During an event held in London on Tuesday, Meta announced its upcoming launch of Llama 3, the next iteration of its large language model designed to power generative AI assistants. This confirmation aligns with a report by The Information published the day prior, indicating Meta’s nearing launch.
Nick Clegg, Meta’s president of global affairs, disclosed the imminent release of Llama 3, slated for within the next month. He hinted at the introduction of various versions of the product, each equipped with distinct capabilities and features, expected to roll out throughout the year.
Chief Product Officer Chris Cox outlined Meta’s strategy to integrate Llama 3 across multiple products within the Meta ecosystem, aiming for broader utility and functionality.
Playing catch-up to OpenAI, Meta has faced pressure to enhance its AI capabilities, particularly after the viral success of ChatGPT. Previous iterations of Meta’s AI assistant received criticism for their limitations, prompting the development of Llama 3 to offer improved accuracy and versatility in handling a wider array of inquiries, including potentially contentious topics.
Joelle Pineau, vice president of AI Research at Meta, expressed ambitions for Llama-powered AI to become the world’s most valuable assistant, acknowledging the ongoing refinement required to achieve this goal. Although Meta did not delve into technical specifics or provide demonstrations, Llama 3 is anticipated to boast approximately 140 billion parameters, doubling its predecessor’s capacity.
Meta Pushes For More With AI
Meta’s commitment to open-source development with its Llama models reflects a distinct philosophy toward AI advancement, aiming to garner favor among developers while fostering innovation.
Despite Meta’s strides in text-based generative AI, the company remains cautious about ventures into other domains, withholding the release of Emu, its image generation tool, due to concerns regarding latency, safety, and user experience.
However, amidst Meta’s push for Llama 3, notable skepticism persists within the company, notably from Yann LeCun, Meta’s chief AI scientist. LeCun champions a different paradigm, predicting the ascendancy of joint embedding predicting architecture (JEPA) over generative AI, advocating for a shift in focus within Meta’s AI product division.
As Meta navigates the evolving landscape of AI technology, the release of Llama 3 signifies a pivotal step forward in its quest for AI innovation and utility.
See also: Symbolica Brings New Strategy: Mitigating The AI Arms Race With Symbolic Models
