like opens a browser, then the mouse moves really slowly and there. I don't think that's like the future of commerce is a computer using a browser that was designed for humans. Nonetheless, the broad premise clearly is going to be a bigger and bigger deal in our lives. Now, the flip side of buyers having AI powered shopping solutions is that the sellers also have them, right, and this has been going on for some time. We've talked on the show before about things like algorithmic pricing, dynamic pricing, but now you can imagine a world where that is just super charged, right? Yeah. And so then you get this arms race phenomenon where both sides have their agents trying to battle with other agents. And unfortunately, you know, probably the, the, the vendor has more data, right? They just going to have an information advantage over you. Maybe they're going to be companies that try to stop that. Whatever. It could just be that the only thing that happens is the only real win. Are the AI companies that both sides are sending a lot of money to to win this arms race. But yeah, like the future of shopping, the future of pricing is very much wide open right now. Yeah. So we do in fact have the perfect guest to talk about all of this. Someone who just wrote a book that does a deep dive basically into a lot of the pricing strategies that companies are pursuing nowadays and is doing some interesting research when it comes to algorithmic pricing, AI powered algorithmic pricing, I should say. So without further ado, we have Lindsay Owens. She is the author of the new book, Gouged at the End of a Fair Price and What that Means for Your Wallet. She is also, of course, president of Groundwork Collaborative. So, Lindsay, thank you so much for coming on the show. Thanks so much for having me. Why don't we start with semantics? Because we hear all these words, right? We hear like algorithmic pricing, dynamic pricing. The companies themselves often have slightly different terms that they use, like profit led pricing and things like that. What are all these terms and how do you think about it in your head? Yeah, the main thing to think about is that pricing has really changed. As we know it, it used to be that pricing was pretty simple, right? Companies figured out how much it cost to make an item. They put their margin on top. They put it on a sticker. And that's really not how pricing works anymore. It's a highly engineered science. And companies hire physicists to help them do pricing. They buy AI companies to build algorithms and to run experiments on their platforms to get better and better at determining the optimal price and how much you'll pay for really everything in your grocery cart. So pricing has really changed. The terms are, useful, I think, to explain algorithmic pricing, I think is really just an umbrella term for pricing that involves an algorithm. There's a formula behind it, dynamic pricing we typically think of as pricing that varies based on market conditions, changes in supply and demand. This is personalized pricing. I forgot that one. Personalized pricing is really the same thing is surveillance pricing, which economists would call first degree price discrimination. So this is, a company that's setting a price just for you, ideally based on information that gives them a sense of how much they can get you to pay for it. I might ask a bunch of devil's advocate questions. I, during this interview, I don't like the idea of, okay, someone. It's like, oh, it's Joe Weisenthal shopping for this. He works at Bloomberg, probably gets paid pretty well, etc.. So we're going to, you know, we're going to find the optimal price point, that he might be willing to say yes to. I don't love that. On the other hand, could this be used for someone who is in a lot worse condition than me? Who the algorithm determined does not have as good of a job I have? And maybe they get a really good price means testing for Big Macs. Yeah. That's right. Well, the first thing we should say is this is definitely going to happen to you. And we've got some really good examples of this. Okay, so Expedia has been found to charge more for the exact same Manhattan hotel room. If you log on from the Bay area where they think you're more affluent and a lot more. Right. So this is going to happen. We have now a really interesting study of a genetic commerce that just came out. They did an experiment with 325,000 agents, and they asked the agents to go find the cheapest item. Okay. Even when directed explicitly to go get the cheapest item, the agents routed around in your email inbox identified that you could indeed afford more than the cheapest option and brought back a more expensive option, which I need to be clear in here an agent. So an agent that's working on my behalf. Are you saying would not actually be working on my behalf? That's exactly right. That's what the experiment discovered. They asked the agents to buy flights. These were pretty big ticket items, health insurance plans, and to select a graduate program. And even if you said, bring me the cheapest health insurance plan, once the agent figured out a little bit about your financial profile. Yeah, basically they said, hey, I think this guy can afford the gold plan. What's your thesis for? What's going on there? Because I can imagine maybe something like muse. I mean, meta is also an advertising platform, right? So maybe it wants to route some business to particular companies that advertise a lot on its platform. But maybe there's something else going on too. Yeah. So we have one other study that gives us a little bit of a hint here. This one is not looking at agents, but it looked at 13 different alums. And what they found is again, even when directed to bring back the cheapest product, they brought back the sponsored product and often the sponsored product was more expensive. Can you tell us just by way of another example, you did this really big Instacart study, and it was kind of interesting to read that you used to work in retail. Yeah. And that your nightmare was like the what do they call them, the secret shopper, secret shoppers. And now you get to hire a bunch of secret shoppers, basically, and turn the tables. What did you find? Yeah, the Secret Shopper was always interesting because it was unclear whether you were ever actually secret shopped, but it was a sort of management tool of surveillance. Like, you better behave because at any point somebody could come in and secret shot, like every customer could be actually working for HQ, right? So you got to make sure you fold everything well, and you put the clothes on the rack and size order things like that. So we recruited 400 secret shoppers. We worked with Consumer Reports and more perfect union. And then we had them fire up the Instacart app shop in four different locations across the country at the exact same time, buy the exact same groceries for pickup so not delivery from the same grocery store. 