This interview podcast features Terry Sejnowski, author of "ChatGPT and the Future of AI," discussing large language models (LLMs) and their implications. The conversation begins by addressing listener concerns about job displacement due to AI, with Sejnowski emphasizing that jobs will change rather than disappear, requiring adaptation and new skills in utilizing AI tools. The discussion then delves into the inner workings of LLMs, their lack of true memory, and the concept of "in-context learning," comparing and contrasting them with human cognitive processes. Finally, the interview explores the potential for LLMs to achieve consciousness and the exciting possibilities of larger neurofoundational models (LNMs) that could potentially simulate brain activity and even lead to a form of digital immortality. A specific example highlights how LLMs, combined with human expertise, can significantly improve diagnostic accuracy in medicine, illustrating the potential for human-AI collaboration.
