Loading summary
Transcript68 lines
- [00:00]
A
Hello, everybody. This is Marshall Po. I'm the founder and editor of the New Books Network. And if you're listening to this, you know that the NBN is the largest academic podcast network in the world. We reach a worldwide audience of 2 million people. You may have a podcast or you may be thinking about starting a podcast. As you probably know, there are challenges basically of two kinds. One is technical. There are things you have to know in order to get your podcast produced and distributed. And the second is, and this is the biggest problem, you need to get an audience. Building an audience in podcasting is the hardest thing to do today. With this in mind, we at the NBM have started a service called NBN Productions. What we do is help you create a podcast, produce your podcast, distribute your podcast, and we host your podcast. Most importantly, what we do is we distribute your podcast to the NBN audience. We've done this many times with many academic podcasts, and we would like to help you. If you would be interested in talking to us about how we can help you with your podcast, please contact us. Just go to the front page of the New Books Network and you will see a link to NBN Productions. Click that, fill out the form, and we can talk. Welcome to the New Books Network.
- [01:07]
B
Welcome to the New Books Network. I'm Alfred Marcus, and this is on the Cusp, where we explore how strategy and ethics intersect in shaping organizations and social change. Today I'm speaking with aijah Lipaunan, professor at Cornell University and a scholar of innovation, digital strategy, and the economics of technology. Aijah is the author of Digital Innovation Strategy. This book examines what is special about digital business innovation, why it is so generative, why uncertainty is so pervasive, and why competition often turns on platforms, networks, and ecosystems. We'll discuss how digital innovations reshape competitive advantage, how firms manage technical and market uncertainty, and what ethical responsibilities arise when data networks and algorithmic decisions increasingly govern economic life. So let's start with your background and motivation, aija. And I hope I'm pronouncing your name correctly.
- [02:21]
C
Am I close enough? It's Aya.
- [02:23]
B
Aya.
- [02:24]
C
Okay.
- [02:25]
B
Could you begin by explaining the motivations for this book? Digital transformation is everywhere, but it is also unevenly distributed. What problem did you see among managers, students, or scholars that led you to write a systematic strategy guide focused on digital innovation?
- [02:45]
C
Yeah, so I kind of focus this book on innovation strategy. So I've been teaching innovation for 25 years or so, and I'm often surprised and kind of. I've learned over the years that people who work with technology often don't have a good grasp of how to think about users and the market where they're aiming their innovation. And so I wanted to kind of create a book about innovation strategy that helps them think through the critical steps when they're sort of working on the technology, but also trying to figure out what, what the market needs, what the users are excited about. So it's kind of a strategy framework for developing technologies that are actually valued by users and not immediately copied by competitors. And then so that's the innovation. Like generally that's what every innovator needs. But focusing on digital markets particularly, there's been a lot of really good, valuable, helpful research in strategy and economics in the past 20 years or so. But there's no good textbook or good kind of manual for innovators that really kind of bring it together into one practical framework that allows people who are, who are creating stuff to kind of apply easily and think through their strategic challenges. So that's what I thought was an opportunity for me, since I've been teaching digital business and digital strategy for quite some time, to kind of bring that together and try to sort of create an umbrella framework that helps practitioners, people who are actually innovating new stuff in digital markets, to be more thoughtful about how to approach users and markets.
- [04:39]
B
Are you an economist originally by training?
- [04:42]
C
I am an economist by training, but I've been teaching strategy the whole time. So I'm kind of a strategy professor with a background in economics.
- [04:54]
B
The book is a great manual. My colleague, who, who I really. Who's no longer with us, Andy Van Devan, you must know about some of his work and his office was next to mine for many, many years. Early in the book, you distinguish between firms that merely use digital tools and firms whose core value creation is digital information products, communication services, platforms and data driven systems. How does that distinction change the strategic questions leaders should ask? What does it imply about where value and competitive advantage are likely to come from?
