In a recent gathering of press and executives on Zoom, Yann LeCun, Chief AI Scientist at Meta, stated that in terms of underlying techniques, ChatGPT, developed by OpenAI, is not particularly innovative. He emphasized that while the technology may be perceived as revolutionary by the public, it is simply well-constructed and not unique in comparison to similar systems developed by other companies and research labs.

LeCun also noted that there are currently several startups that possess similar technology to ChatGPT. Furthermore, he highlighted that the program is made up of multiple pieces of technology developed over many years by various parties and is not a breakthrough achievement. LeCun also emphasized that the use of Transformer architectures and self-supervised learning in ChatGPT, which are fundamental components of the program, are not new concepts and have been advocated for and utilized by other parties prior to OpenAI’s emergence.

In terms of objectivity and wider context, research like this has indeed been going on for many years. However, that research has not been released to the public in an easily accessible manner, or freely (for now) and on such a large scale. Scientists such as Yann LeCun who have had years of experience in the field of AI may not be overly impressed by ChatGPT, however, reading stories in the press and the sheer volume of YouTube videos on the topic makes it feel like the public are extremely impressed.