Rumor: The next version of ChatGPT will reach a previously unseen level of AI
More recently, OpenAI released the fourth generation of the GPT language model at the heart of Microsoft’s ChatGPT chatbot and Bing search, but discussions have begun on the next version. Judging by the rumors, we are waiting for a real revolution.
According to available information, OpenAI plans to complete the training of the GPT-5 neural network by December this year. At the same time, the company allegedly hopes to achieve AGI (Artificial General/Strong Intelligence). The concept of so-called general-level artificial intelligence provides for the ability of a program to think and act like a person.
For comparison, existing neural networks can be described as “weak AI” (Narrow AI), since they are designed to perform and automate some kind of cognitive function with the ability to do it like a person or even better. In the case of “strong AI” (General AI), we are talking about the ability of the program to independently draw conclusions based on incoming information and perform most of the tasks that a person is capable of. There is also a theory that such an AI, as it learns, is able to realize itself. In theory, then comes Artificial Super Intelligence (ASI), which surpasses the intellectual capabilities of any of the people and is able to instantly solve problems of any complexity.
By the way, OpenAI has already announced GPT-4.5 , promising to release an interim version in September or October 2023. The developers promise to address some of the limitations of GPT-4 by improving performance and expanding the way it can be used. In addition, GPT-4.5 will be able to produce longer, more precise, and more connected text responses.