The Post-AI Internet Doesn't Look Great

TEXT START: Since the explosion of usage in generative AI, much of the internet is pretty much unusable.

The Dissection

The text correctly identifies the visible symptom: AI has made information production nearly free, so abundance is consuming signal. Search engines are replacing source discovery with machine summaries; social platforms are filling with synthetic engagement; human curation is retreating into private, subscription-based enclaves.

The author’s workaround—personal sites, RSS, Substack, selective YouTube—is altitude selection. It protects one individual from the sludge. It does not repair the system or preserve an open, reliable information commons.

The Core Fallacy

The text treats the collapse as a temporary usability problem rather than an equilibrium produced by P1. When AI-generated content is cheaper and scalable, platforms have a competitive incentive to produce more of it. Human institutions cannot maintain stable human-only domains across the open internet at scale; that is P2.

The problem is not merely poor moderation. AI also attacks the economic foundation of source-driven publishing by intercepting attention and traffic. High-quality human content may survive, but increasingly as a gated luxury financed by subscriptions, patronage, or institutional ownership—not as the default public layer.

Hidden Assumptions

• Platforms will choose signal over cheap, monetizable volume.
• Human-authored material will remain abundant, identifiable, and economically funded.
• Users can curate their way out without losing access to essential information.
• Verification will remain cheap enough for ordinary users.
• The open internet can be restored without changing who owns the compute, distribution, and filtering infrastructure.

All are weak assumptions. The final hope that the internet becomes usable again has no identified mechanism behind it.

Social Function

Classification: partial truth functioning as a lullaby and transition-management memo.

It makes structural displacement legible as personal irritation. The reader is encouraged to retreat into curated enclaves while ownership and control consolidate above them. This is not necessarily deliberate propaganda, but it domesticates a systemic break: curate your feeds while the economic base underneath information work is automated.

The Verdict

The article smells the corpse accurately but misdiagnoses the cause. The post-AI internet will bifurcate into a synthetic open web optimized for machines, advertisers, and automated agents, alongside expensive trusted enclaves protected by subscriptions, closed communities, and verification services.

Usability can return locally for people with money, expertise, or privileged access. Universal restoration is not coming. The deeper event is that AI is stripping content production, moderation, and discovery of necessary human labor—an early cultural manifestation of P1 and the broader collapse of productive participation under P3. Hope is not a counterforce; control of the filters is.