The article discusses the promise and peril of large language models (LLMs) in artificial intelligence (AI). LLMs, such as ChatGPT, can converse cogently about various topics and have aced exams that serve as gateways for people to enter careers in law and medicine in America. However, there is deep concern that generative AI models are being developed too quickly, and prominent voices are calling for the development of artificial intelligence to be paused. LLMs are a giant exercise in statistics, and their outputs are not entirely predictable. The recent success of LLMs in generating convincing text is due to the coalescence of three things: gobsmacking quantities of data, algorithms capable of learning from them, and the computational power to do so. However, there are good reasons to think that the growth of LLMs cannot continue indefinitely, and the amount of training data available is limited. Despite the risks, having the power of LLMs in many hands means that many minds can come up with innovative new applications, improving everything from medicine to the law.
