LLMs as Meaning Optimization | Dark Hacker News