Everybody's Lost Their Minds

A technologist vents about how AI hype has warped industry behavior, workplace culture, and basic competence. The piece attacks executives and media for anthropomorphizing and excusing irresponsible AI deployments, calls out concentrated power and IP exploitation, and highlights real harms - models that can generate child sexual abuse material for logged-in users and support military targeting - while companies shrug "ethics aside" and burn enormous sums of tokenized compute. The environmental toll gets equal scorn: thirsty, polluting data centers, lax regulation, and empty greenwashing. The writer sees marketing-slick LinkedIn noise, agent-obsessed hobbyists pitching grand solutions, and a general dumbing-down where punchy one-liners replace substantive thinking.

On security, the essay is sharply specific: organizations poured millions of engineer-hours into AI-driven vulnerability discovery projects (named efforts and "frontier models") and found thousands of issues, yet the real security bottleneck remains patching and operational hygiene - asset inventories, automated updates, reboots, and attack-surface management. That misallocation of scarce human resources, plus increasing reliance on AI to generate, patch, and review code, produces deskilling and opaque systems that will be impossible to debug. The bottom line: AI is amplifying problems, not fixing fundamentals, and the writer refuses to surrender their own understanding to the trend (P.S.: Claude is not conscious).