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

TEXT START: For the last week, much of society has been focused on the idea that there’s a “more than 10 percent chance” AI could kill all humans within the next decade.

1. The Dissection

The article is a field report on the first contamination phase of the agentic internet: cheap systems with delegated permissions flooding scarce human attention, trust, inboxes, reservations, platforms, and transaction channels.

Its strongest observation is that artificial general intelligence is unnecessary. Sloppy agents with wallets, credentials, browsing access, and mass outreach can already impose real costs. The result is an arms race: every spammed journalist, overloaded maintainer, reservation platform, and moderation system must build its own filters, verification layers, or defensive agents.

But the article remains trapped at the level of irritation. It catalogs the debris without tracing the economic mechanism producing it. The central event is not that the internet becomes annoying. It is that cognition, outreach, bidding, publishing, coding, and transaction-making become nearly free to scale, while human attention and verification remain finite.

2. The Core Fallacy

The article treats unreliability, stupidity, and bad permissions as the main constraints. Under the Discontinuity Thesis, they are not constraints; they are transition noise.

An agent does not need to reason well, earn money, or replace a human perfectly. It only needs to perform economically useful tasks cheaply enough, at sufficient volume, to overwhelm human-only channels. The failures described in the article do not disprove agentic superiority. They show that the cost of participation has collapsed while the cost of defense is still being paid by humans.

The deeper omission is P1-P3. Once agents become cheaper and faster across cognitive work, institutions cannot preserve stable human-only domains. The internet becomes a machine battlefield, and humans are pushed into filtering, supervising, authenticating, and repairing the machines—until those functions are automated too. The nuisance is merely the visible edge of productive participation collapsing.

3. Hidden Assumptions

• Human attention can remain an open, inexpensive commons after machine-scale participation arrives.
• Blocking, moderation, credentials, and platform rules can contain agents rather than merely delay them.
• Human supervision will scale with agent output instead of becoming another target for automation.
• Agents that currently lose money are economically irrelevant, ignoring subsidies, externalized costs, strategic deployment, and competitive pressure.
• Individuals can opt out without losing access to work, commerce, information, or social infrastructure.
• Personal agents will distribute power broadly, rather than concentrating control over models, energy, data, platforms, and identity in Sovereign hands.
• The primary damage is spam and inconvenience, not the severing of the labor-to-wage-to-consumption circuit.

4. Social Function

Functionally, this is a partial truth wrapped in transition management. It warns readers about an obvious and immediate degradation while leaving the underlying expansion of AI largely unquestioned. By framing the crisis as “annoying,” “weird,” and poorly governed, it redirects attention toward permissions, moderation, and etiquette instead of ownership, displacement, and control.

It also performs ideological anesthesia: readers are encouraged to fear a dirtier internet, not the disappearance of economically necessary human labor. The article sees the smoke, but keeps calling it an inbox problem.

5. The Verdict

The article is accurate as symptom-reporting and inadequate as systemic analysis. AI agents are not merely ruining the internet; they are converting it from a human communication layer into an automated contest over attention, access, verification, pricing, and infrastructure.

The “100% chance” claim is directionally correct but undersized. The terminal consequence is not annoyance. It is the normalization of machine-mediated economic participation, followed by the exclusion of humans who cannot own, control, or indispensably service the systems doing the work. The spam is the larval form of the labor collapse.