Meta, the parent company of social media giants Facebook and Instagram, is taking a bold step into the world of artificial intelligence (AI) with the development of a new, open-source AI model designed to rival the most powerful systems from its competitor, OpenAI.
Sources familiar with the matter have informed The Wall Street Journal that Meta is actively working on this ambitious project. The goal is to create an AI model that surpasses the capabilities of its Llama 2 model, which was introduced earlier this year. Meta envisions this new AI model to be “several times” more powerful than its predecessor.
What makes Meta’s endeavor particularly noteworthy is its commitment to open sourcing the new AI system. This approach will allow other companies and developers to leverage the AI model for various applications, including generating high-level text and performing complex analyses.
To support this initiative, Meta has been diligently building the necessary data centers while securing an ample supply of Nvidia’s H100 semiconductor chips. These chips are among the most potent and sought-after components on the market.
Llama 2, Meta’s previous AI model, was trained using a dataset consisting of 70 billion parameters. In comparison, OpenAI’s GPT-4, which represents one of the industry benchmarks, is estimated to employ around 1.5 trillion parameters.
Meta anticipates commencing the training of this large language model (LLM) early in 2024, with a planned release sometime in the following year. Notably, Meta’s LLM project will likely launch after Google’s upcoming LLM known as Gemini.
While Meta collaborates with Microsoft, a major supporter of OpenAI, to offer Llama 2 on Microsoft’s Azure cloud platform, the company intends to take a self-reliant approach in training its forthcoming AI model on its infrastructure.
Meta’s ambitious AI project aligns with a broader trend in which major technology companies and governments worldwide are racing to develop, deploy, and assert control over high-level AI systems. Initiatives like the United Kingdom government’s $130 million investment in powerful AI chips and China’s legislation focusing on AI reflect the global importance of AI advancements.