Economics of Digital Tech and AI

Basic Economics of Digital Technology and AI

• Understanding the core economic shifts spurred by digital tech and AI adoption is critical
• Consider which costs decrease with new technologies such as ICT

Impact of Digital Technology

• Digital tech reduces various costs, shaping market dynamics (Goldfarb and Tucker 2019)
• Search costs increase with product variety but also contribute to information overload and strategic obfuscation
• Replication costs drop, facilitating the rise of megabundles
• Transportation costs reduce through logistics enhancements, which also decouple value and location
• Tracking costs increase resulting in attention markets and personalized pricing and revive price discrimination strategies.
• Verification costs reduce through scaled feedback mechanisms
• Experimentation costs are lower through digital twins and A/B tests, which promote digital entrepreneurship and layering services

AI's Economic Impact

• Examining how AI influences overall economy focuses on what cost it reduces
• With data and computing, AI systems are efficient prediction machines
• AI lowers the cost of prediction
• Prediction is an input into decision making
• AI increasingly executes prediction tasks more effectively
• Prediction-intense sectors include logistics, adtech, fintech, HR, and scientific discovery

Information Goods and Digital Economics

• Software is a core product: information goods have unique characteristics
• High fixed production costs and low reproduction costs lead to increasing returns to scale
• Economies of scope: data feedback loops enhance applications
• Consumption externalities: user benefits grow as the user base expands
• Digital markets tend to tip toward monopolization and concentration
• Fixed costs are barriers that grow economies of scale
• Network effects serve as positive feedback for incumbents
• Incumbents dominate through killer acquisitions, exploitation and vertical and lateral expansion

Price Discrimination

• Price discrimination sells the same product at different prices (Belleflamme 2005)
• First-degree price discrimination results in personalized pricing
• Third-degree involves group pricing based on age, location, etc.
• Second-degree price discrimination enables product versioning to match consumer willingness based on freemiums, damaged goods, bundling etc
• Digital economy supports personalized advertising

Platform Economics

• Platforms support variety by constraining linkages among components (Baldwin and Woodard 2008)

• Platform architecture separates stable core and variable peripheral components that support core components

• Core components have low variety and high reusability

• Peripheral components have high variety and low reusability

• Interfaces between core and complements are critical

• Find platform architecture in product design, product families, and marketplaces

• Platforms are economic drivers in the digital economy

• Platforms succeed by leveraging digitalization's key ingredients (economies of scope, increasing returns, network effects)

• Architectures supporting low-cost structures are typically used for intermediation

• Affected industries worldwide are platformized, yielding largely platform-based top companies

Types of Platforms

• Multi-sided platforms connect user groups and internationalize externalities

• Multi-sided platforms face the chicken-and-egg problem

• Ignition strategies involve pricing and marquee users

• Integrating a side risks self-preferencing

• Peer-to-peer markets offer market entry to heterogeneous actors

• Accomodation, crowdfunding, rides, and deliveries offer supply an demand

• Ad-funded vs device-funded business models chart cloud vs edge tension and impact welfare

AI Alignment with Key Concepts

• AI systems have the properties of information goods

• Complex with high marginal costs of compute but reusable software

• A few solutions dominate

• Early head-starts retain incumbency

• AI enables versioning, leading to product proliferation

• AI products such as ChatGPT and Claude use interfaces for user applications

AI Accessibility and Market Dynamics

• Market shift moving to off the shelf models and builders of applications
• There is a gradient in AI models releases
• Emerging intermediaries are found in model marketplaces
• Model marketplaces help lower costs for standardizing comparisons

AI Production Factor Market

• Increasing expenses to produce marketable output
• Requirement that medium recoup investments
• Ongoing scaling up discussions