Meta CEO Mark Zuckerberg said on Jan. 18 that the company was bringing its business-focused generative artificial intelligence (AI) research team “closer together” with its Fundamental AI Research (FAIR) team and doubling down on a push to get the technology into products.
Appearing in a Thread video, Zuckerberg announced the change to the company’s AI efforts, including significant investments in specialized computer chips to build and deliver new generative AI models and products. He also said the company had begun training its next-generation Llama 3 large language model (LLM):
“It’s become clear that the next generation of services requires building full general intelligence. Building the best AI assistants — AIs for creators, AIs for businesses and more — that needs advances in every area of AI, from reasoning to planning to coding to memory and other cognitive abilities.”
Zuckerberg explained that to accommodate the push to get generative artificial intelligence into its products, Meta will increase its technology infrastructure. Meta plans to have about 350,000 H100 graphics processing units (GPUs) from chip designer Nvidia by the end of the year.
Meta is combining its two advanced AI research division efforts in a move similar to that of Alphabet’s in 2023, when it brought together its two advanced AI research labs, Google Brain and DeepMind.
Alphabet was trying to catch up to Microsoft and OpenAI, which had beaten Google to market with a highly capable AI chatbot in ChatGPT and made the world’s most potent large language model, GPT-4, available to customers.
Over the years, Meta has conducted significant AI research, covering unsupervised learning, where AI learns patterns without labeled data, to creating AI software that outperforms top humans in the strategy game Diplomacy. The company has also achieved advancements in machine translation and computer vision algorithms.
Meta’s GenAI team created Llama 2, a powerful open-source language model. While not as advanced as OpenAI’s GPT-4 or Google’s Gemini models, Llama 2 is favored by developers for building cost-effective and customizable chatbots compared to using other models.