Meta’s AGI Ambitions: FAIR Aims to Rival Open Source AGI

In a strategic move, Meta is stepping into the forefront of Artificial Intelligence (AI) research with its Fundamental AI Research (FAIR) team. The company, led by Mark Zuckerberg, has set its sights on developing Artificial General Intelligence (AGI), a concept often described as human-level AI.

Meta’s Vision for AGI

Zuckerberg, in a recent Instagram Reel announcement, provided insights into Meta’s long-term vision. The aim is not just to create advanced AI systems but to delve into the realm of AGI, where machines exhibit intelligence comparable to or exceeding human capabilities.

View this post on Instagram

A post shared by Mark Zuckerberg (@zuck)

The journey towards AGI is a gradual process for Meta, and Zuckerberg doesn’t define it with a specific threshold or a singular moment. Instead, he envisions AGI evolving over time, emphasising the breadth of intelligence, including reasoning and intuition.

FAIR’s Role in AGI Development

At the core of Meta’s AGI initiative is its AI research group, FAIR. This team, dedicated to fundamental AI research, is undergoing a strategic shift, being moved closer to the unit responsible for developing generative AI products across Meta’s various applications. The goal is to facilitate breakthroughs that can directly impact the billions of users within Meta’s ecosystem.

While the industry is witnessing intense competition for AI talent, Meta is taking a proactive approach with FAIR. The company is not only investing in attracting top-tier researchers but also amassing significant computing power. By the end of the year, Meta plans to own more than 340,000 of Nvidia’s H100 GPUs, a critical component in building generative AI.

Meta’s Stance on Open Source AGI

What sets Meta apart in the AGI race is its commitment to openness. Zuckerberg emphasises the importance of open-sourcing AGI responsibly when funding FAIR. This approach, according to him, addresses concerns related to unequal access to opportunities and value. In a landscape where some companies are moving away from open-source models, Meta’s stand aligns with a vision that fosters collaboration and widespread benefit.

$META Zuck just uploaded an AI progress update video to Threads, here are the major bullet points

• Merging two internal AI teams (FAIR and GenAI)
• Will have ~350k Nvidia H100s by end of year – in addition to others
• Doubling down on open source & wearable technology pic.twitter.com/ASyUbggou4

— Stocktwits (@Stocktwits) January 18, 2024

Zuckerberg draws a subtle contrast with other approaches, particularly mentioning the trend of companies starting off as open but later deciding to keep valuable AI developments private in FAIR. He suggests that Meta’s strategy of openness serves not only safety and responsibility concerns but also aligns with a broader vision of equal access.

AGI in the Context of Meta’s Larger Strategy

Despite the emphasis on AGI, Zuckerberg clarifies that this move does not signify a pivot for Meta. FAIR’s recent focus on the metaverse and Augmented Reality (AR) through products like the Ray-Ban Meta smart glasses remain intact. Zuckerberg envisions a future where AI and humans interact seamlessly, emphasising that AI will become an integral part of daily life, especially through smart glasses.

@robotshowers

Meta, driven by Mark Zuckerberg, is taking a giant leap in the AI race. Today, Zuckerberg revealed plans to build and open source their own Artificial General Intelligence (AGI), aiming to achieve or surpass human intelligence in most fields. This ambition is set to transform chatbot interactions, creator tools, and business applications, pushing AI’s boundaries further. In a recent video, Zuckerberg emphasized the need for advanced AI to develop the next generation of services. He envisions a future where Meta’s AI, open-sourced and widely available, can be a boon for everyone. The company’s commitment to this vision is evident in their aim to amass about 350,000 Nvidia AI chips by year-end, showcasing their serious investment in processing power. The strategy includes consolidating Meta’s two major AI research groups, FAIR and GenAI, to accelerate progress. Zuckerberg links this AI vision closely with the metaverse, predicting that by the decade’s end, many will frequently interact with AIs through smart glasses, like the Ray Ban Meta glasses, which already incorporate AI for hands-free communication and media capture. This announcement marks one of Meta’s most significant commitments to AI, at a time when tech giants like Microsoft, Google, and Amazon are intensifying the AI arms race with their innovative tools and visions. But with great power comes great responsibility, and concerns about potential unintended harms from these AI advancements are voiced by some tech skeptics. Earlier this year, Zuckerberg announced a new top-level product group at Meta focused on turbocharging AI tool development. Since then, Meta has released tools and resources to help users understand AI’s impact on their app experiences.

