The current discourse surrounding artificial intelligence oscillates wildly between utopian visions and apocalyptic warnings, often obscuring the practical realities of its development and integration. This tension formed the core of a recent discussion between technology analyst and former a16z partner Benedict Evans and General Partner Erik Torenberg on the a16z podcast, where they dissected the true scale and impact of AI by drawing parallels with historical technological shifts. Their conversation, rooted in Evans's latest presentation, "AI Eats the World," offered a grounded perspective on what AI is, what it isn't, and what lessons history might offer.
Evans begins by noting a peculiar linguistic phenomenon: the term "AI" itself is fleeting. "In actual general usage, AI seems to mean new stuff," he observes, highlighting how technologies once deemed artificial intelligence, such as machine learning or databases, are eventually absorbed into the fabric of everyday computing, losing their "AI" designation. This cyclical nature suggests that much of what we call AI today will, in time, simply become "more software."
The true challenge in assessing AI's future lies in its fundamental unpredictability. Unlike previous technological advancements where physical limits were often understood (e.g., bandwidth, battery life), AI's theoretical boundaries remain largely unknown. This inherent uncertainty fuels what Evans terms "vibes-based forecasting," where even leading figures in the field offer diverging opinions on AI's capabilities, from Sam Altman’s conviction of PhD-level researchers to Demis Hassabis’s more cautious stance.
For most, the practical application of raw AI models remains elusive. Though platforms like ChatGPT boast hundreds of millions of weekly active users, Evans provocatively asks why "five times more people look at it, get it, know what it is, have an account, know how to use it, and can't think of anything to do with it this week or next week." This disparity underscores a crucial insight: the transformative power of AI will largely manifest through its integration into specialized products and user-friendly interfaces, much like Excel revolutionized accounting or the graphical user interface simplified computing.
This leads to a "schizophrenia" in the AI conversation, where the grand pronouncements about artificial general intelligence (AGI) often overshadow the more immediate, tangible developments. While some envision AI as an autonomous, super-intelligent entity, others focus on its utility as a tool to enhance existing workflows. The real impact, Evans suggests, will be found not just in entirely new applications, but in how AI redefines and optimizes existing industries, from law firms leveraging AI for discovery to marketing teams generating hundreds of assets with greater efficiency.
However, the rapid influx of capital into AI also carries the risk of a bubble. "Very new, very, very big, very, very exciting, world-changing things tend to lead to bubbles," Evans cautions. While over-investment can accelerate technological progress by funding ambitious projects, it also creates an environment where exuberance can outpace reality. The question then becomes whether AI will primarily foster net-new trillion-dollar companies, akin to Google and Facebook emerging from the internet and mobile shifts, or if its gains will largely be captured by existing tech giants. The precise balance remains unclear, but history suggests that true paradigm shifts often create unforeseen winners.
