OpenAI's World of Bits (WoB) is a research project aimed at training AI agents to interact with computer interfaces using a keyboard and mouse, just like humans do. The goal is to enable AI to complete web-based tasks such as form filling, information retrieval, and website navigation. Initially developed in 2016, WoB was considered ahead of its time. However, with advancements in large language models like GPT, the project has gained renewed interest. The concept involves giving neural networks access to screen pixels as input and allowing them to control keyboard and mouse actions as output. This approach could potentially lead to AI assistants capable of performing complex digital tasks autonomously, bridging the gap between AI capabilities and human-computer interaction.