LAGOS, Nigeria (VOICE OF NAIJA) – Meta Platforms is currently in the process of developing a cutting-edge artificial intelligence system with ambitions to match the capabilities of OpenAI’s most advanced model.
This new AI endeavour from the parent company of Facebook is expected to be ready for deployment next year, boasting performance several times superior to its existing commercial model, Llama 2.
Llama 2, Meta’s open-source AI language model introduced in July, is delivered through Microsoft’s Azure cloud services, positioning it as a contender against OpenAI’s ChatGPT and Google’s Bard.
While the specifics of this planned AI system are subject to potential changes, its primary purpose is anticipated to facilitate the development of advanced text generation, analysis, and other output services for various companies.
The report further mentions that Meta’s timeline includes the commencement of training for this new large language model in early 2024.
READ ALSO: Meta, LG Partner For 2025 Quest Pro Microdisplays Challenging Apple
Meta has yet to respond to a Reuters inquiry for comments on this development.
In recent times, businesses and enterprises have shown significant interest in the emerging generative AI market, driven by the need for enhanced capabilities and streamlined business processes, following the introduction of OpenAI’s ChatGPT in the preceding year.
In a related development, Bloomberg News reported in July that Apple is also actively pursuing AI solutions akin to OpenAI’s ChatGPT and Google’s Bard.
Apple has taken steps to create its own framework, known as ‘Ajax,’ for the development of large language models. The company is also in the testing phase for a chatbot, internally referred to as ‘Apple GPT.’
Furthermore, in August, Meta unveiled an AI model capable of translating and transcribing speech in numerous languages, marking a potential milestone in enabling real-time communication across linguistic barriers.
The model, named SeamlessM4T, can support translations between text and speech in nearly 100 languages, and it offers full speech-to-speech translation for 35 languages.
This integration combines technologies that were formerly available only in separate models.