Imagine an AI that doesn't just write poems or translate languages, but can actually design complex machinery like a jet engine. That's the tantalizing possibility hinted at by new research exploring the use of Large Language Models (LLMs) in energy and power engineering. Researchers are experimenting with LLMs to tackle a classic engineering challenge: gas path analysis for gas turbines. This involves intricate calculations to optimize the flow of air and fuel through the engine's components—compressor, combustion chamber, turbine, and nozzle—for maximum efficiency. The challenge? LLMs, despite their impressive language skills, aren't inherently good at math or physics. They can hallucinate information and struggle with complex reasoning chains required for engineering problems. To overcome this, researchers developed a clever "dual-agent" system. One agent, guided by a technique called "ReAct" (Reasoning and Acting), figures out which calculations are needed. The second agent then uses specialized tools—essentially pre-programmed physics equations—to perform those calculations. The results are promising, but not perfect. Smaller LLMs struggled to use the tools correctly, while larger models showed more aptitude. Even the most advanced models stumbled when faced with multi-component problems, highlighting the difficulty of translating real-world physics into AI instructions. However, with some human guidance, the LLMs showed potential to correctly solve these complex problems. This suggests that with further refinement, LLMs could become valuable assistants for engineers, helping them design and optimize complex systems like gas turbines. The future of AI-driven engineering is still taking shape, but this research offers a glimpse into how LLMs, combined with human expertise and physics-based tools, could revolutionize how we design and build the machines of tomorrow.