META Platforms Introduces Llama 3 and Image Generator

On Thursday, April 18th, Meta Platforms introduced early editions of its latest large language model, named Llama 3, along with an up-to-date image generator. The image generator instantly alters pictures in response to user commands as Meta strives to challenge the leading generative AI company, OpenAI.

Meta AI, the state-of-the-art virtual assistant offered for free by the company, is being promoted as more advanced than its competitors’ versions, and this upgraded model will be made available to users.

The assistant will have a more confrontational encounter with ChatGPT, the popular AI chatbot supported by Microsoft, due to a separate website and increased visibility in Meta’s apps, including Facebook, Instagram, WhatsApp, and Messenger.

Meta is working diligently to compete with OpenAI’s leading position in generative AI by making the technology accessible to its vast user base of over a billion people. This initiative involves revamping the computational resources and combining previously separate research and development groups.

Chris Cox, Meta’s Chief Product Officer, mentioned in an interview that Llama received three new programming abilities from Meta and was provided with images as well as text this time around. Yet, presently, the model can only generate text outputs.

In the future, later editions will boast advanced reasoning skills, enabling them to construct intricate multi-phase plans. As indicated in Meta’s blog posts, forthcoming releases will additionally offer “multimodal” capabilities, meaning they can generate both written content and visual graphics.

Cox explained that the ultimate objective is to alleviate burdens and simplify life in various aspects, be it dealing with businesses, composing texts, or arranging travel plans.

Cox suggests that incorporating images while training Llama 3 for this year’s Ray-Ban Meta smart glasses update, a collaboration with Essilor Luxoticca, will enhance the glasses’ capabilities. This upgrade enables Meta’s AI to identify objects viewed through the lenses and answers queries regarding them.

Meta has teamed up with Google to enhance its collaboration with Microsoft’s Bing, allowing for the direct inclusion of real-time search results in the Meta AI assistant’s responses. This partnership comes as Meta’s AI assistant prepares to enter more than a dozen markets beyond the US, although the company faces ongoing challenges in complying with European privacy laws.

Meta makes available for free commercial use models like Llama 3 for developers, including its advanced AI assistant, Meta AI, which Zuckerberg described as “the most intelligent you can freely use” in a promotional video, implying intense competition. However, this approach brings up safety concerns.

Read More

Sorry. No data so far.

2024-04-18 21:40