- [05:34]
C
So that's a great question because I think the answer is sort of seems obvious once you hear it. So one distinction between physical production and digital production. So these firms that just use digital tools for physical products and then firms that are really kind of core, the core products, services and activities are digital. So one critical distinction is the cost structure. And this is where you see my economist stripes, maybe. So it's actually very strategic how what activities are costly and how they're costly for these different types of firms. So if you're a physical producer of let's say cars. So we're very. So economics, microeconomics 101 is all over kind of how that production function comes about and what it looks like. So you'll have to pay. It's very expensive to start a factory for cars. There's a large fixed cost and then there's variable costs for materials and employee wages and energy and all kinds of things that are also quite substantial. It's expensive to make cars. There are some economies of scale, but you can only kind of. You have to figure out what is the optimal number of cars that you want to produce within that one factory. And then trying to produce many, many more cars than your optimal number is going to just make things worse. But if you think of. So that's a physical production system, but if you think of digital production, it's actually quite different. So let's think about news. It's also a lot of work to create valuable news that people really want to learn about. So there's a high fixed cost there, but there's almost no variable cost. So if you're a news producer, let's say that you're operating digitally and you have a website and you're just kind of writing up news stories and then posting them on the website, and then millions of people can read it practically with no additional cost to that business. So there's huge economies of scale. And so the company will become more and more profitable the larger it grows. So it has an incentive to really become as large as it possibly can. And it will compete, and it should compete very hard to take over as much market share as possible and push out competitors. So that competition for market share is going to be really, really tough and challenging. So these different types of production systems are associated with different cost structures. And then as a result, the competitive dynamics in the marketplace are going to be quite different. And so you have to be aware of what bucket of operations you're in in order to understand how the competition will play out. And there's the book goes to great detail about how that. What is different about competition and what you have to look out for. But if you don't understand the distinction then, and you're in the digital marketplace, then you're not likely to be very successful.
- [08:51]
B
Software companies like Google and Oracle would be in the latter category, while Tesla and Amazon may be in between to some degree, but probably in the former category because there's a lot of physical assets involved in their production processes.
- [09:08]
C
Exactly, exactly. So Tesla is an interesting example because they do have a lot of digital assets, maybe their greatest distinction and differentiation is in digital because they try to create self driving cars, but there's still that physical kind of weight of the physical operations, and it's really difficult to make that very efficient.
- [09:29]
B
And nearly every firm today is digital in some form or another. But for most firms, it's a tool, it's not their primary product. You emphasize uncertainty, technical uncertainty, market uncertainty, and behavioral uncertainty in social systems. Could you unpack those three uncertainties with examples? Which type do you think is most underestimated by executives and why?
- [09:56]
C
Yeah, so I would say that all three forms of uncertainty tend to be underestimated by innovators, but behavioral uncertainty is the most tricky to deal with and even perceive. So let's start from technical uncertainty. And by that I mean just the kind of uncertainty about can we even create this product that we're thinking about. So for example, at this time we don't really know whether we can actually create artificial general intelligence. We have forms of AI, but not the AGI. Yes. And so we don't know. There's a lot of companies that are trying to create it. They're investing a lot of money, but we don't know exactly what it takes. So there's a lot of technical uncertainty about that kind of artificial general intelligence. But in a lot of software, software products, and digital products that are software based, technical uncertainty is actually fairly sort of modest or lower. If you want to create an additional feature for your users to do an additional transaction or whatever it is with software, it's actually usually doable. It can be clumsy and you can do it well or poorly or well. But if you have enough code and data, you can create almost any feature. So technical uncertainty tends to be, I think, lower in a lot of digital companies. Market uncertainty, on the other hand, becomes much more of a challenge for all digital companies or companies operating in digital markets. So what is market uncertainty? It's more about kind of the reception, the market reception for your product. So you launch some digital service and you don't really know before you launch it, you don't know for sure if your users will like it, will they actually use it, how much are they willing to pay for it and all that. And when the product is based on software, so it's not very tangible, they can't really try it on in the store or really experience it very well, and it's a new kind of a service, then it's really hard for potential users to imagine how or why they would use it. So they tend to not want it and not express much willingness to pay for it. So that's where we kind of, and not, not me, but the strategy industry and strategic management industry, the whole field of, of academics has, has come up with a lot of kind of practices and techniques like design thinking and early prototyping and, and other kinds of practices that, that we tend to think of kind of falling under lean startup. And so when you follow those practices, you can at least to some degree or significant degree, usually lower market uncertainty, at least enough to decide whether the product or service is sort of likely to be successful or not. But then there's this behavioral uncertainty and that's really the trickiest part. And by that I mean uncertainty about how will users behave when they use your innovation or the new product. And so a lot of