Summary:

- The article discusses the challenge of playing chess against large language models (LLMs) and the mystery surrounding "instruct models" - AI systems that can follow instructions and complete tasks.
- It highlights the impressive chess-playing abilities of LLMs, which can outperform even the best human players, and explores the potential implications of this technology.
- The article also delves into the inner workings of instruct models, which are trained to understand and follow natural language instructions, and how this capability could be leveraged in various applications.