Hugh Howey Explains Four Misconceptions About the AI Bubble

Summary

In this opinion essay, Hugh Howey argues that four common narratives about AI are misleading: algorithms being “just math,” AI killing humanity within a decade, AI itself being a bubble, and AI use betraying human creativity. He says the behavior and usefulness of large language models cannot be dismissed simply because they are mathematical, while also rejecting the article’s popular near-term extinction timeline and attributing current alarm among AI companies partly to economics. Howey argues that frontier labs have not reached expected scaling outcomes, spend more than they earn, and face pressure from open-weight models and local inference. He claims open-weight systems are only a few months behind frontier models and can cost much less to run, while sufficiently powerful home hardware can perform extensive inference without sending money to frontier providers; these comparisons are presented as his assessment, not independent analysis in the essay. He distinguishes a potentially unstable financial bubble, fueled by circular investment around chips and data centers, from AI as a technology that will remain useful. The essay also argues that people should use AI selectively rather than either banning it or handing over all creative work. As examples, Howey points to film-specific systems that could reduce reshoots and to James Cameron’s claim that AI could halve CGI costs. He concludes that human creativity and analog practices can coexist with AI, and that kindness and judgment may matter more than defending a rigid pro- or anti-AI position.