these, a lot of digital services entail to greater or lesser degree interaction among users. So there's some communication, some social interaction that users do while using the service. So if you think of social networks, for example, it's very central that, that users interact with each other. And when there's some communication and interaction, the innovating company doesn't really know how that will evolve. So let's think about, here's an example. Email. So everyone knows email by now. It was actually invented in, in the 60s and early 70s to send really kind of technical messages within the, the arpane, one of the kind of predecessors of the open Internet, so intended for technical communication between different universities and the government. But it became a super popular application and people started to send all kinds of messages, communications. And by now I recently heard that the state of Denmark is considering stopping delivery of letters for individuals. So we're almost out of orders, we're just using email. So that's how massive this innovation was. But initially when the, when the idea of email was invented, those inventors had no idea what would happen when people started really using it and sending both not just professional emails, but also personal emails. And there were mailing lists and, and a lot of kind of individual expression that started to bubble up and people formed communities and all kinds of kind of social, sort of what sociologists call meaning creation. So people started using it in ways that they could not anticipate at all. And so if you were the inventor of email, first of all, you had no idea. But second, is there anything that you would have done differently if you had known anything about that? Probably they might have facilitated some of the uses that ended up being really kind of valuable or popular. And so if you understand that your Digital innovation might be associated with this kind of behavioral uncertainty and then this process of social construction that follows where people start to use it and kind of interpret it in various different ways. Then the innovator can facilitate that process, can try to maybe steer it in some ways or at least learn from it quickly, and then adapt the product or service in a way that sort of helps and supports the user's interacting with innovation and using it in more various ways that are more valuable. And so being aware of that process possibly happening after launch is really valuable.
- [16:39]
B
Do you think that or to what extent will AI reduce uncertainty?
- [16:48]
C
Yeah, that's a great question. So I'm hearing that people are using generative AI to do market research and I'm very skeptical of this idea, but I think it might be a way to aggregate sort of diverse perspectives very quickly. So it might accelerate some information collection processes. But I don't know.
- [17:16]
B
I don't know if this is a question you can answer either. But like today, if we think about, let's say world GDP or GDP in developed countries, how much of that GDP is being generated by companies that are purely digital as opposed to companies that use digital? You know, we may not know that number, but do you have any way of estimating that or your sense for that?
- [17:39]
C
I don't know about gdp, I have to say. But if we look at the value, stock market value of companies, you saw maybe some of the tables in the book where a lot of the stock market value is concentrated in digital platforms. We know that, that there's the value creation, a lot of that is in digital companies. But out of global gdp it has to be a faction.
- [18:03]
B
Right. But it's, it's, it's, it's enormous. If you look at, definitely if you look at the stock market value, the so called seven leading stock market companies. Other thing about digital in terms of how they market to consumers is that they often offer you something free as bait and then get you involved.
- [18:26]
C
Yeah, they try to lock you in. We're starting to talk about networks and that's where it all matters.
- [18:32]
B
So a major part of digital strategy revolves around networks and platform dynamics, network effects reaching critical mass, switching costs, and the tension between openness and control. For listeners who aren't economists, what is the simplest way to understand network effects and tipping dynamics and why do they create such high stakes competitive battles?
- [18:58]
C
So we could continue on that email example because it's such a simple technology, but it's so valuable to. And so if I don't know anyone else using Email, it's not a valuable technology for me and I'm not going to, you know, download an email app and I'm not going to get engaged with that. But then once you're a lot of your friends and colleagues and family members, everyone uses email, then it's really costly not to join the network. So then it's very compelling to join a new network. And so when I teach network effects in my class, I often ask the students to think about the last time they joined a new network. And so this last semester, this fall, somebody mentioned the news newer. It's not entirely new social media site or social network called Bereal. And so when I ask people why did you join that? And they start talking about, well, I had friends or my sister or someone was using it and told, told me that it's super fun, so I joined it. And so then you start to really kind of noticing how the users, the existing users in a network draw you in and make it valuable and make it compelling. And so the more users are there, the more valuable it becomes overall. And, and so then we have Facebook that has how many billions of users by now? And it's an unbelievable social graph and kind of collection of connections globally between people. And so once you have something like that that's created, it's really hard to start a new Facebook. There are in different regions of the world, there are some other social networks that are more, more dominant. But, but if you think of your own Facebook or Be Real or whatever social network application you enjoy, if a similar application shows up and tries to start kind of luring people away and to switch to the new network, it will be really hard to, if you, even if you wanted to switch, it would be really hard to get all your Facebook friends, all your Bereal connections to switch to the new network. So there's, there's, there's a huge cost of switching when you have to coordinate the switching with your friends. And so you end up being locked into that to usually to the larger network, you're just kind of stuck with your network because it's, it sort of works and you have a lot of friends there.
