OpenAI and Meta Preparations are said to be underway for the release of more advanced AI models that will be able to help solve problems and take on more complex tasks, with OpenAI's COO Brad Lightcap telling the Financial Times that the company's next version of the GPT will make advances in solving "difficult problems," such as inference.
Source Note: The image is generated by AI, and the image is authorized by Midjourney
Meta executives also said the company's upcoming Llama3 model is on a similar development trajectory and is expected to be launched in the coming weeks. Joelle Pineau, Meta's vice president of AI research, said the company is working to figure out how to make the model communicate, reason, plan and have memory. Representatives from Meta and OpenAI did not immediately respond to Business Insider's request for comment.
Giving AI models the ability to reason and plan is the key to achieving artificial general intelligence (AGI).AGIIn a major step forward for AI, both Meta and OpenAI have claimed that their goal is to realize AGI, a development that could be worth trillions of dollars to the companies that achieve it. In February, former Meta executive and virtual reality pioneer John Carmack called Artificial General Intelligence (AGI) the "big gold ring" of AI and predicted it would be a trillion-dollar industry by the 2030s. While definitions of AGI may vary, the simplest definition is a type of artificial intelligence that can perform at a human level or higher on a variety of tasks.
Some experts have raised safety concerns about developing technologies that surpass human intelligence. Leading researchers, including AI godfathers Yoshua Bengio and Geoffrey Hinton, have urged people to consider the risks of AI to humans. Elon Musk, a longtime AI skeptic, recently estimated that AI will surpass humans within two years. Musk said that the "total amount of sentient computation" -- a concept that may refer to AI thinking and acting independently -- will surpass all humans within five years.