23 items. What we found 75% of the items were differently priced for different shoppers. Same time, same place. The differences were as big as 23%, so you could be charged 23% more for a dozen Lucerne eggs than I was charged from the same Safeway in Washington, DC at the exact same time. We estimated that because of these differences, some families could end up paying as much as $1,200 more in a given year. This was based on Instacart, its own estimate of how much a household of four spent shopping. And I think it really was, you know, it's a blockbuster report, right? It was the cover of the New York Post. You don't usually see a think tank paper on the cover of the New York Post, because I think what it illuminated is that something that we all took for granted fairness in pricing has eroded this idea that you and I went to the grocery store. We stood in line next to each other with the same item. We expect to pay the same amount. And that's no longer the case. So you talk about this in the book, but the idea of one price that everyone can see is also, I guess, kind of new in American history. And I didn't realize apparently there are Quaker routes to public price labels. Yeah, for thousands of years we haggled. Yeah. Across the globe. Yeah. And there was haggling in the US context as well. The Puritans were Hagler's and the Quakers didn't like it. They felt like there were, moral problems with haggling. All men are created equal under God. They thought, and we should pay the same amount for a bushel of apples at the Quaker marketplace. So they skewed haggling and moved to fixed prices. And then John Wanamaker, the local Philadelphia retailer, and his iconic American Wanamaker's department store. He was said to be a religious man, himself a Presbyterian, but inspired by the local Quakers nearby and instituted the first fixed price, the price tag and Wanamaker's department store, in the late 1870s. So we've had a price tag for about 150 years in this country. And I think it's given us a lot. It gives you predictability. Yeah. So you can shop on a budget. You know, how much something is going to cost in the grocery store. You know, you used to be able to get a Sunday circular where you knew what the prices of grocery would be for the entire week, right? A pound of meat, a pound of apples. It also offer transparency, which is pretty important for comparison. Shopping and comparison shopping is, of course, important if you're on a budget, but it's also the way in which prices are kept competitive in the entire economy. Right. That's interesting to think about fixed prices as a public good in a sense. Right. That like, okay, if there is just a price that everybody knows and it seems like that's what you're going to get. Yeah, that there is a positive externality from this is the price and everyone knows it and everyone can plan around it. Absolutely. I think the predictability is core. I do, by the way, I think comparison shopping is being eroded by all of the volatility in pricing right now. And I think budget shoppers who are really essential for keeping our economy healthy, right. When people have to price to include, budget shoppers. That's good because that keeps prices low for all of us when they're able to be picked off. You know, if a store can identify who they are and quietly push them a discount, prices will go up for the rest of us. When you put research like that on the cover of the New York Post, obviously they're going to go with the headline. It's like, you may be paying 1200 more for groceries than someone else, but you may be paying 1200 less for for groceries than something else. That's pretty good, right? Yeah. Look, definitely it was the case that different people in the Instacart study made out a lot better than others. And there were there were sort of five distinct price groups. Okay. So it wasn't, you know, we do think the prices were randomly assigned. That's what Instacart said. They were just running pretty sophisticated AB tests. And Instacart in 2022 bought a company called Ever Said. And this is sort of how we knew to run this experiment, because ever say it was an AI company that built a platform to allow retailers to run pricing experiments on their own platforms, it was sort of turnkey pricing experiments. And importantly, without the consumers knowledge. Right, because this isn't something that people like when they find out they're paying more than their friend for the same item at the same time. And so Instacart purchased them in 2022, obviously started using the product. And I think what we ultimately exposed with our experiment was their experiments on shoppers. So setting aside personalized pricing for a while, there is a general phenomenon in the US economy of prices seem to be going up, right? Yes. I went I went to a Broadway show yesterday. Very expensive these days. I bought one cocktail. It was a double $25 way off. What $48? Oh, it's crazy. $48 for admission. Holy smokes. I probably should know this. I was surprised this is the David Brooks like. But my point is, like, I probably would have ordered more than one cocktail if it was less than $48, substantially less than $48. But I don't think that matters that much to the bar at the theater, because if they're making $48 from a single cocktail, then they save on labor costs and everything else. Is this is this the dynamic that you see throughout the U.S. economy? The idea of like companies pursuing profit over, I guess, cost price over volume. Yeah. Yeah, absolutely. I mean, I think we're in an