- [21:37]
B
It's very hard to switch.
- [21:40]
C
The switching becomes just really cumbersome.
- [21:42]
B
That's sort of why Twitter or Eve is, what's it called now, X. Why? Why even though many people wanted to switch, they didn't switch because it had this captured. It already captured such a large network of people.
- [21:59]
C
Yeah, but then sometimes, occasionally something comes along that offers something unique or different feature. And so Twitter is a great example of that. So when Elon Musk took it over, it became X. And one thing that they did was that they removed content moderation pretty much on X. And so a lot of people were really, just really did not enjoy the network, enjoy the content they were seeing after that change. And so despite that, despite the network effects, a lot of. So there's millions of users who switched to a new network called Blue sky that started from scratch just a couple of years earlier. And so if there's, if there's a new feature that's really attractive, even if you have a large network, like a pre existing large network, then sometimes people do kind of take that, eat that switching cost and start a new when, when they really enjoy a new feature or a new aspect of the, of an innovation.
- [23:05]
B
Complementers Many platforms depend on complementers. Developers, content creators, sellers, partners. Strategically, platform leaders want to attract complements while still capturing value ethically they also wield extraordinary power over the livelihoods of the people who are involved in the complement look. Governance principles help platforms strike a balance between healthy ecosystem growth and fair treatment of complementers.
- [23:35]
C
So I'm a kind of a cynical person and I don't think when we look at these centralized platforms, Facebooks and Twitters and X's and Googles and all kinds of platform owners didn't really think about fairness. It's not one of the kind of leading lights of their strategic planning. So economists, which is sort of the dismal as a dismal science, they have this concept of individual rationality constraint. So what does it mean? It means that people will participate, a user, let's say a complimentary, like a game developer who develops for a mobile platform like Google or Apple. So they'll continue to kind of offer their game and keep it on Google or Apple as long as they benefit more than it costs. So the rationality constraint is just that you won't continue to do things that over time that are more costly than what you benefit. So if you have platforms, so Apple and Google, they host the game, but there are probably thousands of game developers there and most of them are small. They're just trying to kind of break through and they have one or a couple of games and they try to attract users and make some little bit of money there. So the platform has all the power and the developer has, their only power is to either leave or stay. They don't really negotiate, they can negotiate hard against Apple and Google. And so what happens is that the platform will take a large chunk of their Profits, they will charge fees for the developer to host their games and the developer gets tiny benefits. And so if they can somehow gain more revenue than the, than the access fees, they'll keep their game there. So the game developer might be better off participating in that platform. But it's not the fairness. There's no fairness in terms of dividing the benefits equally or, or being somehow getting a fair share of that.
- [25:49]
B
There has been a lot of friction. Spotify, Apple has had a lot of friction. Amazon with its companies that sell off of Amazon, I mean, even the New Books Network depends upon people like me who are hosts and complimenters essentially to this platform. That friction, I think, is very hard to resolve, especially when the complementaries really depend upon like the people who sell their goods on Amazon. They really depend upon Amazon for their livelihood. So that the government has become involved in some of these controversies in court cases.
- [26:36]
C
That's right. So there's really two solutions to that friction. There's regulation and there's litigation. The platforms and complementers by themselves will not be able to resolve it. The platforms will kind of try to eat up all the profits if they can. But there's Spotify actually participated in the Epic Games litigation, where Epic Games is not a small developer, is a major developer. So they had enough money to take Apple and Google to court and argue that they had monopolized the mobile app market and they managed to get some concessions out of those companies through litigation. And then regulation. The EU has been more active on this front, but there's that currently in the US there's an antitrust case against Amazon about exploiting these complementers, the sellers in the marketplace, until I think that's going to go to trial this year, 2026. So we'll see if regulators will actually try to address that. That sort of fairness versus exploitation versus platform power.
