I'm still not convinced by LLMs or chatbots - they're just glorified autocomplete tools that mimic human language without actually understanding it. Where's the substance?