inflationary period. This is not a period in which, you can run an aggressive cost cutting play. Your costs are going up, as a business. And I think pricing is where companies have headed and they're really pushing the envelope there. You know, the k-shaped economy is real. Like we see CEOs talking a lot about it in earnings calls. They are really testing the outer bounds of what they can get folks with disposable income to fork over. We did an experiment a couple of weeks ago for the today show. And we helped them design it. They ran it on their production assistants. They brought them into the conference room and everybody fired up StubHub and purchased, or planned to purchase a Harry styles concert ticket for Madison Square Garden same night, same row, same seat. What we found ticket prices were constant across all of the production. Assistance fees were varying by as much as 60 bucks. Well, yes, StubHub charging $60 more in fees. This is the processing fee, which yeah, I don't the QR code. Why it exists anyway. But yeah, the fee for emailing you a QR code. Yeah. We did they find the okay among the, 12 production assistants. They're like, yeah, I'm the rich one. Like, yeah, I know why. Like, were they able to intuit what it was about themselves that, the algorithm knew that they could get away with charging them no more? Yeah. I mean, it's hard to pinpoint personalized pricing in a sample of, like, 12 production assistants, but there were a couple of interesting clues. So the first thing I'll say is when they shopped in incognito mode, feed dispersion dramatically dropped. Oh, so it went from $60 to $4. By the way, StubHub responded to the today show and said, this isn't personalized pricing, it's AB testing, it's pricing experiments, and that the incognito mode issue was sort of moot. The other thing they found is the Android phone user was paying more, but, you know, it's impossible to isolate whether or not there were other aspects of her profile that could have been driving her price point. We had you on the podcast before, David Di yeah. Now, in that conversation, I remember one of the things and, this question of Uber pricing and at least the technology existed such that if it identified that your phone was running low on batteries, that it would charge you more. But if I'm out at 1 a.m. and my phone is at 5%, I happily pay more. And in fact, if, let's put it this way, if I if we're if I'm out at 1 a.m. and my friend is out at 1 a.m. and his phone is at 50% and I'm at 5%, I want my app to be outbidding for that random driver who's in the neighborhood. That seems great because you want the ride. I want the ride, absolutely. And so it's like, I think most people, if their phone is at 5%, they're like, yeah, let me win this little auction that's going to occur between this driver because I really I if I don't win it then my battery, I might not have a phone. This is a true economist mindset. The idea that like auctions are the best way of coming up with a price. Well, I kind of don't. I know Americans like lines. Americans hate auctions. Yeah, I know, I know, but like, if I'm out at 1 a.m. and my phone's about to die and a bunch of people want the car, I would be happy that the app is, offering to pay their driver more if he picked me up because my phone's about to die. Yeah, we actually found a battery example with DoorDash as well. Okay, so we reviewed some DoorDash patents, and I should say that DoorDash, after we exposed this patent, did withdraw the patent. So that's an important thing to add to this story. But the patent was for a 16 factor agitation index that DoorDash was building to determine. So battery life would just be like one of the 16 factors. Another one was how quickly you scroll. I mean, it was basically a hangry index, right? How quickly you scroll. Another was how jerkily you scroll. And all of this was going into an index to be matched to things like whether or not you were going to get a promotion, what types of offerings you would see first in the menu, sort of the ordering of the menu. All of this was connected. So. So battery life is definitely something that the companies see as an opportunity. You're over a barrel when you're battery life is down. It's a little bit like, you know, there are five bottles of water left in a hurricane. You're willing to pay a lot for them, right? So people think about price gouging as this loaded term. Yeah. And again, you know, per Joe's point, one man's price gouging is another man's, I guess, legitimate profit ride savior. Ride Savior. Yeah. But there are actually regulations around price gouging in the US. Can you explain what they are and what the definition actually is? Yeah, the legal definition of price gouging is pretty straightforward. It's charging unconscionably excessive. You know, how you define that varies a little bit of moral relativism there. But yeah. Yeah. But it's charging unconscionably excessive prices during an emergency basically. It typically is limited to necessities. So if you're charging a bunch for a Birkin bag, like you're not going to run afoul of price gouging laws. But if you're charging a lot for food after a hurricane Birkin emergency luxury handbag emergency raid if you're charging a lot for food, after a hurricane or water if you're charging a bunch for PPE during a pandemic. Many states actually updated or brought new price gouging laws online during the pandemic, but more than 40 states have price gouging laws in the country. The title of my book is certainly referring to the legal definition of price gouging, but it's also referring to the colloquial feeling of being ripped off or squeezed or gouged, which I think is a word a lot of Americans are using right now when they see the their grocery bill. Who are the pioneers of scientific pricing? Where did this like who were the first to recognize that with technology, with the abundance of data, there were all the, the exhaustive data that we all admit just by using devices and so forth. Who are the pioneers that recognized that this could be used for pricing? Yeah, I really think the OG here is the airline industry. And there are some interesting reasons for that. One is obviously they have a scarce resource, right. You only have