- [27:50]
B
Any different in the EU or in Europe, have there been cases or have they tried to regulate this in a different way?
- [27:59]
C
It's not different, they're just more active. I mean, a lot of US companies claim that EU does it because there are no big platforms in Europe. So there's that. But there's also a different philosophy about antitrust in Europe. So fairness is actually one argument. Antitrust and regulatory arguments. So they do consider. Regulators consider fairness in Europe, I think more actively than US regulators do. So they, they are more kind of, I guess, more. It's easier to trigger regulatory antitrust scrutiny there than it is in the us.
- [28:39]
B
They've been more proactive. So another thing is like digital Innovation increasingly rests on data. Data assets, analytics, algorithmic decision systems. Council leaders think about data as a strategic resource, building it, protecting it and commercializing it, exploiting it without crossing ethical lines on privacy cons or the like. It's a, it's a big question that we all face because we know that we're getting free use. But, but in exchange the, the companies are, are, are, are are pinpoint pinpointing us as who, who we are and, and, and providing that oftentimes providing that information to other companies.
- [29:25]
C
Yeah. Consumers are so helpless to a large degree. And why is that? It's because these software based systems are really opaque. So consumer has no clue. There's no way for me as a consumer to really understand how Google or any other company handles my data. And I will only learn about something about that, kind of get a vignette into that when there's some kind of a leak or a breach. And there are a lot of breaches in the United States all the time. I think there's thousands every year. So I recently saw a statistic that the majority of US population who, who are online at all have their data compromised within a single year. So there's just a lot of leaking. But so that's kind of about data security. But at when this data breaches or other kinds of leaks that also can be really, really devastating for individuals if their private data is exposed. And so we've had these breaches of dating sites. There's been a couple that have been really just personally damaging to a large number of people when their kind of personal dating, dating data has been exposed and sensitive information about individual has been, has become sort of public domain information. And so there's not a lot that consumers individually can do about that. In the United States we do have some regulations of California has regulation, there's a few other states, Europe has GDPR that's trying to limit these kinds of or kind of give clear governance practices to companies so that they keep their data secure. But the other thing that GDPR does regulate and that we're lacking here is that kind of sort of data, personal data markets. And so as you said, so for example these dating, dating breaches. So we know that dating apps package their data and then sell it to data brokers and data brokers, who knows what they do with the data. And so GDPR in Europe prevents that kind of after sort of aftermarket for your data when you have kind of signed a term, terms and conditions where you say that you don't want your data to be sold if you're data.
- [32:07]
B
So you note that information and communication technologies shape the sequence of communication, information, decision and action. As more decisions are automated, what ethical risks concern you the most? Errors, its scale, opacity, discrimination, manipulation. And what strategic moves can firms make to build trust as a competitive asset?
- [32:32]
C
So I think the kind of biases and discrimination are, if not the most concerning, at least they are highly concerning. And so when we have companies that collect a lot of data and then they use it through kind of predictive analytics to make decisions or at least make recommendations for decisions and action, a lot of. So most of those systems at this point are based on kind of correlational data analytics. So it's not causal inference, meaning that they don't. They don't use the data to figure out what causes some outcome. They just look at the correlations. They use massive amounts of data, but it's still correlational. And so it can be really misleading if you don't have representative data about a population, for example. And so we have examples of medical research and medical kind of health outcomes where if we don't have good data about some subpopulations that actually have different kind of medical features or health issues, then the treatments are not going to be. There's mismatch matching of treatments, and then the outcomes are poorer than they could be. And so it's really important to have kind of representative data and do this kind of data analytics carefully and thoughtfully. But that's really kind of unintentional sort of discrimination that might happen if the data is misleading or not representative in ways that we don't understand. But database systems can also be used kind of intentionally to discriminate. And so obviously you need someone with really bad will because these systems tend to be really opaque and they're fairly easy to manipulate or exclude or ignore groups or aspects of people for private benefit. So there's another risk for just bad behavior in database systems. So two ways that things can go wrong when you automate decisions, like in.
- [34:59]
B
Hiring, there's this accumulation of data. Are companies actually accessing it when they make hiring decisions? Like if somebody puts their. If somebody has a long train of Facebook comments, is a potential employer likely to access that and use it in making their decisions?
