Imagine designing a computer chip not by writing code, but by simply describing what you want in everyday language. That's the groundbreaking idea explored by researchers at Johns Hopkins University, who used OpenAI's ChatGPT to design a recurrent spiking neural network (RSNN) chip. Why spiking neural networks? These networks mimic the way the human brain works, offering the potential for smaller, faster, and more energy-efficient AI. The team used a conversational approach with ChatGPT, breaking down the complex RSNN design into smaller, manageable modules. They described each module in plain English, and ChatGPT translated these instructions into Verilog, the standard hardware description language. This process wasn't without its challenges. The researchers had to refine their prompts and troubleshoot errors through numerous iterations, highlighting the need for clear communication even with advanced AI. But the results were remarkable. ChatGPT successfully generated the Verilog code for the entire RSNN, which was then validated on a field-programmable gate array (FPGA) and even fabricated into a physical chip using SkyWater 130nm technology. The chip performed impressively in tests, achieving high accuracy in tasks like image classification. This research opens exciting possibilities for the future of chip design. Imagine a world where anyone, regardless of technical expertise, could design custom hardware simply by describing their needs. While challenges remain, this work demonstrates the potential of AI to democratize hardware development and accelerate innovation.