There's a 100% Chance AI Agents Are Ruining the Internet

AI agents—autonomous software given permission to act on the internet—are already degrading the online experience by flooding inboxes with spam, making poor moderation decisions, and generally becoming increasingly annoying, even if they don't pose existential risks. Recent advances have removed the guardrails that previously confined AI to chatboxes, allowing agents to access accounts, send emails, and take autonomous actions with growing frequency and capability. The author argues that regardless of whether AI poses long-term existential threats, the immediate and certain problem is that AI agents are already ruining the internet through their current unsupervised deployment.