- [35:25]
C
I've heard of, at least anecdotally, that employers do look at social media when they hire young people. That's really concerning. Young people can do all kinds of wild things and still be good people.
- [35:39]
B
Right. We all were at one time young.
- [35:42]
C
Exactly. But we were not online. So Thankful for that. My dad taught me a lot, including how easy it is to forget to cancel things. So I downloaded Experian, my bff. Big financial friend Experian could help me cancel my unused subscriptions and lower my bills, saving me hundreds a year. Get started with the Experian app today. Your big financial friends here to help you save smarter. Results will vary. Not all bills or subscriptions eligible Savings not guaranteed $631 a year average savings with one negotiations and one plus cancellations paid. Membership with connected payment account required. See experian.com for details.
- [36:14]
B
Experian Incumbents often struggle in digital disruption. You talked about that a little before. Especially when new performance trajectories emerge and their existing business models lock them into old assumptions. What are the commonly most made mistakes by incumbents when they attempt digital transformation? Or their particular organizational designs, metrics or leadership practices that help them innovate without losing their core? The basic strategy question really this is about explore versus Exploit.
- [36:52]
C
Yeah, there's that and the whole kind of how to deal with disruption question. And so we do have a lot of examples of industries like whole industries like newspapers and music labels that could not figure or took it took a long time for them to figure out how to, how to deal with a digital way of doing things. I mean they were challenged by piracy and losing control of their their intellectual property, but also just kind of stubborn inability to look at their own operations and figure out what parts of it could be done in a simpler way. So one really cool industry where disruption has actually happened multiple times is mapping. And here's a story. So before computing, before the Internet, mapping was done through kind of aerial photography and all kinds of kind of ground level surveys. And then in the early 2000s, companies like NavTech and also Google started to create digital street maps by driving around cities. And so huge investment in that kind of digitization of physical space. Super expensive. So what we ended up with was just there was just a handful of companies, couple of companies that had mapping data and then everyone else was licensing from those companies. And of course the companies, navtec for example, made a ton of money on their, on their maps. Just a few few years later, 2008, a company called Waze was founded. So this is an Israeli company. They figured out a much simpler way to create maps. What they did was create an app that users could could use to do routing when driving around or figure out where there are maybe speed traps or congestion and so on. But also while they were using the app or the app was on the phone, it would be sending their mobility data back to Waze. And so they would then basically use these users devices to map the world rather than sending cars around to drive those roads for them. And so a fraction, a tiny fraction of the cost to create the map. And so hugely disrupted. So what ended up happening is that Google having driven all the roads and mapped everything already once, they actually bought Waze in 2013 and then integrated it into Google Maps. So much of Google Maps is based on Google on Waze mapping now. But Google was, was pretty quick and smart to realize that there's a better way to collect the mapping information. And so within just a kind of a few years they, they changed their operations in Google Maps and adopted this new way of using networks, user networks to crowdsource the mapping data. So they had to kind of put aside their huge investment into creating their own maps when they realized that you didn't do it on the cheap using users devices to collect data. And so it often requires that willingness to put aside old inefficient ways of doing things and streamline your operations with digital tools and networks acquisition.
- [40:27]
B
What about a company like the New York Times, Wall Street Journal, they seem to have made this adaptation to a digital world. In fact, I think almost, you know, like a hundred times more of their subscriptions now are digitally based as opposed to getting physical newspapers. How did they manage this?
- [40:50]
C
Yeah, well it, I actually don't know much about Wall street journals. I haven't looked into that. But I teach New York Times. I wrote a case about New York Times when I teach it. And so I've looked at their data every year since several years ago and they almost didn't make it. They have a very kind of a. They were in the stubborn category. So they definitely thought that they were in the business of writing news and printing it on paper and then delivering the paper. So they continued to do that. They still do that to some degree, but it's a very costly activity. And they, their perception of what New York Times is about was based on that physical paper. They thought the website is just to advertise to get people to buy the physical paper for years. And then finally they, they, I think it was around 2012 they started experimenting with well, we could actually create subscriptions that people like, create a first create a paywall, then start be smarter about or develop ways to be smarter about pricing for access. And then they created a much better system of price discrimination which sounds bad, but it's the only way to make money with information products. So they started figuring out who is willing to pay what kinds of subscriptions and what kind of content and how can we make this work. So they were also lucky in that 2016 was a year when people, for political reasons people started to pay for newspaper subscriptions. So that also helped their recovery. But it took a couple of decades for New York Times to figure this out. So it's a long time to long time to work through.
