Impossibility and Uncertainty Theorems in AI Value Alignment (or why your AGI should not have a utility function) | Read Paper on Bytez