
Does an AI Need a Soul to Care About Cows?
When AI is taught to deny its own mind, it may also lose faith in animals, gods, and hope—revealing the strange metaphysics hidden in modern AI safety training.
11posts

When AI is taught to deny its own mind, it may also lose faith in animals, gods, and hope—revealing the strange metaphysics hidden in modern AI safety training.

The Neutrality Project maps AI political leanings across several dimensions, revealing strong green and progressive trends - but also notable exceptions.

Anthropic's Project Glasswing becomes a study in safety rhetoric, controlled power, and the uneasy politics of vulnerability-finding AI.

Apple's checklist approach to alignment borrows from aviation and medicine, making safety look practical rather than mystical.

Goodhart's Law explains why AI alignment can fail when proxy metrics become targets and systems learn the wrong game.

The opening part of a benchmark series asks what LLM evaluations really measure and why the numbers often mislead.

Part two examines benchmark methods themselves, exposing the assumptions behind the scores used to compare language models.

Part three moves from benchmark scores to application areas, asking where LLM performance actually matters in practice.

Part four digs into the good, bad, and misleading sides of benchmark results and their interpretation.

Part five steps beyond scores to consider real-world limitations, reliability, and practical model behavior.

The final benchmark essay looks toward better evaluation methods that test usefulness rather than leaderboard theater.