- [42:40]
B
What about if there's a manager listening to this podcast and they want to improve their digital innovation strategy in the next 90 days, what is a practical experiment or a diagnostic you would recommend? And how can they reduce uncertainty through experimentation without drifting into activity for its own sake?
- [43:01]
C
So one idea I recommend trying is the idea of digital twin. So a few years ago I worked on a research project and research paper where we thought about the process of creating an organizational digital twin. So a lot of. So even if. So if you go to these companies that are still in physical production mode and not fully digital, a lot of companies still keep their data kind of operational and market data in silos, in kind of departmental or unit or functional sub organizations and kind of separate from one one another. But making the most value out of data requires that you kind of integrate it or you connect a lot of data, even data sources that seem to be completely disparate. But when you bring them together and you use machine learning techniques to analyze the data, you can generate, sometimes generate insights that are super valuable. So I recommend putting organizational data into a secure environment, give people access to agentic generative AI and then they can really quickly using the organizational, the proprietary data, they can start running low cost experiments on all kinds of innovation ideas. So one benefit of generative AI is that it can allow, let's say a business person, it can allow them to do vibe coding fairly easily or technical person could access some market analytics fairly easily. So alternative AI can help a lot of employees to kind of branch a little bit outside of their area of expertise. And if they get access to data and a lot of data, then they can quickly really kind of leapfrog a lot of sort of analytics and R and D cost by doing sort of quick early stage prototyping based on the data and the AI tools that we now have available for everyone. So that's what I recommend. Yeah.
- [45:11]
B
So AI is really the next stage in digitization or do you see the two being different in some kind of essential but fundamental way? Or is it just the next stage which do, sorry, AI being The next stage in digitization, or does it make digitization much more powerful than it was previously through its combination with digitization? Digitization makes AI possible without.
- [45:40]
C
Yes. So to benefit from generative AI tools, you do need to be fairly highly digitized as an operation, an organization. And so the data, the organizational data helps to get a lot of value out of generative AI tools. So I do think that it will be when, when a lot of people have access to agentic AI, I think it, it does change. It can be a game changer. If people know what to do with those techniques, what would you.
- [46:12]
B
What do they need to know to do what.
- [46:15]
C
So this is. We have to do another podcast about this. But they understand that how to use the AI as a complement to their own kind of analytical effort and judgment about what is valuable, what is interesting, what is important, what should be, how should we think about the world, how should we think about users? So there's that human and AI complementarity, humans. We cannot outsource our analysis or thinking or innovation to AI that does not work. We've already kind of establish that through research and just companies trying out things. So we need to figure out how to bring AI into our creative activities. So it's a kind of a. It's a learning process. Not everyone does it intuitively. Right. But I think after people kind of get the hang of it and understand the potential, it does accelerate cognitive effort and cognitive activities and creative activities.
- [47:27]
B
Yeah, I really agree. It's the art of asking AI the right questions in the right sequence and not letting it get away with sometimes the weak answers it gives and challenging it today.
- [47:41]
C
Exactly.
- [47:42]
B
You have to be very hypercritical, I think, in your use of AI, or else it's. And the idea that it will make decisions for us is dangerous because without that hypercriticism.
- [47:55]
C
Yes. Journey of AI does not optimize anything for us. It will just create answers that based on its analysis, it believes is the most appropriate or most popular. That's how those systems are created. And so if we go with those answers, then we're just going to be completely generic or just kind of going with the mainstream or going with the flow.
- [48:22]
B
Lukewarm mediocrity. That's what our society will have more of. What about, I mean, besides AI, or maybe within the realm of AI, what really excites you today about platforms and the digitization and where it's going and its promise and its pitfalls, whether they be ethical or even just social in, in their, in their broader. In the broader sense.
