Meta Platforms has unveiled the early versions of its latest large language artificial intelligence (AI) model, Llama 3, and an accompanying real-time image generator, as part of its efforts to beat rivals like OpenAI in the generative AI market.
Llama 3 boasts enhanced capabilities, including computer coding and image processing, with plans for future versions to support multimodal outputs. Meta’s Chief Product Officer, Chris Cox, noted that the company’s long-term goal has always been to streamline users’ tasks, from interacting with businesses to planning trips.
Meta plans to integrate these models into its Meta AI virtual assistant, which CEO, Mark Zuckerberg, has touted as “the most intelligent AI assistant that you can freely use.” The updated AI assistant will receive heightened visibility within its Facebook, Instagram, WhatsApp, and Messenger apps, alongside a dedicated website aimed at challenging Microsoft-backed OpenAI’s ChatGPT.
On the site’s landing page, users will be able to explore various tasks such as creating vacation packing lists, seeking homework assistance, and generating artwork of the New York City skyline.
Meta’s push for generative AI products has seen the tech giant invest heavily in computing infrastructure and the merging of research and product teams. The company has been openly releasing its Llama models to developers as part of its efforts to catch up with OpenAI, though this has sparked safety concerns regarding potential misuse.
Moreover, Meta has announced a partnership with Google to incorporate real-time search results into its assistant’s responses, expanding its reach to over a dozen international markets. However, challenges lie ahead, particularly in Europe, where stricter privacy regulations and the impending AI Act pose compliance hurdles.
