In a surprising announcement, Mark Zuckerberg, the CEO of Meta Platforms, Inc. (formerly Facebook), took to Instagram Reels to unveil a significant development in the world of artificial intelligence (AI). Meta is making a bold move into Artificial General Intelligence (AGI) with a commitment to develop it as open source. This means merging two of Meta’s prominent AI research teams, FAIR (Facebook AI Research) and GenAI, to collectively work on creating the general intelligence system. 

Zuckerberg’s announcement

Zuckerberg’s announcement came as a shock to many. This move represents a significant shift in the approach to AGI development. Traditionally, companies have pursued AI advancements in a competitive and often secretive manner. Meta’s vision is based around openness and collaboration.

Vision for open source AGI

Zuckerberg detailed a long-term vision where Meta intends to build general intelligence and open source it responsibly for the benefit of humanity. He emphasized the need for advancements across all facets of AI, including reasoning, planning, coding, memory, and other cognitive abilities. 

The emphasis on comprehensive AI development clearly demonstrates the importance of creating AI systems that are highly capable and ethically responsible. This initiative meets the broader industry trend towards responsible AI, focusing on issues like bias mitigation, transparency, and accountability.

Infrastructure and development

To achieve these ambitious goals, Meta is investing heavily in infrastructure and development. They are currently working on training Llama 3, an advanced AI model, and constructing a “massive compute infrastructure” to support the training and development of AGI.

One staggering statistic for Meta’s commitment is their plan to have 350,000 Nvidia H100 GPUs in operation by the end of the year. This massive computational power is indicative of the scale of investment required for AGI research and development.

Integration with the Metaverse and smart devices

Zuckerberg also discussed the integration of AGI with the metaverse, a virtual world that is gaining increasing prominence. The metaverse can be pictured as a shared digital space where people can interact, work, and socialize, and presents a natural environment for AI to thrive. 

To complement this, Meta is actively working on smart devices, such as their Ray-Ban smart glasses. These glasses will let AI see and hear what the user does, creating intuitive interactions between humans and AI. This development perfectly matches Meta’s broader goal of making AGI a ubiquitous presence in our daily lives.

Addressing concerns and skepticism

Following Zuckerberg’s remarks, OpenAI CEO Sam Altman at the World Economic Forum commented on the existential risks posed by AGI. These concerns are primarily based on the need for responsible AGI development and collaborative efforts, which Meta’s open source approach aims to address.

It’s worth noting that Meta’s chief scientist, Yann LeCun, had expressed skepticism about the near-term feasibility of AGI. This skepticism details the complexity and challenges involved in AGI development, making Meta’s ambitious endeavor all the more significant.

A bold step towards a shared future

Meta’s foray into open source AGI shows AGI is not the exclusive domain of tech giants but a shared resource for the benefit of all. This initiative reflects a commitment to responsible and comprehensive AI development, with a focus on transparency and collaboration. The combination of AGI within the metaverse and smart devices holds the promise of transforming how we interact with AI in our daily lives.

Alexander Procter

February 22, 2024

3 Min