After OpenAI, Google and Musk’s company xAI, Meta boss Mark Zuckerberg has also announced plans to develop artificial general intelligence (AGI). The current boom in artificial intelligence is essentially based on more specific language or image models.
On Thursday, in a video on the short messaging service Threads, Zuckerberg said, “Developing the best AI assistants, AI for creatives, AI for business, and so much more requires advancements in all areas of AI, from thinking to planning, scheduling” to memory and other cognitive skills.” This requires “comprehensive general AI.”
Zuckerberg doesn’t get into specifics anymore. The question of when his company will introduce such an AGI to the market also remains open.
However, it is controversial among critics how seriously Meta is pursuing the approach of opening up its model to everyone. In the terms of use, Facebook’s parent company restricts the commercial use of its Llama 2 model for services with more than 700 million active users and excludes various areas of application. This is unusual for an open-source approach.
Zuckerberg’s announcement is likely to put pressure on OpenAI in particular. His first major goal was not actually the development of the successful chatbot ChatGPT, but rather the development of the first AGI.
At the World Economic Forum in Davos, OpenAI chief Sam Altman remained as vague as Zuckerberg about his plans and the current state of development. An AGI will be more powerful than humans in many areas, he said.
We’ll see how it turns out.