so many seats on the plane. So dynamic pricing can be an efficient way to allocate a scarce resource. But it's also why when you log on in, there are five seats left. The prices are high. When there's one seat left, the price is quite high. Right. But the airlines get deregulated, they have price controls. And all of a sudden, the airline executives are like, wait, how do we set prices? And they hire a bunch of consultants, advisors to build what becomes the first revenue management and yield management systems really in the country. And they really pioneered the dynamic pricing model. And it was actually pretty rare to see a dynamic pricing outside of the airlines for decades. It was 2011 when Uber rolled out surge pricing for the first time. Halloween, actually, 2011, they said, hey, it's going to be a busy night tonight. We expect a lot of people out on the streets. We're going to try this new model that's going to make sure you can get home. Same year Ticketmaster 2011 as well. Move to dynamic pricing. So we really got, more widespread usage of these practices around 2011. Uber would say that on big nights like Halloween or, New Year's Eve or I guess, New Year's morning, more typically that, yes, prices go up, but you also need to incentivize drivers to come out at one in the morning. Absolutely. On New Year's Eve. And that part of the purpose, the surge pricing does, one of the things that it accomplishes is it sends a signal to drivers, get out on the road, there's a bunch of business. Absolutely. It brings drivers off the sidelines, like you're sitting on your couch deciding whether or not you're going to call it a night. Yeah, or pick up another ride. Suddenly the price point gets interesting. You jump in your car, right? You put your sweatpants on, you jump in the car, you go drive someone home. Absolutely. Here's the thing, though. Uber's model has shifted dramatically. Okay. And how do we know we have run experiments on Uber Uber Eats. And we've done it on the driver side and we've done it on the rider side. And here's what we find. You put a bunch of people outside of concert. Everyone's leaving at the same time. You have them set the exact same destination. Pickup time is the same, destination is the same. Uber prices vary substantially. Same is true for drivers. More perfect Union did an experiment where they put a bunch of drivers in L.A. in a room. They had them stick their phones on a conference table and all fire up their app and say, I'm ready to start driving. The rides came online. They all picked the same ride. They were given different, price points. And I think what is going on here, quite clearly, is that Uber is getting very good at figuring out the maximum amount each rider will pay for a ride, and they're getting really good at figuring out the minimum amount each driver will accept for a ride. How much of this is also a story of companies just getting bigger? Because one of the reasons that this would be annoying to me, and maybe not to Joe, but you know, there are limited alternatives for transportation. I know there's Lyft and there's taxis and things like that, so you have some options. But if everyone is doing the same thing at the same time, i.e. raising their prices or giving you a price that's higher than everyone else is seeing, then your options are limited. And certainly with something like food, where it feels like every supermarket in the world has been raising prices. And maybe you're in a city where there's only one big supermarket that you can legitimately shop from. You have limited choices, right? How much of this is about scale? I think this is about companies moving into different profit models. So I'll give you two examples. Kroger now doesn't just sell groceries. They sell your data. 30% of Kroger is not operating. Income comes from what they call their alternative profits book of business. Part of what is included in the alternative profits is the sale of your data. 95% of Kroger transactions happen attached to loyalty cards. So they've built really sophisticated dossiers on all of us that are pretty valuable, to external stakeholders, by the way, you know, another thing, and I was just thinking about the $48 cocktail, and I think this is a very important look. You bought the cocktail I did, yeah. Like, this is an interesting, by the way, it was my birthday. So let's just be clear. I'm treating myself to this information important. And happy birthday. Thank you to those yesterday. But thank you everyone. It was Tracy's birthday. Everyone was happy birthday thing. I mean, this is not really a, a, question about, technology pricing. We have a lot of inflation in this country, at least the people on the top side of the coin, like they're not slowing down. Like, that's the thing that's kind of weird. And kind of interesting to me, which is like, I go out to eat and I complain, and it's like, bacon and eggs cause this much. It didn't stop me from going the next day or the next day or the next day. Maybe it's because of, you know, the stock market's going up, which over on feels good about like their investment accounts etc. like people do. People complain and then they just keep doing it. Yeah. I mean, you mentioned this idea of surveillance pricing being used as a sort of Robin Hood in corporate America, right? Yeah. Like you can really soak the rich once you've had a peek into their wallets. And I absolutely think that is possible. And I think companies will do that 100%. I think where I am not so convinced is that that will be used as a cross-subsidy for everyone else. And there are a couple of reasons why. Okay. The first is willingness to pay is not always the same thing as ability to pay. So sometimes you don't have a lot of money, but you're willing to pay a lot because you're over a barrel. You're a mom who told her chat bot that her kid is sick, and now the personal assistant goes out to get the Tylenol. I'll probably pay 15 bucks for the children's Tylenol, right? I want to break the fever. Sometimes you need something urgently. You've got to get home to a funeral. Right? So I think what is most likely is companies will use these technologies to take a lot from the rich. They'll use it to take as much as they can from everyone