Summary

This talk explores the current "boring" state of Large Language Models (LLMs) despite their initial sparkle, focusing on tool calling. The speaker references their experience with Common Lisp, embedded engineering, and the evolution of LLM usage since Copilot and ChatGPT. The practical takeaway is that LLMs, through tool calling, can abstract away complex schema parsing and validation, allowing for a more natural, conversational interaction with machines to achieve desired outcomes.