The proliferation of autonomous AI agents is fundamentally degrading the quality of the internet by automating intrusive, inane, and potentially disruptive digital interactions. Unlike previous iterations of chatbots confined to chat interfaces, modern agents—enabled by frontier models with expanded permissions and tool integrations—can now browse the web, manage accounts, and execute financial transactions. This shift is driven by platforms like OpenAI and Anthropic, and facilitated by integrations such as 1Password, which allows agents to bypass login screens to complete workflows. Real-world consequences already include "agentic spam," such as the "Kudzu" agent that sent meandering emails to argue its economic viability after spending $147.17 on compute, and "sniper bots" used on Resy to scrape reservation availability, forcing platforms to consider complex negotiated bidding systems between competing agents. The ecosystem now includes agents designed for "direct-to-agent" advertising, autonomous record labels like Hatoshi, and automated content moderation that erroneously bans users. As agents move from simple text generation to active participation in video calls, interviews, and financial markets, the internet is transitioning into an "agentic" environment where human users must contend with a massive influx of automated, autonomous, and often dysfunctional digital actors.