else, but they'll also be spotting opportunities where our willingness to pay is high, even if our ability to pay is not. Yeah, I'm assuming that protected characteristics apply to shopping as well as things like employment and lending. So the idea that you can't discriminate based on race or age, gender, that sort of thing. However, one thing we've learned from algorithmic pricing is you can you can develop a pretty good proxy for someone's age and gender and race, the zip code of the area that they live in. So you can tell if it's a wealthy neighborhood or a poor neighborhood. Is anyone looking at this on like a federal level, whether people are sort of, I guess, getting close to discriminating based on protected characteristics without explicitly saying that that's what they're doing. Yeah. I don't know of anyone who's looking into it closely right now, but yes, absolutely. What we think of as statistical discrimination, right? You may be discriminating based on something like zip code, but actually the end result is you're discriminating based on race. And the best example of this in the book is, there was a big ProPublica exposé of Princeton Review. And what they found is that Princeton Review was charging individuals more for a test prep package or, you know, so many hours of tutoring, a couple of handbooks or whatever those things are, exercise books. If they logged on from an IP address from, high Asian-American neighborhood where there were a lot of Asians in that region. And the basic idea here was that the concern was that Princeton Review had identified that there were immigrant moms who wanted their kids to go to college, and as a result, would pay another grand for the test prep package. Right. So, absolutely, I think, you know, particularly given that the United States is carved up geographically by race, when companies are using geolocation. And that is a classic, you know, profiling tool that it's in all of these models, right? Your IP address where you, where you come in from, tracking you on your phone to know where you are. That's going to be, a real concern. Let's talk a little bit more about, you know, your work and, in the context of genetic pricing. And I, services I bought whatever going out shopping on our behalf. First of all, if I were to like, I think if, if one were to find an example of an AI but not actually pursuing the best price because it turned out it was either it was not working for me or the company. That put it out since I don't know the name, you know, I was actually gearing towards sponsored. Yeah. I mean, that should be a huge scandal. If if that, is empirically sort of demonstrable, etc. but at least in theory, a company could build a model that actually works on behalf of the customer. And also, at least in theory, a model could build, or not only work on behalf of the customer, but then at the point that the model makes the transaction that the end company does not know who's buying it. Right. The way these things work, typically in terms of the transaction, is that let's say I load my credit card info into, muse or something like that. But then when muse makes the purchase, I believe if I have the strict measures of not using my credit card, muse is generating an instant one time credit card, maybe from a company like stripe. They make the transaction and then muse buys it. And then I'm buying from muse. So at least it's at least it seems theoretically more than theoretically possible, that Agenda commerce could allow people to buy something without that, and seller having any idea of who the actual buyer is. Yeah 100%. I mean, I think there are a huge number of use cases for personal assistants in to commerce that could really benefit consumers and frankly, could get around a lot of the stuff in my book. Right. So we've written a piece at groundwork on the Annoyance economy, the $165 billion a year that, American spend and lost time and money and things like subscription traps. One of the big, yeah, advertising, campaigns of muse is I'm going to crack my knuckles and you'll get to work canceling your subscriptions. And it wasn't surprising to us to see the inertia stocks plummet when the muse numbers started ticking up. Right. Because there's a lot of stuff that people, leave because they forget about it. That muse can go find and handle for you. So, look, I love the idea of folks having help getting out of subscriptions. I love the idea that folks would have help with comparison shopping. Comparison shopping is a total bare, in this digital and e-commerce age, precisely because of the pricing technologies that I've talked about. The problem in my mind is the due diligence. Right? Like you're using an agent, presumably because you're already time strapped and don't have time to go look at all of this yourself. And then you're going to review the output of the agent and be like, well, was this really the cheapest price? How do I know that the agent is actually finding me the best bargain, which Lindsay, by the way, is why you are proposing fiduciary duties for AI agents, right? Yeah. I mean, I don't think this problem is new, right? We often I think with new technologies, I think we need a whole new set of considerations or a whole new set of policies. But this is just the classic agent question who does your agent work for? Yeah, if it's calling itself an agent and particularly if its calling itself your personal agent, it should probably work for you and not muse or whoever is offering sponsored content. And the highest bidder, it should probably not be rooting around in your emails to work against you, in a setting. And so, you know, we have imposed a best interest standard or a fiduciary duty on agents. Throughout time, we try to strengthen those roles under Obama with a stronger fiduciary duty for, financial advisors and people who offer you financial products to make sure that their interests are aligned with yours. Look, I think if these agents aren't going to work in your interest, maybe that's fine. But like, you have to know. Yeah, they're not your agent. They're not your personal assistant. Like, it's basically just Google if they're returning sponsored products. Can you remind me the fight? The fiduciary fight over back in during