OpenAI CEO Sam Altman confirmed the rumors about the development status of GPT-5, the widely believed upcoming version of the immensely popular GPT series. Since the release of the GPT-based ChatGPT chatbot, the language model has become the fastest-adopted technology in recent times.
GPT-3, the earlier version of the language model released on November 30, 2022, was considered an effective achievement in natural language processing and artificial intelligence. GPT-3 is frequently utilized in various applications such as chatbots, language translation, and content creation.
During a discussion at an MIT event, Altman said work on GPT-5 had not yet begun. He said, “I think we’re at the end of the era where it’s gonna be these giant models, and we’ll make them better in other ways”. As the capabilities and precision of such technologies increase, the safety measures must increase as well. He then explained how people are more focused on the next iteration and bigger parameter numbers like the gigahertz race in the 90s and 2000s.
The chart above shows the growth in the number of parameters in different models.
OpenAI has just released GPT-4 on March 14, 2023. The capabilities of this new generation are yet to be fully comprehended and utilized before putting the spotlight on the next generation. At the moment, OpenAI is focusing on expanding the capabilities of GPT-4, especially the safety measures, and squashing bugs for which they"re offering a $20,000 cash bounty.
For those who want to try the latest version of the GPT by OpenAI, they can use the Bing Chat feature.
Source: Amy Sterling (YouTube) via The Verge and TechCrunch