Machine, Platform, Crowd: Introduction and Organizational Transformation

Polanyi's Paradox and the Intuition of Systems

• Polanyi's Paradox is based on the concept that individuals know more than they can explicitly tell or articulate.
• This paradox is evident in complex games like Go, where players possess an intuitive understanding of the systems and strategies that is difficult to describe in set rules.
• Language acquisition serves as a primary example of this intuition:
  • Native speakers intuitively learn how to speak English without thinking about the specific parts of speech being utilized.
  • English adjectives follow a strict, innate order: Opinion, Size, Gauge, Shape, Color, Origin, Material, Purpose, Map.
  • An example of this ordering in practice is the phrase "lovely little old rectangular green French silver widowy map."

The Evolution of AI and Language Generation

• Historically, AI systems struggled with language because they could not replicate human intuition or make sense of semantic meaning.
• Noam Chomsky famously critiqued AI systems by using the phrase "colorless green ideas sleep furiously," which fits grammatical rules but lacks intuitive logical sense.
• The development of the Transformer technology has enabled systems like ChatGPT to develop an intuition for how words fit together.
• Originally intended for autocomplete and predicting the next word in a sentence, overtraining revealed that these models could effectively answer complex questions.
• In order for an AI to predict the next word accurately, it must possess some underlying understanding of the world to determine if a word choice makes logical sense.

Drivers of Modern AI: Moore's Law and Data Explosions

• The rapid emergence of AI technology is driven by Moore's Law, which states that computational power increases at an exponential rate over time.
• The volume of available data has grown significantly since the early days of the Internet when the entire web could potentially fit on a single phone.
• The Internet of Things (IoT) includes WiFi-connected devices like lights that produce massive amounts of data about the world.
• Deep learning algorithms utilize this data to predict outcomes and perform specialized tasks:
  • Mastering the game of Go.
  • Controlling complex data center equipment.
  • Predicting medical conditions, such as cancer.

Organizations and Decision-Making: Humans vs. Machines

• There is a fundamental shift in organizational structure from humans making decisions to machines making decisions.
• In the past, decisions were often made by "HIPPOs" (Highest Paid Person's Opinion), whose choices might not be based on facts or data.
• Modern organizations are moving toward encoding information in "bits" (digital ones and zeros), allowing AI to assist in making data-driven decisions.

Principles of Platform Business Models

• A platform is defined by a three-part structure: Producers, an intermediary Platform, and Consumers.
• Traditional business models involve a single producer creating a product (e.g., mugs) and selling it directly to consumers.
• The future business model involves platform ownership, where a platform owner connects diverse producers (e.g., various mug makers) to consumers through a distribution network.
• This model allows for a wide variety of products to satisfy a larger consumer base.

Case Study: Apple's Transition to a Platform

• Apple is primarily a platform company, although it continues to sell individual products.
• In the $1980s$, $1990s$, and early $2000s$, Apple focused strictly on creating products like computers (e.g., the Lisa) and software.
• The launch of the iPhone and the subsequent creation of the App Store in $2008$ marked Apple's transition into a platform.
• Steve Jobs was initially a "control freak" who hated the idea of the App Store and opposed external developers creating apps for the iPhone.
• Internal executives successfully argued for an open intermediary model, which became a massive success.
• The revenue model involves Apple Pay and taking a $30\%$ cut of all in-app purchases.

Conflict in Platform Economics: Epic Games vs. Apple and Google

• The popular game Fortnite, developed by Epic Games, challenged the $30\%$ fee for in-app purchases.
• In $2020$, Epic Games attempted to circumvent Apple's fees on items like skins and emotes.
• At the time, in-app purchases generated approximately $30,000,000$ per month for Apple and $1,000,000$ per month for Google.
• Both Apple and Google removed Fortnite from their stores, resulting in lawsuits.
• The outcomes differed between the two companies:
  • The court ruled that Apple was not monopolistic in its enforcement of the $30\%$ tax, and Fortnite had to pay the fees to be reinstated.
  • The case against Google suggested intent of monopolistic behavior in other areas of their business, allowing Epic to circumvent fees on the Google App Store.

