OpenAI CEO Shares Reality of Upcoming GPT-4 Model
OpenAI CEO Sam Altman reveals that the upcoming GPT-4 model will not achieve Artificial General Intelligence and dismisses rumors of a massive 100 trillion parameter count.
OpenAI CEO Sam Altman sets realistic expectations for the upcoming GPT-4 model by confirming that it does not achieve Artificial General Intelligence (AGI). He also definitively dismisses circulating rumors that the new model contains 100 trillion parameters, steering the conversation away from pure size and toward functional improvements.
While Altman stops short of explicitly labeling GPT-4 as a multimodal model, he strongly hints that multimodal capabilities are arriving in the very near future. Unlike current OpenAI products that rely strictly on text interactions, a multimodal system processes and generates text, images, and audio, allowing users to converse with the AI using voice commands.
Altman envisions this multimodal evolution as a foundational technological platform that rivals the rise of the mobile ecosystem. He predicts that this natural language interface allows the computer to execute complex tasks simply through conversational dialogue, which ultimately sparks an explosion of entirely new business models and ventures.