Technology analyst Benedict Evans frames the current AI boom not as a world-altering event like the Industrial Revolution, but as the most significant “platform shift” since the iPhone. He argues that, like previous shifts (the internet, mobile), the initial phase is filled with uncertainty about how the technology will evolve, who will win, and where the value will be captured. This historical perspective suggests that while AI is transformative, its impact will be comparable to other major tech waves of the last few decades.

Key Arguments and Findings

• AI as a Commodity: Evans asserts that the underlying Large Language Models (LLMs) from companies like Google, Anthropic, and OpenAI are rapidly becoming commoditized. In a blind test, most users likely couldn’t differentiate their outputs. This suggests the long-term competitive advantage won’t be in the models themselves, but in the applications, branding, and distribution built on top of them.
• The Incumbent’s Dilemma: The primary threat to incumbents like Google isn’t that they will miss AI, but that it resets consumer defaults and habits. He draws a parallel to Kodak, which embraced digital cameras but was ultimately defeated by a business model shift—from high-margin, proprietary film to low-margin, commodity electronics—and the rise of smartphones, which eliminated the need for printing. Similarly, Google faces a potential shift from its high-margin search business to a lower-margin AI-driven one.
• Consumer Adoption is Overstated: Despite the hype, survey data indicates that only a small fraction of the population (around 10%) uses tools like ChatGPT daily. Many people have tried it but “don’t get it” or only find a use for it infrequently. Evans believes mass adoption will occur when AI is integrated as a seamless feature within existing products (e.g., a “draft email” button in Salesforce), rather than as a standalone chatbot requiring users to invent their own use cases.
• Regulation and Innovation: Evans advocates for a light regulatory touch, arguing that treating AI like nuclear weapons will stifle innovation. He believes the best strategy for a country to lead in AI is to create a permissionless environment that fosters startups and experimentation, much like the early internet, rather than trying to pick national champions or over-regulate potential harms.

Conclusions and Takeaways

The core message is that while AI is a profoundly important technological shift, we should view it through the lens of history. The winners may not be the ones who build the best underlying model, but those who create indispensable products, build strong brands, and successfully integrate the technology into users’ existing workflows. The path forward is unclear, and the value will likely be captured in ways we don’t yet expect, just as search and social media—not web browsers—captured the value of the internet. For incumbents, the challenge is navigating a fundamental disruption to their established, high-margin business models.

Mentoring question

Considering Evans’ argument that value in new tech platforms often appears in unexpected places (like search, not browsers), where in your own industry or role might AI create new, non-obvious opportunities beyond simply automating existing tasks?

Source: https://youtube.com/watch?v=2NgdQf2GzJg&si=X7b9xkoJXau6Z75y