Just two weeks before the release of GPT-4, OpenAI CEO Sam Altman stated that the company has not yet achieved true Artificial General Intelligence (AGI), which is the ultimate goal for OpenAI. AGI refers to an AI system with capabilities at least on par with humans across a wide range of fields. Current versions of ChatGPT, such as GPT-4, have not yet reached this level of capability, leading Altman to declare that true AGI remains unattained.
However, recent rumors suggest that the upcoming GPT-5 model may have the potential to achieve AGI.
**Note: The following information is based on rumors
GPT-5 to Complete Training by December 2023
Runway CEO Siqi Chen, whose startup company focuses on machine learning and artificial intelligence research, recently shared a message claiming that GPT-5 is scheduled to complete its training by December 2023 and is expected to achieve the AGI that OpenAI has been striving for. This has led to heated discussions on whether AGI has truly been realized, with the consensus leaning towards affirmation.
The news has garnered significant attention, although AI technology is not without its flaws. Consequently, discussions around the completion of GPT-5’s training in December have been overshadowed by debates on the topic of AGI.
Chen later clarified his statement, emphasizing that he was not claiming that achieving AGI through GPT-5 is the consensus within OpenAI, but rather that there may be individuals within the organization who believe AGI has been realized.
At this point, it is impossible to determine the accuracy of these rumors. Considering GPT-4’s performance, the rate of iteration may continue to increase, with each iteration potentially resulting in significant breakthroughs. As such, the realization of AGI remains a possibility.