ChatGPT/GPT-3 is a powerful language model, designed to perform language-based tasks, but it is not an artificial general intelligence (AGI). Despite this, it is transforming the way we communicate with machines and is poised to become a crucial bridge between humans and machines, as well as machines to machines.
Language itself is not equivalent to intelligence, but it serves as a universal glue that connects knowledge across different disciplines, allowing individuals and groups to operate effectively as a society.
Language is a Universal Glue
As Bing connects search engines to language models, and users connect ChatGPT’s output to StableDiffusion, we can see the potential for a future in which e-commerce apps allow users to search and purchase items via natural language conversations, among other possibilities.
This interconnected network of apps, expert systems, and other AI tools will operate much like human society, with different components working together. While it may not be considered an AGI in the strictest sense, it will undoubtedly be able to handle a wide range of complex problems and operate independently of any single organization, much like the internet itself.
General Interoperability Across Humans and Machines
To truly understand the future possibilities of these models, a deeper understanding of their potential roles is necessary. As engineers work towards developing more powerful models, they are acutely aware of the current limitations and areas for improvement. It is important to view ChatGPT as just one component in the wider context of advanced language models and their potential impact on the field of artificial intelligence.
(Original ideas by Jacky Ma, edited by ChatGPT)