Chapter 1 Introduction

On November 30, 2022 a digital product with the curious name of "ChatGPT" was released to the public by OpenAI, a company not widely known at the time. The company had been founded since 2015, bringing together an elite group of the world's top AI researchers with the goal of finding a safe way to achieve Artificial General Intelligence (Tamkin et al., 2021), an elusive holy grail of AI research, referring to highly autonomous AI systems that perform better than humans at most kinds of work (Morris et al., 2024). After ChatGPT was released, in a matter of days it broke records for the fastest widespread adoption in the history of consumer software, reaching more than 1 million users in five days and 100 million users in two months (Teubner et al., 2023).

What makes ChatGPT's story particularly fascinating is not just its explosive growth but how it exemplifies a profound pattern that is the focus of this book - one that challenges our fundamental assumptions about innovation and value creation. OpenAI created something whose applications surprised even its own creators. As users worldwide began sharing their ChatGPT experiences on social media, they revealed uses that no one at OpenAI had anticipated. Mira Murati, then Chief Technology Officer at OpenAI, captured this phenomenon in an interview (Lynch, 2024):

It's one thing to statistically predict the performance of the model, it's another thing to actually see the capabilities when you test them across different domains and see that the models can rhyme, they can do extremely well in biology or math tests," she states. "That is completely different, when you actually bring it in the real world and see how the statistical performance translates into actually solving hard problems.

As you will see throughout this book, product creators being surprised by their own products is a common theme in a pattern we refer to as generativity: a system's ability to channel unpredicted innovations from a vast variety of contributors.

The artificial intelligence technology behind ChatGPT is not limited to the ChatGPT user interface. In fact, even before ChatGPT was released as a consumer product, OpenAI was selling the services of its models to developers through Application Programming Interfaces (APIs) which OpenAI has made incredibly easy to use. As of late 2024, around 3 million software developers around the world were building their own tools with these models or using the services of the OpenAI models in existing software tools (Metz, 2024). Because each developer uses a unique API key to access the AI models, API usage is tracked by OpenAI, and the company charges a rate for each individual API call.

The nature of OpenAI's products allow the company to profit from the innovations of others on a massive scale, without having to manage or even know much about those innovations internally, let alone make any predictions about their success. This mechanism, which we refer to as the generativity advantage, is an incredibly powerful engine of value creation and profit that has been the key driver of success for some of the world's most valuable companies in the digital era. Surprisingly however, generativity as a driver of competitive advantage has remained severely underappreciated in discussions of business strategy and entrepreneurship. Youngjin Yoo, a world-leading scholar of information systems, argued years ago that the "age of generativity" is upon us and that it calls for new management practices, as well as new theoretical models and frameworks to guide those practices (Yoo, 2013). It is time for the world of business strategy to catch up.

This is the gap addressed by the present book. While the book is mostly geared toward businesses that can be described as "digital platforms," many of the arguments apply to other types of businesses as well. By digital platforms, I mean digital products that facilitate the interaction of multiple stakeholders whose contributions can extend or increase the value of the product (de Reuver et al., 2018).

The core arguments of this book are as follows: a) the properties of generative technologies conceptualized in technology generativity theory can be viewed as intentionally achievable outcomes or design choices that can be pursued with strategic intent as features of generative products; b) the properties of dynamic markets conceptualized in economic theories of knowledge-generating economies can be pursued with strategic intent as design parameters or features of generative marketplaces; c) generative products and generative marketplaces, especially in combination with each other where generative marketplaces form around generative products, allow firms to tap into the generativity advantage: the mechanism of profiting from the distributed innovations of others on a massive scale. This involves taking user-innovators seriously as profit-generating entrepreneurs and helping them succeed and build sustainable businesses around your platform; and d) the implementation of a strategy to intentionally pursue a generativity advantage is not for everyone and is far from trivial: the path sometimes requires specifically going against certain approaches that are considered "best practice" in strategy and entrepreneurship and often involves strategic trade-offs along the way. These core arguments form the organizing logic of the chapters in this book.

Throughout the book, many anecdotes and cases are referred to that either illustrate examples of generative products and marketplaces, or illustrate certain product or marketplace features that can be conducive to generativity.

The Messiness of Generativity

The generativity advantage is fundamentally different from traditional approaches to innovation that rely on the company's own innovative capacity or its ability to identify and acquire promising innovations from others. A generativity advantage derives from creating the conditions for an entire ecosystem of external innovators to generate value in ways that benefit the platform owner, without requiring the owner to anticipate or directly manage those innovations.

The implications of this approach are profound. Traditional strategic thinking emphasizes control, prediction, and direct management of innovation. But the generativity advantage flips this logic on its head. It suggests that in certain contexts, the most powerful competitive advantage comes from deliberately creating conditions for unplanned, distributed innovation. This is not merely about "open innovation" or "crowdsourcing" - it is about architecting systems of uncontrolled distributed value creation, and channeling a fair amount of that value back to the platform creator.

However, this lack of control over the innovations that happen around your product is not without its costs. Generativity is often messy and noisy, and is not always desired. An example of a non-generative digital platform is Uber. You need it to do something specific (get from point A to point B), and you would rather that the app does this reliably and predictably, rather than for the app to introduce and encourage all sorts of unpredictable innovations from its drivers. Even for a product as generative as the smartphone, a loss of control is not always an intuitive goal to pursue. Steve Jobs was famously against allowing third-party apps on the iPhone, initially resisting what has hence become Apple's greatest money-making machine (Stier, 2024).

OpenAI has had to deal with its own share of the messiness that comes with generativity. The first method in which OpenAI opened its ecosystem to external innovations was through the API. But many entrepreneurs and innovators who started building with the API were ridiculed for building simple "GPT Wrappers," a term used in a somewhat derogatory way to mean that these products were not adding much value beyond what ChatGPT could do itself, and were not really sophisticated or hard-to-build accomplishments to be proud of. Nevertheless, many entrepreneurs who persisted and moved fast were able to profit handsomely from such products, although there was and continues to be uncertainty around which applications will be replaced by features in OpenAI's own products and which applications will continue to provide unique value.

By early 2025 after the release of the DeepSeek models that pointed toward a rapid commoditization and severe cost competition at the model layer of generative AI, it became apparent that much of the profit margin was going to be in the application layer, where there is much more room for diverse and unique ideas. Pieter Levels, the famous indie entrepreneur who profited handsomely with his PhotoAI product at the application layer, tweeted1 in reaction to the DeepSeek developments that "I can't believe we all accidentally betted right on making GPT Wrappers, even though everyone trolled us for it for years, and are now living The Good LifeT." But the story around the messiness of masses of external innovations got even more interesting when OpenAI opened up the ecosystem to non-developers.

The Saga of Laundry Buddy

In its first attempts to create a managed marketplace around its products that went beyond APIs, on the occasion of OpenAI's first developer day in...