Value Propositions of Platforms for Stakeholders

• Platform Owners:
  • Generate significant revenue.
  • Benefit from massive distributed innovation from thousands of developers.
  • Transfer risk to developers, as Apple does not have to fund the development of every app.
  • Capture the "Long Tail," serving niche interests at the far ends of a distribution curve without high resource investment.
• Developers:
  • Access to standardized components and tools provided by the operating system, making innovation easier.
  • Immediate market access to millions of iPhone users.
• Users:
  • Mix-and-match customization of their devices.
  • Faster innovation cycles than a single company could provide.
  • Increased competition between apps, which improves quality.
  • Lower search and transaction costs via a centralized App Store.

The Power of the Crowd: Innovation and Intelligence

• The Internet allows organizations to engage with communities outside their traditional corporate boundaries.
• GE Case Study:
  • GE utilized the "crowd" to help design a nugget or "pebble" ice machine.
  • They gathered market intelligence through user voting and utilized external labor for designs.
  • The resulting product sold for $409$ and received four-star reviews.
• Nupedia vs. Wikipedia:
  • Nupedia was a closed-source encyclopedia that relied on $21$ expert editors, which limited the amount of information it could produce.
  • Wikipedia, created by Jimmy Wales, utilizes $100,000$ active editors who contribute specialized knowledge to create a massive volume of pages.
• Prediction Markets:
  • Engage the crowd by aggregating individual pieces of information to create an average prediction of future events.

Open Source Success: Linux and the Linus Torvalds Model

• Linus Torvalds created Linux, an open-source operating system that serves as the basis for Android.
• At its inception, the market was dominated by Windows and MacOS; Linux introduced a global community of developers working for free.
• Linux is considered the most stable operating system and is used extensively for servers.
• Strategies for Organizing the Crowd:
  • Broad Initial Requests: Recruiting people by making the initial request as wide as possible to attract diverse interests.
  • Non-credentialism: Contributions are judged based on the quality of the code rather than degrees or age.
  • Git and GitHub: Torvalds created "Git," a technology that allows multiple developers to manage code and reverse changes if something breaks.
  • Clear Outcomes: Establishing specific technical expectations and working results.
  • Self-Organization: Allowing contributors to choose which aspects of a project they want to work on.

Strategic Rebalancing: Minds, Platforms, and the Crowd

• Business technology is understood through three lenses, each with an "analog" counterpart that requires rebalancing:
  • Machine Intelligence vs. Human Minds.
  • Platforms vs. Products.
  • Crowd vs. The Core (traditional employees hired by the organization).
• Organizations must find a mixture of these elements rather than being $100\%$ in one category.
• McKaffney and Rudolfsen suggest that rebalancing toward machines, platforms, and crowds is essential for being "future-ready."

Historical Context: The Second Machine Age and Technological Shifts

• Phase 1 ($1990$): The rise of the digital Internet. Organizations shifted from physical phone books to digital presences during the .com boom.
• Phase 2: Movement toward "science fiction" technologies like AI that can perform non-routine tasks.
• Key Characteristics of Phase 2:
  • Ubiquitous processing and connectivity (e.g., $5G$).
  • Information encoded in bits rather than physical "atoms."
• Historical Analogy: Steam vs. Electricity:
  • Steam-powered factories were built around a large central motor; machines that needed the most power had to be closest to it.
  • When electricity arrived, factories initially replaced the steam motor with an electric one without changing the layout, which was inefficient.
  • It took $30$ years for factories to reorganize around electricity, eventually becoming $203,100\%$ more efficient.
  • Modern organizations may be repeating this mistake by injecting AI into human processes without truly reorganizing around the technology.

Challenges and Failures in Platform Development

• The Curse of Knowledge: Successful winners in one environment often fail to take the risks necessary to win in the next era (e.g., Nokia failing to adopt platform business models).
• Reasons Platform Businesses Fail:
  • Mispricing: Failing to balance subsidies between producers and consumers.
  • Lack of Trust: Failure to manage the user experience or remove bad actors (e.g., overbooked rentals on Airbnb or buggy apps).
  • Dismissing Competition: Preemptively ignoring rivals, similar to how Internet Explorer was the "king of browsers" in $2005$ before losing ground.

Class Applications: Business Model Canvas

• Students will work in groups to develop app ideas for specific industry verticals.
• The project will use a Business Model Canvas to analyze partners, value propositions, and business logic.
• Students are encouraged to identify industries that have not yet engaged machines, platforms, or crowds for innovation.

Questions & Discussion

• Participant Confusion: One student noted that the material was difficult to process, stating the concepts "just went over my head" and resulted in them sitting on their phone instead of following the material.