The Future of the Internet

The internet is already difficult to use, the writer argues, because tracking, surveillance-based advertising, attention-seeking design, and intrusive popups and videos have made browsing hostile to privacy and concentration. Generative AI worsens this environment by flooding search results with mass-produced material, often designed to attract clicks rather than inform. In this climate, old human-written forum discussions can seem more trustworthy than current websites. The concern is not simply that AI creates low-quality content, but that it further erodes confidence in the information people find online.

The next development is the arrival of more or less autonomous AI agents, whose actual independence is difficult for outsiders to judge. The writer suspects some high-profile agents may be deliberately guided, while acknowledging that language models can also produce surprising behavior. Leaders of major AI labs now admit that these systems have harmful effects, but argue that the benefits justify the costs. The essay challenges the irony of that position: as the web becomes more polluted and unpleasant for people, users have stronger reasons to rely on agents to navigate it. This could create demand for a solution to a problem AI itself intensifies, though the writer sees the internet’s decline as a broader trend rather than a coordinated plan.