Mark Zuckerberg announced on Thursday that Meta is planning to build its own artificial general intelligence (AGI) and release it as open-source (open AGI) software and make it freely available to the general public.
“Our long-term vision is to build general intelligence, open source it responsibly, and make it widely available so everyone can benefit,” he added.
Artificial general intelligence, commonly known as AGI, refers to a form of artificial intelligence that matches or exceeds human intelligence in nearly all aspects. It would have the capacity to learn, reason, adapt, and execute any intellectual task comparable to a human.
In a video shared on Meta’s social network Threads, Meta’s founder and CEO also emphasized that achieving the best AI for chatbots, creators, and businesses requires advancements in various aspects of AI.
“It’s become clearer that the next generation of services requires building full general intelligence,” he said in a personal video. “Building the best AI assistants, AIs for creators, AIs for business and more—that needs advances in every area of AI, from reasoning to planning to coding to memory and other cognitive abilities,” Zuckerberg said.
To support this ambitious initiative, Zuckerberg said Meta plans to have a substantial array of computing power in its cloud facilities by the end of 2024. He highlighted the goal of having around 350,000 Nvidia AI chips (Nvidia H100s) or approximately 600,000 H100 equivalents, including other GPUs, by the same timeframe.
“We’re building a massive amount of infrastructure. At the end of this year, we’ll have ~350k Nvidia H100s — and overall ~600k H100s H100 equivalents of compute if you include other GPUs,” Zuckerberg said in the video.
Zuckerberg’s massive AI compute capacity, when calculated, is expected to reach 4.8e+16 transistors, amounting to 48 quadrillions. With this computational power, Meta aims to continue training Llama 3 and pursue an exciting roadmap of future models responsibly and safely.
The company also intends to bring together its two major AI research groups, FAIR and GenAI, to accelerate its work in the field of artificial intelligence.
“We’re bringing Meta’s two major AI research groups (FAIR and GenAI) closer together and growing both teams to accelerate our work.”
At the same time, delving into artificial intelligence doesn’t imply that Zuckerberg has given up on his metaverse dream. He firmly believes that the two, AI and the metaverse, are intricately connected.
He clarified that embracing AI does not signal a departure from Meta’s metaverse ambition. He sees a connection between the two, envisioning that, by the end of the decade, many people will interact with AIs frequently throughout the day using smart glasses like those being developed with Ray Ban Meta.
“The two major parts of our vision—AI and the metaverse—are connected. By the end of the decade, I think lots of people will talk to AIs frequently throughout the day using smart glasses like what we’re building with Ray Ban Meta,” he said.
Meta’s latest Ray-Ban glasses, powered by artificial intelligence, enable users to make calls, send messages, and capture videos hands-free.
While the pursuit of AGI has been a persistent objective in AI research, it remains largely theoretical and comes with substantial technical and ethical hurdles. The company’s intentions to develop AGI also raise heightened concerns regarding potential misuse and unintended consequences.
Zuckerberg’s latest announcement reflects a significant commitment to doubling down on artificial intelligence, aligning with other major tech players like Microsoft, Google, and Amazon who continue to unveil new AI tools and visions. Despite the excitement, some tech skeptics express concerns about potential unintended harms resulting from these revolutionary products by big companies and newcomers like OpenAI.