Nvidia CEO Jensen Huang recently made headlines by personally delivering the world’s first DGX H200 to OpenAI, a leading AI research organization based in the US. The DGX H200 is an advanced integrated AI platform that is powered by the most powerful GPU currently available, the Nvidia H200. It serves as the successor to the highly acclaimed DGX 100 system, offering a comprehensive hardware and software solution for tasks such as generative AI, natural language processing, and deep learning recommendation models.
The hardware was warmly received at OpenAI by President Greg Brockman and CEO Sam Altman. OpenAI is widely recognized for its efforts in developing safe artificial general intelligence (AGI) with the aim of benefiting all individuals globally. The organization has already achieved notable success with projects like GPT-3 and ChatGPT, which are large language models that have revolutionized the field of artificial intelligence.
One of the most striking aspects of this news is the training of GPT-4 on approximately 25,000 Nvidia A100 GPUs over a period of roughly 100 days, costing between $2 to $3 million US dollars. With the new DGX H200, OpenAI is gearing up for the training and development of GPT-5, which is scheduled for release in mid-2024. GPT-5 is anticipated to surpass its predecessor GPT-4, leading to speculations about its official name. There is a possibility of an interim version like GPT-4.5 before the launch of GPT-5, which could be an upgraded version of GPT-4 Turbo.
The upcoming GPT-5 is poised to be a multimodal model capable of processing both visual and text inputs and outputs. It is expected to combine various AI capabilities such as natural language processing, image recognition, sensor data interpretation, pattern recognition, predictions, abstraction, and reasoning. This fusion of technologies aims to create an AI system with human-like capabilities, enabling it to perform tasks with dexterity. Popular AI tools like Siri, Alexa, Watson, and ChatGPT fall under the category of Artificial Narrow Intelligences (ANIs), highlighting the advanced level of development in the field.
In conclusion, Nvidia CEO Jensen Huang’s delivery of the DGX H200 to OpenAI marks a significant milestone in the realm of artificial intelligence. The collaboration between these two entities promises groundbreaking advancements in AI research and development, paving the way for the emergence of sophisticated AGI models like GPT-5 in the near future. Stay tuned for more updates on the latest innovations shaping the AI landscape.