♬ you not the same – tilekid

Meta’s commitment to the metaverse is evident in ongoing initiatives, such as the development of Llama 3, the next-generation AI language model. With capabilities extending to code generation and advanced reasoning in FAIR, Llama 3 reflects Meta’s ambition to lead in AI model development.

Also Read: Elevating Australia’s 2024 Outlook with the Top 10 Anticipated Tech Trends

Industry Dynamics and Meta’s Competitive Edge

The AGI announcement comes at a time when major players in the tech industry, such as OpenAI and Google, are also pursuing AGI. Meta’s aggressive investment in both AI talent and computing power positions the company as a formidable competitor in the evolving landscape of AI research.

Zuckerberg acknowledges the fierce competition for AI talent, describing it as more intense than ever. Meta’s ability to attract top researchers and its substantial investment in AI infrastructure, including a massive number of NVIDIA H100 GPUs, signal the company’s determination to stay at the forefront of AI advancements.

Meta FAIR’s Integration and Impact

The integration of FAIR with the team working on generative AI products clearly is a strategic move to ensure that breakthroughs in AI research directly translate into improvements for Meta’s vast user base. By bridging the gap between fundamental research and practical applications, Meta aims to push the boundaries of what AI can achieve within its ecosystem.

$META Mark Zuckerberg provided a big AI update.

🧠 Aiming to create AGI.
🤝 FAIR & AI product teams merge.
🔓 Llama 3 will remain open-source.
🤖 Buying over 340K $NVDA H100 GPUs.
🕶️ AI Hardware is coming (smart glasses). pic.twitter.com/yd6dK3cDCg

— App Economy Insights (@EconomyApp) January 18, 2024

Yann LeCun, Meta’s Chief AI Scientist, expresses optimism about the impact of these changes, suggesting that the integration will accelerate progress in AI. FAIR is positioned as a sister organisation to GenAI, the product-focused division, creating a synergy that LeCun believes is essential for advancing AI research to the next level.

AGI and Open Source Alliance

Meta’s advocacy for open-source AI is not an isolated effort. The company, alongside IBM, recently launched the AI Alliance, a group promoting their vision of open-source AI. This move places Meta in opposition to some industry players advocating for more closed approaches to AI development.

The alliance’s objective aligns with Zuckerberg’s belief that openness addresses concerns related to the concentration of valuable AI technology. By releasing AI models openly, Meta aims to contribute to a collaborative environment where advancements are shared for the benefit of a broader audience.

Our Final Say: Meta Is Stepping Up the AI Race

At the moment, no one really knows what a perfect AGI will look like when it’s ready in the coming years. For a very long time, a utopian vision of having a human-like AI has been envisioned in several science fiction content, such as Iron Man’s Jarvis. With the recent success of AI for different purposes, possibilities of something like this have become brighter than ever before. 

The caveat is the lack of trust on the Big Tech to drive such innovations. Open source seems like a noble concept in a more democratic version of the internet but can the likes of Facebook and others be really trusted with your personal data? We’ve already seen them leveraging grey areas in regulations to milk user data around the world. Moreover, Big Tech has a monopoly on almost all aspects of consumer tech – hardware and software. With AGIs becoming concentrated to these firms, it may pretty much be old wine in a new bottle. 

As far as Meta goes, it owns the world’s largest social networking platform Facebook, Instagram, and WhatsApp. The company always enjoys a massive user base and possesses the power to push mass adoption of anything new instantly. If done right, Zuckerberg may soon have another massive hit under his regime. But we’re not sure how this will pan out with all his metaverse pivots and subsequent products and services as they still remain niche and novelty for many so be sure to stay connected by following us on our website for first-hand updates on Meta’s FAIR story and the unfolding chapters of human-machine capabilities.