Meta on the Move: Next Super AI to Rival OpenAI’s GPT-4

With rapid advancements in artificial intelligence technologies, it is unsurprising that Meta, the tech giant previously known as Facebook, has embarked on a mission to train an AI model with capabilities believed to rival OpenAI’s GPT-4.

As reported by trusted sources, the company’s venture into an ever-competitive field spells exciting implications.

The Upcoming AI Challenger

Meta is reportedly making efforts towards developing a new, large language model with training expected to begin in 2024. 

The magnitude of this endeavor is backed by their increasing acquisitions of Nvidia H100 AI-training chips.

The company’s ultimate goal, it appears, is to advance the creation of AI tools capable of emulating human expressions.

To enable this, the company has assembled a team dedicated to building the new AI model.

Notably, in an offer likely to excite the scientific community, Meta CEO Mark Zuckerberg hints at these AI tools being initially free for companies and researchers.

Code Llama: A Step Towards the Goal

In paving the path towards this ambitious goal, Meta’s concentration on AI tools is encapsulated by its recent launch of Code Llama. 

This AI tool, designed for code generation and debugging, demonstrates the company’s commitment to innovation in AI.

Code Llama, developed from the building blocks of Llama 2, is a code-specialized version that has been extensively trained on code-specific datasets.

It promises efficiency and productivity boosts for developers and simplifies code learning for beginners.

Looking Ahead: Generative AI at the Forefront of Meta’s Goals

Zuckerberg announced earlier this year about the formulation of a “top-level” product team focused solely on generative AI. 

As AI technologies rapidly develop, the envisioned rivalry of Meta’s new AI model against OpenAI’s GPT-4 suggests the dawn of an interesting era in the AI world.