Bounded Rationality for LLMs: Satisficing Alignment at Inference-Time | Read Paper on Bytez