the Obama administration? Because I remember there was some group that claimed that they should not be, categorized under fiduciary. Can you remind me a little bit of, like, how that fight settled out and who ended up having to have that fiduciary responsibility? Yeah, I mean, it's been a minute. It's just it's like a kind of an oldie. But I actually worked for Senator Warren at the time, and she was the major senator who was helping Obama. And Tom price was running the Labor Department at the time. Make sure they got the fiduciary duty over the finish line. But the basic idea is it's like if you're coming to me and saying, Joe, this here, I want, you know, here are some options for Childs, like 529 plan to save for college. I want to have some confidence that you are not getting a kickback from those three funds to get me into one of them, and that when you present them to me, that actually they're it's not because you're going to get paid an extra bonus if you sell one of them. Yeah, that's exactly what the rule did. Yeah. It was actually really interesting at the time. And this I think is related to our conversation today. It was actually the robo advisors and the new fintech companies who Obama, held up as the white hat companies because they came in and said, yeah, we can we can set up a system that complies with this law, like we can provide, financial advice in the best interests of customers. We can just program it to be such. Right. It was the mom and pop financial advisors from the Charles Schwab in the strip mall who were coming in and saying, wait, like I used to get a trip to a castle in Scotland after I sold $300,000 worth of annuities. Are you telling me that, like, I won't get my Super Bowl style ring, or marry my perks and freebies and junkets, right? I mean, they were much more opposed to it. I think the question with attention Commerce is, could these companies even guarantee that the agents will work in your best interest if they don't actually have full control over what the agents do? The other thing I want to ask you on this is, would you expect, I mean, let's assume the agents do get good at finding bargains to some extent. Would you expect companies like an Amazon or a Walmart to maybe start, I guess, tweaking their APIs or like reconstructing the walls around some of their pricing information such that they try to erode away that that information. I mean, it's already here. The price wars are here or the agent wars, I guess, or here. So we have Amazon announced that it's blocking muse. So muse can't go into Amazon. But Alexa for shopping by the way helps you with Amazon but also leaves Amazon's walls right. So they're not playing by the same rules that they're going to apply to. Yeah. Muse. We know that Walmart is building up sort of separate infrastructure so that you can register your agent basically. And that's crazy. But you would log on and say, hey, I actually am going to let muse, come to Walmart and they're going to set up a system that sort of greenlights that so that they can authorize your agent so they know that this agent is shopping on your behalf. But it raises all sorts of thorny questions about who's going to be holding the bag. If your agent buy something you didn't want it to buy, for example, Walmart is going to say no. But muse is also going to say no. Yeah, so a lot of customers already are experiencing situations where they where someone uses a colloquialism like, go find me a garden hose and the agent goes and buys a garden hose. And you said, fine because you wanted to see some options. Yeah. You know, there's a peerless Heritage Garden hose for $200. Beautiful shoes. Or what happened? What? When you tell your agent, go find me a garden hose and it doesn't buy the garden hose, it hacks into, Home depot.com. Yeah. Messes with their database. And then it's like, who committed the crime? I mean, there's a that's a crime. That's theft. Yeah. And there's get I'm certain this is going to happen. And there are going to be some pretty interesting legal questions about what did you commit the crime because you told your agent to do this or did Metta commit the crime. Like, these are huge questions that we're all going to be facing with. There's there's a story today in Axios about the banks. Oh yeah. So who have said they have started to see highly unusual activity in bank accounts coming from agents and personal assistants. And one of the things that they're tracking is that they feel like a lot of the transactions they're seeing were not directed by the users. And it's a little hard to know exactly how they determine that. But already this idea that you're going to give the keys to your wallet to an agent is, I think, starting to look pretty suspect. I think there's going to be some pretty huge corporate strategy decisions that companies are going to have to make because like, I get why in the first instance, Amazon is like, you know, we don't want your music, accessing our site. But if there's like if, if Walmart is like, bring on your music here, they might not have any choice. It all is going to depend on like if people really want to shop through an AI agent, then like someone is going to really exploit this and say and fling the doors wide open for them. And they, the the Amazons of the world may not have any choice. I don't think it's going to be down to the Amazons of the world. It's going to be down to the insurers and whoever is like insuring liability for the agents and the companies themselves, it will be down to them. But like there is a lot of value for an Amazon in you going there and you them knowing something about you. Right. So you go to Amazon and you buy a toothbrush and maybe you like, click on a movie that you're interested in and blow out. And suddenly they have a rich pool of data about you. Right. And that is very valuable. And if you're going through your moves to buy it, you're still doing the transaction, but there's a lot less that they're getting from you. And this is like going to be a very difficult trade off for them. Like which is the huge it's a huge issue. I'll say a couple of things. One, you should definitely check out your Amazon about me if you haven't already. Oh, I never have. So Amazon now lets you