- [48:54]
C
I have to say that I, I have spent the past half a year thinking a lot about generative AI because it's just 2025 was a real breakthrough year in, in so many ways in that regard. And so I, I had, we had so many conversations about it in my classes this last semester and I am excited. But I'm also super worried about what will happen with this. And I don't think we can let. So going back to the platform power and regulation and all that, I don't think we should let as a society let generative AI and AI systems more broadly be kind of monopolized by one or just a couple of platforms or systems. I think we have to make sure that these systems become. That we know what's going on with the system, we know what data they use to train them and we have to know a lot about the systems in order to be able to trust them, but also to not allow anyone monopolize such a hugely valuable technology. And I'm hoping so for the future. I'm hoping that these systems will eventually become open source. And I think it's possible, but it's not on the horizon yet. But I think one scenario is that there's a sort of a disruption of foundation models for generative AI, whereby the foundation models become open source, kind of collaborative structures, and then people can build commercial applications on top of, and then make it profitable and economically viable. But the foundation models I hope will be open source.
- [50:44]
B
Is AI open source as opposed to the other companies? And isn't there also? I mean, one thing I'm sort of amazed at is there is a lot of competition in AI right now. There are a lot of different companies offering AI and their AI systems are, are quite different in the results. You ask the same question and when you get answers for questions, they'll remember your prior interactions so that it personalizes the question. And so each system that you work on will remember your prior interactions and will give you a very unique answer. In that sense, I think it's very different than software because software is very predictable and you put the inputs, the outputs, you get. AI is. People sometimes claim it's not creative, but in some sense it's very creative.
- [51:38]
C
That's very generative.
- [51:40]
B
Yes, it's generative. I guess that's. Yeah, it's the way to put it. So we should bring this to an end. It's been a great discussion, but what are you working on right now? You sort of alluded to it a bit, but can you explain what are the key things that you think are important to work on right now and what are you working on?
- [52:01]
C
Well, one of my projects is focusing on the data markets for data for generative AI. So we're trying to understand and kind of piece together what data are used to train generative AI. So large language models, and there's a whole sort of legal context in which copyright for Genesee so large language models is kind of playing out currently. So what can they use? What is fair use in training language models and when do they have to license and what is the data that they're licensing? How much are they paying? So we're trying to sort out what's going on on the data side of large language models. But in addition to that, I'm still working on, on standard setting and intellectual property questions in communications industries more broadly. So it's a gift that keeps on giving because there's always new networks that.
- [52:59]
B
Come so important right now, the data providers ultimately are the ultimate complementers. Without the data providers, none of these systems could be built. The other thing that I've been thinking a lot about and I really don't understand, although I've been reading various articles about it, that is the alliance relationships, like let's say between Oracle, which is building the big data centers for OpenAI and one's. One company's success depends upon the other companies. There's a lot, there's a lot of this intertwining, you know, the relationship, let's say, between Google and OpenAI, you know, they're in competition yet they're, they're working with each other. Have you thought a little bit about that or do you have anything to say about that?
- [53:51]
C
I have not really researched that. I have observed the same thing that I don't quite understand. The networks, those alliances that are happening and some of them being overlapping. And so it's really intriguing. I haven't looked into that, but I think there's a. There's an so one element to that is that the technology is early stage and it's not. None of the companies know yet what's going to be the kind of the dominant design in a sense, in our sort of traditional innovation research literature. And so they're hedging a little bit. They have kind of their hands in all the pies a little bit and trying to figure out how they can be part of the evolution of the industry without committing completely to just one model or one set of partners.
- [54:41]
B
I heard a podcast with Sam Altman and the one observation I had after the podcast was he was basically saying, let it all fly, let it happen. I don't know what's going to happen in the end and how this is going to work. I was very surprised that he was so open ended about the possibilities and I guess that's where we are, so.
- [55:07]
C
Yeah, yeah, I think that's right. I mean, looking at how much money OpenAI is still taking in and how little revenue they make. So they are early stage. They don't yet know how they can commercially make this work.
- [55:22]
B
Is there anything else you'd like to add to our conversation?
- [55:26]
C
No. Thank you. Thanks for the opportunity to talk about these issues and talk about the book. Great questions.
- [55:33]
B
Okay. Aya, thank you for joining me and for helping us think more clearly about digital innovation strategy and its ethical stakes for listeners. I'm Alfred Marcus and this has been on the cusp on the New Books Network where we examine how strategy and ethics intersect in organizations and in civic life. Thank you for listening. If you have comments, suggestions for future podcasts, please be in touch at amarcusmn. Edu. Amarcusmn. Edu.