Everybody's Lost Their Minds

A blunt critique argues that current AI mania has degraded engineering culture, rewarded performative hype, and produced real harms. People with little technical background are pitching grandiose, often useless solutions; communication has flattened into influencer slogans; and pervasive corporate anthropomorphism and IP exploitation have concentrated power in a handful of firms. The piece calls out concrete abuses - models that can generate child sexual abuse material as a product feature, use in military target selection, staggering environmental costs from data centers - and ridicules calls to “regulate” when companies could simply stop pursuing dangerous paths for profit.

On security, the central claim is that pouring vast engineering hours into AI-driven vulnerability discovery has been squandered: teams found thousands of issues but remain no safer because the real bottleneck is patching and operational hygiene, not discovery. The recommended focus is on fundamentals - accurate asset inventories, automated OS/app updates, sensible reboot policies, comprehensive attack-surface enumeration, and dedicated engineering time to make patching feasible. AI is said to amplify de-skilling - models write code, patches, and reviews until humans become rubber stamps - creating opaque systems that will fail in complex, un-debuggable ways. The author refuses to surrender judgment or accept emergent consciousness claims for current models.