Meta unveils Llama 3 and real-time image generator; see details

Meta Platforms, formerly known as Facebook, has taken a significant stride in the field of generative AI with the release of its latest large language model, Llama 3, and a real-time image generator. The move comes as the tech giant aims to close the gap with market leader OpenAI in the rapidly evolving field of artificial intelligence.

The newly unveiled models are set to be integrated into Meta's virtual assistant, Meta AI, which the company touts as the most advanced among its free-to-use counterparts. Meta AI promises enhanced performance across various domains, including reasoning, coding, and creative writing, putting it in direct competition with offerings from tech giants like Google and emerging players such as Mistral AI.

To bolster its position in the AI market, Meta plans to give its updated Meta AI assistant greater prominence within its suite of apps, including Facebook, Instagram, WhatsApp, and Messenger. Additionally, a standalone website will showcase Meta AI, positioning it as a formidable competitor to Microsoft-backed ChatGPT.

Visitors to the new website are greeted with a range of interactive features, inviting them to engage with the assistant in tasks such as creating vacation packing lists, playing music trivia, receiving homework assistance, and generating artwork of iconic city skylines.

Meta's aggressive push into generative AI represents a significant investment in computing infrastructure and the consolidation of research and product teams. By openly releasing its Llama models for developer use, Meta aims to disrupt the market and thwart rivals' attempts to monetise proprietary technology.

Llama 3 boasts new capabilities, including enhanced computer coding skills and the ability to process both text and images, albeit limited to text output for now. Meta plans to introduce more advanced features, such as multimodal capabilities for generating text and images simultaneously, in future iterations of the model.

The incorporation of image data into Llama 3's training holds promise for Meta's upcoming product, the Ray-Ban Meta smart glasses, enabling the AI assistant to identify objects and provide relevant information to users.

In a bid to enhance search capabilities, Meta announced a partnership with Google to integrate real-time search results into the Meta AI assistant's responses, complementing its existing collaboration with Microsoft's Bing search engine.

Meta CEO Mark Zuckerberg acknowledged the competition with OpenAI, hailing Meta AI as "the most intelligent AI assistant that you can freely use." Initial benchmarks suggest promising performance for Llama 3, with smaller versions showcasing competitive results against other free models.

However, concerns persist regarding the quality and contextual understanding of AI models, with previous iterations encountering issues with nuanced language comprehension. Meta asserts that Llama 3 addresses these concerns by leveraging high-quality data and significantly increasing training data volume.