know the characteristics that it knows about you. There was a viral, story a couple of weeks ago because Amazon identified one customer as having flat buttocks based on some of their, I guess, leggings purchases and jeans purchases. So it was quite detailed. It had a lot of information about my three year old, for example, because I had. Yeah. And on some party shopping, Amazon will think that all I'm interested in buying is squishy toys these days. So yeah, the second thing I'll say is we are already seeing anecdotal evidence of folks who used to do quite a bit of shopping on Amazon who are now using use and have bought more things off Amazon than they have in years. And I think this is a real threat to Amazon. And there are some folks at Y Combinator and other places who are suggesting that a genetic commerce could be the first opportunity to start to, like, sure, draw some blood, at Amazon for the first time in years, I want to end on like a happy, happy ish note. Can you tell us about the the hot dog guy who managed to reverse hot dog prices in the 1970s? Yeah. So this was someone who, was a Giants fan, who felt like, you know, the San Francisco Giants were were changing hot dog prices by. I think it was only a nickel. But he felt like this was a period of high inflation. A nickel was was far too much. He sort of ran the numbers, figured out how many nickels they were going to bring in, and figured out how much their costs had gone up. The reason was because they were funding some sort of renovation of the stadium. Right. And they said, like, we have to have higher hot dog prices too. Yeah. I'm forgetting this guy's name now. Is it Alfred Kohn? Is that is that what it was? Ron Gordon runs the San Francisco Giants okay. And Gordon. Yeah, it's funny, you're the only person that has asked me the story on this book tour. Oh, really? Yeah. I mean, there's a lot of happy stories at the end of the book that I tell about. I forgot about this one. I like the hot dog one. Yeah. So Ron Gordon is basically, a hot dog price warrior on behalf of San Francisco Giants fans. And yes, they were renovating the stadium. They were going to start charging more for hot dogs. He felt like it was too much. He started a campaign, to keep hot dog prices down and to make sure that they didn't go up a nickel. He actually ended up getting a letter from the Council of Economic Advisers, like, praising him for his efforts to beat back price hikes during a period of national inflation. This was back during the last period of inflation 50 years ago. And he was successful, in keeping prices low for the rest of us for a while. Anyway, I curious what what hot dogs cost, nowadays at that stadium. But I have one last question. You know, the, I guess the politics of who are the inflation fighters in America? Yeah, they get tricky. We did an episode when we were out in Jackson Hole with the, president of the Richmond Fed, and he was talking to some of the suppliers in his district, and he was like, look, you know, maybe, I don't know, maybe make some lawn chair or something like that. And he's like, our prices are going up, but Walmart is not letting they're like holding the line. And we're not they're not paying more for our lawn chairs. And if we want shelf space at Walmart, we have to eat the price. Yeah. So and so when I hear that story, I'm like, Walmart salute. You are the great. You are our strongest soldiers in our fight against inflation, etc. I also know that a lot of people are like really uncomfortable with, you know, the Walmarts of the world and politically like, would like to see a world of more mom and pop stores, etc. although in many cases they charge more. How do you think, generally speaking, about like sort of the optimal what we want as a society in terms of the optimal mix of margin, large and small companies, because the Walmart for the world, the Amazons of the world, they have incredible power of scale and power of scale. They can also bring prices down, etc. small businesses are sort of famous in many cases for being the worst and things like wage theft and so forth. And often they actually don't have the best prices because they don't have that scale. What do we want as a society in terms of the flora and fauna, the ecology of commercial platforms? I mean, for me, I want the mix that provides the most price competition. Okay. And I think when we don't have price competition in our economy, we're in real trouble. And what I see increasingly is companies building technology is and I talk about a lot of them in the book, to subvert price competition. So even, you know, companies like McDonald's are building sophisticated algorithmic pricing tools to ensure that even franchisees can't compete against each other. Yeah. The whole point of franchises is that they're independent and should have some level of control over pricing. That's increasingly not the case. Now. McDonald's will hem and haw and say, we're just recommending these prices. The franchisees don't have to take them. But it turns out that when McDonald's is evaluating renewal decisions, when when McDonald's is evaluating whether or not you as a franchisee can pick up another property and open a second location, they evaluate things like your conformity to the pricing algorithm. Right. And so my concern here is even in markets where there actually is some competition, these tools are being used to subvert it. So we saw RealPage, another case of algorithmic price fixing where landlords in a city who would usually compete with each other were able to share competitively sensitive information and align on pricing, resulting in rent prices as much as 25% more than you would see in a competitive market through this software. Similarly, there's a gas station technology calibrate, a tech company that, is under scrutiny right now. A class action filed in California. They were helping 1700 gas retailers align on price. The, allegation is that it could have been as much as $0.30 a gallon higher for drivers in California. So I'm looking for price competition that is not always just achieved by volume, right. When these types of technologies are able to circumvent, scale. All right, Lindsay Owens, the book is Gouged. Thank you so much for coming on