OK, what's the difference between coherence and representation theorems?

A debate explores the distinction between coherence and representation theorems in rational decision theory. Coherence theorems show that irrational preferences lead to predictable losses or exploitation, like in 'money pump' arguments. Representation theorems show rational preferences can be modeled as maximizing expected utility. The discussion uses Savage's theorem and the Complete Class Theorem as examples, highlighting a nuanced and often blurred line between these foundational concepts in philosophy and AI safety.