Meta reportedly building a new AI model to compete against OpenAI’s ChatGPT

Meta, formerly Facebook, is processing a caller artificial quality (AI) exemplary to vie with OpenAI, the Wall Street Journal said connected Sept. 10.

According to that report, Meta intends for its caller AI exemplary to beryllium respective times much almighty than its existent model, Llama 2. The institution besides expects that the caller exemplary volition person akin capabilities to GPT-4, OpenAI’s competing model.

GPT-4 went unrecorded successful March 2023 and was extended for wide usage successful July 2023. Facebook’s competitor, however, volition not statesman grooming until 2024.

Current reports suggest that it volition beryllium imaginable to usage Facebook’s upcoming exemplary to make services that nutrient blase text, analysis, and outputs.

Meta’s upcoming exemplary volition apt beryllium released arsenic an open-source merchandise disposable astatine nary cost, successful enactment with the company’s existent AI releases and stated goals.

The institution is besides gathering caller information centers and acquiring chips required to execute grooming tasks. According to WSJ‘s reporting, Meta is buying NVIDIA H100 Tensor Core GPUs—the company’s astir precocious exemplary for AI training—which are successful very precocious demand.

Meta presently offers Llama 2

Though Meta’s caller AI merchandise volition beryllium a important measurement forward, it is not the company’s lone merchandise to date. After partial releases and root codification leaks successful aboriginal 2023, Meta afloat released Llama 2 successful July 2023 for probe and commercialized purposes.

Meta has galore different AI efforts underway, galore of which are aimed astatine the public. The institution launched a developer adjunct instrumentality successful August and announced AI-powered advertisement services successful May. The company’s chatbots are besides expected to spell unrecorded soon.

Incidentally, Meta CEO Mark Zuckerberg and different tech leaders are acceptable to sermon artificial quality with the U.S. senators successful a bid of upcoming forum events, according to CNBC. Other tech leaders acceptable to be see Tesla and Twitter/X enforcement Elon Musk, Microsoft co-founder Bill Gates, and OpenAI CEO Sam Altman. Those forum dates volition statesman successful September and tally connected an ongoing basis.

The station Meta reportedly gathering a caller AI exemplary to vie against OpenAI’s ChatGPT appeared archetypal connected CryptoSlate.