Mark Zuckerberg has invested heavily in AI through Meta Platforms, with several important initiatives and business investments.
In March 2024, Mark Zuckerberg personally called AI researchers at Google’s DeepMind for a recruitment at Meta. This is how serious the company was in developing its AI strategy. For them, Meta used very unusual hiring practices; offering roles without formal interviews, and inflating salary offers to attract top talent.
In July 2024, Meta made an open-source approach to AI by launching AI models such as LLaMA, for Large Language Model Meta AI. This strategy was meant to make powerful AI tools more accessible. It would pit Meta against well-consolidated players like OpenAI and Google.
In June 2024, Meta announced new AI-powered capabilities for WhatsApp to improve the efficiency of businesses. The updates put a lot of focus on automating customer interactions and giving businesses more advanced tools to manage communications, including building better chatbots and AI-driven customer service. These innovations are targeted at simplifying operations and helping businesses “get business done” more effectively by making WhatsApp the core of all communication means. The new AI features further underscore Meta’s continued commitment to embedding AI within its popular services.
Mark Zuckerberg announced Meta’s AI plans in January 2024 by detailing how the company intends to buy about 600,000 GPUs through the end of the year, including about 350,000 Nvidia H100 GPUs. That will be a base of infrastructure for AI services and models being developed, including those aimed at General Artificial Intelligence capabilities.
In September 2023, Meta introduced the latest Ray-Ban-Meta Smart Glasses. These smart glasses combine fashion and technology: it’s possible to shoot photos and videos, broadcast events in real-time, and use all of Meta’s AI capabilities hands-free. With improvements to audio and camera, better integration with Meta’s ecosystem boosts augmented reality.
In September 2023, the Chan Zuckerberg Initiative launched an ambitious AI effort for progressing biomedical insights. This project focuses on using AI to enhance cellular research for quicker discoveries in biology and medicine. CZI seeks to change just how scientists comprehend cellular functions and diseases by combining AI-driven tools and data, with the intent of fostering innovative solutions to accelerate health improvements.
In August 2023, Meta released Code Llama, an AI tool developed to enhance coding workflows. This is a large language model tool that uses text-based prompts for code generation and discussion. It supports languages such as Java, PHP, C#, and Python and is free to use under the Llama2 community license for research, personal, and commercial applications. Code Llama Code completion or debugging. Available in three sizes: 7B, 13B, and 34B parameters. The 34B model yields better code assistance but is more GPU-intensive.
Imagine with Meta AI, in December 2023, Meta announced several new AI features, one of which was “Imagine with Meta AI,” a generative AI model for creating images. Trained on 1.1 billion public images from Facebook and Instagram, this lets users create images from text prompts. Available first on Meta’s chat apps, its popularity moved it out to a standalone website: imagine.meta.com. The tool generates four images, watermark included, visible at 1280×1280 pixels, to ensure transparency in the images it makes. In addition, Meta plans to add an invisible watermark for further improvements in traceability for AI-generated images.
In December 2023, Meta and IBM announced an AI alliance that would help accelerate artificial intelligence development. What the collaboration does is bring both companies together to realize the potential of large-scale AI model development and make the AI infrastructure better. This deal will bring together Meta’s experience with AI models such as Llama and combine it with IBM’s computing and enterprise solutions to power innovation into a set of emerging areas, including natural language processing and enterprise AI solutions. The alliance is expected to improve AI-powered services for businesses while increasing Meta’s leadership in AI technology.
In April 2022, Meta acquired Presize.ai, a startup dealing in AI-powered body measurement solutions for online apparel retailers. Presize.ai leverages computer vision and deep learning by having users find the right size of clothing through the analysis of videos of themselves for the perfect fit reducing returns. This acquisition by Meta is expected to fall in line with the company’s efforts to enhance its e-commerce and AI capabilities and support more personalized shopping experiences across its ecosystem.