Meta Is Engaged on a New, Extra Highly effective AI System to Rival OpenAI: Report

Meta Platforms is engaged on a brand new artificial-intelligence system supposed to be as highly effective as probably the most superior mannequin provided by OpenAI, the Wall Avenue Journal reported on Sunday, citing folks acquainted with the matter.

The Fb mother or father is aiming for its new AI mannequin to be prepared subsequent yr, the Journal stated, including it will likely be a number of occasions extra highly effective than its industrial model dubbed Llama 2.

Llama 2 is Meta’s open supply AI language mannequin launched in July, and distributed by Microsoft’s cloud Azure companies to compete with OpenAI’s ChatGPT and Google’s Bard.

The deliberate system, particulars of which may nonetheless change, would assist different firms construct companies that produce refined textual content, evaluation and different output, the newspaper reported.

Meta expects to start out coaching the brand new AI system, often known as a big language mannequin, in early 2024, the report added.

Meta didn’t instantly reply to a Reuters request for remark.

Companies and enterprises have flocked to the nascent generative AI marketplace for newer capabilities and refining enterprise processes for the reason that launch of OpenAI’s ChatGPT late final yr.

Bloomberg Information reported in July that Apple is engaged on AI choices just like OpenAI’s ChatGPT and Google’s Bard, including that it has constructed its personal framework, often known as ‘Ajax’, to create giant language fashions and can be testing a chatbot that some engineers name ‘Apple GPT’.

In August, Meta launched an AI mannequin able to translating and transcribing speech in dozens of languages, a possible building-block for instruments enabling real-time communication throughout language divides.

See also  Geologists Efficiently Forecast Volcanic Eruption Utilizing Supercomputers

The corporate stated in a blog post that its SeamlessM4T mannequin may assist translations between textual content and speech in almost 100 languages, in addition to full speech-to-speech translation for 35 languages, combining expertise that was beforehand out there solely in separate fashions.

© Thomson Reuters 2023


Affiliate hyperlinks could also be routinely generated – see our ethics assertion for particulars.