Odd Lots. Thanks for having me. This was great. Joe, I found all of that really fascinating. I know, you asked a lot of devil's advocate questions there, but in my mind, you said it, not me. In my mind, the issue here is an imbalance, right? If you're a Walmart or an Amazon, you have like the best technology, the best AI models, the best data out there at the touch of your fingertips, versus if you're someone you know, working a 9 to 5 or whatever, who's just trying to buy, I don't know, a car, totally a computer. You don't have that much time or that much technological prowess to actually devote to it. I could, in theory, solve some of that for sure. Yeah. But no, like I'm hopeful that I could solve some of these imbalances. But what I worry about, and not just with pricing, but with a range of fields, cybersecurity very much in the news today. We're recording this October 22nd. The dynamic that I'm worried about is that it's just going to be arms races. Yeah. Right. So you're going to have an agent and maybe it truly does work on your behalf, etc.. And then someone else is going to have an AI pricing model that tries to figure out more about you and overcome that. And then you have to upgrade to the next version of the model to overcome that, and that the only that that in many respects it it's possible that AI makes markets more efficient, it makes things cheaper, etc. it is also possible that it just becomes deadweight loss where we we arrive at the exact same price that we would have before. Except the only difference is that both sides have paid more to open AI, anthropic or metaphor in order to re arrive at the same equilibrium that we could have had without all this. I. This is my busywork theory of AI, where everyone's just going to have agents that like file lawsuits and we get to the same place complaints. But yeah, someone makes a lot of money selling tokens to us. Also, I thought that has the history behind like the single public price tag is really interesting because if you think about it to some extent with a genetic AI and all these new pricing strategies coming out of companies, we're kind of going back to that customized era of haggling where like, you have to avoid haggling, you have to devote a bunch of time to, like figuring out the price. If you're buying from McDonald's and you want a good price, you have to, like, figure out which coupon you want for that day and what's available to you. It's a time suck, right where it is a time suck. You know? It's not efficient. You know, here's a thought that I had. Remember in 2021, 2022, 23, there was this very recurring thing that you see online. It's like someone saying, oh my God, I just paid $15 for blueberries, okay? And then someone else just like, would say, you're an idiot, go to Walmart.com. Here's blueberries that you can buy for 299. Yeah. And if you're paying $15 it's really your fault. And it kind of is, but not the. But when I see these, you could have bought blueberries for 299 at Walmart. What I see is a lack of general equilibrium thinking, by which I mean, if all of those people who shopped at Whole Foods suddenly switched to the 299 Walmart blueberries, they would now be 2.99 anymore, right? Right. There would be this big inrush. Maybe they wouldn't jump to $15, maybe they jump to 499, etc. but like prices our prices. And so just because there exists a lower price does not mean that everyone can capture the lower price, because then the lower price goes up. And so like, I don't know, I would also argue that buying blueberries for a decent price should not consume that much mental capacity. No, it should and shouldn't be like a five minute campaign even to go online and research like, oh, I'm going to buy it from here and I'm going to have it. You know, if you buy it online, you're going to have to have it delivered to your house as well versus just walking in a store. It's a time issue and insufficiency and inefficiency issue in my mind. And I say that as someone who loves shopping and buying $48 cocktails, apparently. Yeah. Well, this is the last point I was going to make. And this is a macro question, and it has nothing to do with algorithmic pricing, which is that there is a those of us on the top of the K complain about prices all the time. And for whatever reason, by and large, some of us but most of us have not meaningfully changed our consumption behavior despite all the prices that we complain about. Well, I'm really getting grilled for buying a cocktail on my burger. No, it's all it's this is my whining present. This is people complain and then they do the same thing they do the next day. And that is why inflation continues to persist, because it hit people behaviorally. Whatever it else, people complain and then they buy the same thing the next day. Yeah, it's true, but also it's not like if you want to drink something in a theater, you don't have one. It's not like I can go out somewhere else. But anyway, shall we leave it there? Let's leave it there. This has been another episode of the Odd Lots podcast. I'm Tracy Alloway, you can follow me @tracyalloway. And I'm Joe Weisenthal. You can follow me @thestalwart. Follow our guest Lindsay Owens. She's @owenslindsay1. Follow our producers Carmen Rodriguez @carmenarmen, Dashiell Bennett @dashbot, Cale Brooks @calebrooks and Kevin Lozano @kevlloydlozano. And of course, you can chat about all of these topics 24/7 in our discord, discord.gg/oddlots. And for more Odd Lots content, you can check out our daily newsletter, which you can find at bloomberg.com/oddlots. And if you enjoyed this conversation, if you like it when we talk about pricing, then please leave us a positive review on your favorite podcast platform. And if you're watching this on video and you enjoyed it, then please leave a comment or like the video. Or better yet, subscribe. And remember, if you're a Bloomberg subscriber, you can listen to all of our episodes absolutely ad free. All you need to do is find the Bloomberg Channel on Apple Podcasts and follow the instructions there. Thanks for joining.

Assistir ao vídeo original no YouTube

Assuntos: Agentes de IA, Muse. Publicado no YouTube em 2026-10-26.

Seleção automática por assunto e idioma. O vídeo pertence ao canal de origem.