Meta Platforms, formerly known as Facebook, is not new to jumping on the hype train of emerging technologies. Upon its rebranding from Facebook to Meta, the social media giant had launched its own metaverse in 2021. As metaverse hype slowed down drastically, Meta’s virtual world reduced expansion as they realized increasing costs and a major loss in user engagement. 

Arguments that highlight difficulty in accessing hardware for the metaverse, as well as being difficult to use and costly, make a valid point that technology is not up-to-speed with adoption expectations.

The advent of generative AI and AI tools, however, is proving to have a different impact compared to the metaverse. Since November 2022, the world has been enthralled by AIs capabilities and how it opens doors to new opportunities. In August 2023, Meta unveiled Code Llama, its new open source, commercially licensed large language model (LLM) specifically geared towards programming. 

It is built on top of Meta AIs Llama 2, which launched in July earlier this year in partnership with Microsoft.

In their research publication, Meta expresses, “we release Code Llama, a family of large language models for code based on Llama 2 providing state-of-the-art performance among open models, infilling capabilities, support for large input contexts, and zero-shot instruction following ability for programming tasks.”

Their ‘family’ of LLMs include a 7-billion, a 13-billion, and a 34-billion parameter model, all of which are trained on 500 billion tokens. The 7B and 13B models are able to run on fewer graphic processing units (GPUs) which is a great boon considering the limited supply of hardware there is to support the demand for high-performing GPUs.

While the rise of new generative AI and AI tools is exciting, there are inherent risks which Meta addressed. As with all AI coding tools, according to a research team affiliated with Stanford, “engineers who use AI tools are more likely to cause security vulnerabilities in their apps.” This is due to code appearing correct, however it raises security concerns by triggering compromised software and employing insecure configurations.

Meta’s announcement highlights that “Code Llama is designed to support software engineers in all sectors - including research, industry, open source projects, NGOs, and businesses. But there are still many more use cases to support than what our base and instruct models can serve. We hope that Code Llama will inspire others to leverage Llama 2 to create new innovative tools for research and commercial products.”

Users can download Code Llama today through Github, or directly from Meta AIs website.