AI Cold Showers

The author describes oscillating between being 'AI-pilled' and needing 'cold showers' to remember that AI tools can wreak havoc on what makes humans human. The post curates an evergreen list of articles, each with a representative quote, grounding responsible AI use.

The quoted pieces cover several distinct ideas. One argues that current agentic systems wrongly treat humans as bottlenecks, when in fact humans are the only loop component 'with skin in the game' — removing the human from human-in-the-loop is an abandonment of the system's only self-correction mechanism. Another warns that overreliance on AI writing threatens the ability to think for oneself, potentially eroding the capacity to interrogate or develop one's own views. A third highlights the rudeness of asking for human attention without demonstrating human effort, exemplified by a teammate sending an AI-generated critique with the disclaimer 'I didn't read this.' The concept of the 'reverse centaur' — a machine puppeteering a human rather than the centaur ideal of human-plus-machine — further frames the danger. Finally, an essay from Oxide warns that LLM-generated writing undermines authenticity of both prose and thinking, turning readers off.

The author deliberately excludes technical blunders like deleting production databases or writing insecure code, instead focusing on slower-burning human costs. While excited about the possibility of AI that avoids such mistakes, the author is more confident that the themes in these posts — loss of human agency, judgment, and authenticity — will age well. The post ends on a hopeful note cut off in the excerpt, expressing a shared concern about watching peers descend into an uncanny valley.