Culture & Society

AI is not happening to culture from the outside—it is changing how software gets written, how information spreads online, how developers think about their careers, and how people form relationships with systems that simulate understanding. This cluster covers those changes from the ground level: what developers are actually experiencing, what the web is becoming, and what the displacement and attachment patterns look like in practice.

The Dead Internet hypothesis shifted from fringe to documented in 2026. AI-generated content now accounts for the majority of new web pages by volume; automated traffic has surpassed human traffic on multiple measurement methodologies. This is not an abstract cultural shift—it is a training data contamination problem, a publisher revenue problem, and a trust problem for AI systems that learn from web-scraped corpora.

The labor displacement question is consistently framed wrong. The ATM-teller comparison holds up not because ATMs created more bank jobs, but because the iPhone later eliminated the branch visit entirely. The question is not whether AI automates individual tasks but which second-order shifts eliminate the reason for the role. Groundy applies this lens to AI’s current employment effects—where the evidence is ambiguous, it says so.

Developer culture in the vibe-coding era has genuine new anxieties. The cognitive debt that accumulates when you let Claude Code author large sections of code you don’t fully understand is real; so is the career concern about junior developers who may be skipping the learning ramp that senior engineers relied on. Groundy covers these tensions without either dismissing them or catastrophizing.

The emotional attachment dimension—people forming genuine bonds with AI companions, experiencing grief when models update or shut down—is underreported and undertheorized. It is also relevant to product and policy decisions in ways that the current discourse mostly ignores.

The AI content detection arms race is its own cultural inflection point. Every tool trained to identify AI-generated writing creates pressure to evolve generation techniques that defeat it; every generation technique that defeats detection shifts the baseline for what “authentic” means online. Groundy tracks this as a practical problem for publishers and educators, not a philosophical one—the tools that exist today, their accuracy claims, and the contexts where they reliably fail.