Making sense of the debate over AI psychosis

The tech industry has entered a period of profound ideological intensity, recently characterized by a phenomenon critics are labeling "AI psychosis." This term describes more than just simple enthusiasm for new software; it refers to a distinct psychological and strategic state observed among Silicon Valley’s elite. At its core, this state is defined by a singular, near-religious fixation on Artificial General Intelligence (AGI) and a tendency to prioritize speculative, existential risks over the tangible, material impacts of current AI deployments. As leading CEOs pivot their entire corporate identities toward these far-flung futures, the boundary between visionary leadership and detached delusion is becoming increasingly blurred.

This current climate did not emerge in a vacuum. It is the culmination of years of "longtermist" philosophy and effective altruism—intellectual movements that have deeply permeated the executive suites of OpenAI, Anthropic, and Google. Historically, Silicon Valley has thrived on lucrative "disruption," but the scale of the current rhetoric is unprecedented. Where once founders spoke of connecting the world or organizing information, they now speak of birthing a new form of digital life that could either save or end humanity. This shift suggests a departure from traditional business cycles into a realm where high-level executives function more like prophets than stewards of shareholder value.

The mechanics of this "psychosis" are driven by a feedback loop of massive capital requirements and competitive pressure. Developing frontier models requires billions of dollars in compute, which in turn necessitates a narrative of world-changing importance to justify the investment. To maintain the flow of talent and capital, CEOs must promote the idea that their work is of cosmic significance. This creates a strategic necessity for hyperbole, which eventually shapes the internal culture of these firms. When a CEO begins to believe their own marketing—treating a sophisticated statistical model as a burgeoning consciousness—the resulting decision-making process can become volatile, leading to internal fractures and sudden pivots.

The industry-wide implications of this state are significant, particularly concerning the misallocation of regulatory and cognitive resources. By focusing the global conversation on sci-fi scenarios like rogue superintelligences, tech leaders successfully divert attention from more immediate, pressing harms. Issues such as algorithmic bias, mass copyright infringement, and the environmental cost of data centers are often relegated to the background. This "psychosis" therefore serves as a potent, if perhaps unintentional, shield against pragmatic regulation, focusing the eyes of lawmakers on a hypothetical future rather than the messy present.

Furthermore, this ideological fervor is beginning to create a rift within the labor market. Engineers and researchers are increasingly forced to choose between "accelerationist" camps, which view speed as a moral imperative, and more cautious factions. This polarization complicates the corporate governance of companies that are essentially running the modern world’s infrastructure. If the leaders of these companies are operating under a set of assumptions that do not align with physical or economic reality, the risk of a "bubble burst" extends beyond mere finance into the very reliability of the systems we are becoming dependent upon.

Moving forward, the primary factor to watch will be the collision between AI idealism and quarterly earnings. As the astronomical costs of AI development continue to climb, investors will eventually demand more than visions of a post-scarcity utopia; they will demand a return on investment. Whether these "AI-obsessed" leaders can pivot back to practical utility, or whether they will double down on their existential narratives as a way to explain away missed targets, will determine the next era of the technology sector. The ultimate test of Silicon Valley’s current state will be whether its most powerful figures can distinguish the software they have built from the myths they have created about it.