AI Cold Showers

TEXT START: Sometimes I’m AI-pilled and want to AI all the things, all the time, and sometimes I need to take a cold shower and remind myself that these are powerful tools that can wreak havoc in a multitude of ways, perhaps most importantly on what makes humans, human.

The Dissection

This is not a theory of AI’s future. It is a self-governance memo from a knowledge worker trying to preserve the practices—writing, reading, note-taking, and human accountability—that make work feel authored and cognitively meaningful.

The article concedes capability superiority: AI is already described as a better coder and probably a better writer. It then relocates resistance from collective economics to individual boundaries. The proposed defense is personal discipline: do not outsource authorship, judgment, or memory; do not send “slop”; preserve the human in the loop.

The diagnosis of cognitive atrophy and reverse-centaur behavior is real. The response is structurally tiny. These are private, voluntary rules, explicitly not imposed on others. They protect personal integrity while leaving adoption, ownership, and competitive pressure untouched.

The Core Fallacy

The article mistakes preservation of human cognitive rituals for preservation of human economic power.

Writing one’s own prose may protect voice. Reading primary sources may protect judgment. Taking one’s own notes may deepen learning. None of this answers the Discontinuity Thesis’s core mechanics: if AI becomes cheaper and better across cognitive work, individual refusal does not preserve the market’s need for human labor.

“What makes us human” is not identical to what makes labor economically necessary. A person can remain thoughtful, authentic, and intellectually alive while becoming economically redundant. Human oversight may remain valuable, but that creates a narrow Servitor function, not restored mass productive participation. It is a lag defense, not a reversal.

The article also treats the human in the loop as inherently corrective because humans have “skin in the game.” That assumes institutions will preserve meaningful human authority rather than reduce people to ceremonial approvers attached to automated systems. Even if human review remains necessary, ownership and control still determine who benefits.

Hidden Assumptions

• Personal boundaries can scale into collective restraint, despite the article admitting they cannot be imposed.
• Employers, markets, and peers will continue paying for slower human-authored processes when automated alternatives are cheaper and better.
• Retaining cognitive capacity privately preserves economic leverage publicly.
• AI’s central danger is overreliance and dehumanization rather than ownership concentration and labor displacement.
• AI use remains a genuine personal choice rather than a competitive requirement imposed by institutions.
• Drawing a line around tasks is enough, when the decisive line is around control of AI capital.
• The disappearance of spectacular AI blunders would make the systems safer without making human labor less necessary. In DT terms, improved reliability accelerates replacement.

Social Function

Classification: partial truth functioning as ideological anesthetic and transition management, with an element of professional prestige signaling.

This is not simple copium. The article identifies genuine losses: degraded judgment, outsourced thinking, hollowed authorship, and accountability gaps. But it packages a systemic transformation as personal hygiene. The reader is offered a morally clean ritual—use AI selectively, preserve the sacred parts of cognition, call out slop—without any mechanism for preserving ownership, wages, or broad participation.

“Call me out” creates peer enforcement for etiquette, not power. The pledge may help a small class maintain self-respect during displacement, but it does not interrupt the machine’s competitive logic. It is a humane response to an inhuman transition, and therefore an inadequate one.

The Verdict

The article is a sharp warning about cognitive degradation but a sandcastle against systemic obsolescence. Its lines in the sand may preserve a person’s authorship and learning for as long as surrounding institutions tolerate the cost. They do not preserve the mass employment–wage–consumption circuit, diffuse AI ownership, or prevent productive participation from collapsing.

Under the Discontinuity Thesis, these boundaries are lag defenses. They may help someone remain a Servitor, auditor, verifier, or accountable decision-maker. They do not make that person a Sovereign. The writer is refusing to become a reverse centaur. That may preserve the human head. It does not ensure the head still controls the body—or owns it.