The AI Bubble Explained

Identifies four widespread misconceptions about AI and rebuts them with technical, economic, and cultural arguments. The first misconception - “AI is just math” - is challenged by pointing to emergent behavior in large networks (neurons akin to logic gates, scale creating qualitatively different capabilities) and to historic analogies (computers, engines) showing that mathematical systems can transform labor and society. The claim that AI will kill us within a decade is rejected as alarmist: existential risk is framed as a long-term concern (decades to a century) while the immediate problem is economics. Companies chasing AGI are burning unsustainable sums, models aren’t scaling linearly, and open-weight models (examples named) plus local inference on consumer hardware are rapidly eroding frontier firms’ moats; pauses and pronouncements from CEOs are driven by financial fear, not purely safety concerns.

Argues the current situation is a financial bubble that can burst even as the technology itself endures and reshapes daily life. The author highlights a “shell game” of money circulation (with a few hardware suppliers profitably subsidizing the rest) and warns of recessionary fallout, while predicting enduring productivity gains - from autonomous personal agents to email elimination - alongside societal pushback and niche demand for non‑AI spaces. Concludes that using AI is neither a moral betrayal nor an invitation to full surrender: nuance and selective adoption are required.