Covering Disruptive Technology Powering Business in The Digital Age

IBM Adds to AI Portfolio, Set To Host Meta’s Llama on Its watsonx AI Platform
August 11, 2023 News


Written by: Martin Dale Bolima, Tech Journalist, AOPG.

IBM has recently announced that it will be hosting Meta’s Llama 2 on its enterprise Artificial Intelligence (AI) platform. Llama 2 is an open-source AI initiative recently launched by Facebook’s parent company to supposedly democratise AI and open it to further innovation.

IBM watsonx is an AI and data platform in its nascent stage, having been officially launched by IBM only in May of this year. It is designed to address the challenges businesses face in scaling and operationalising AI and can help organisations “roll out models in an accurate, scalable and adaptable manner,” according to Kitman Cheung, Director of Technical Sales – APAC, IBM Technology Software.

By hosting Llama 2, watsonx will be able to help businesses integrate AI into their business workflows. It will also get them early access to Llama 2, which is being touted by Meta’s founder, Mark Zuckerberg, as being pre-trained by “more than 1 million human annotations” with “applied supervised fine-tuning and reinforcement learning with human feedback” to improve safety and quality.

“Open-source drives innovation because it enables many more developers to build with new technology,” Zuckerberg said in a Facebook post announcing the launch of Llama 2. “It also improves safety and security because when the software is open, more people can scrutinise it to identify and fix potential issues. I believe it would unlock more progress if the ecosystem were more open, which is why we’re open-sourcing Llama 2.”

A Response to Growing AI Demand

Less than a year since OpenAI’s ChatGPT took the enterprise world seemingly by storm, more and more companies are now trying to leverage AI by incorporating it into their workflows and processes—with the end goal of enhancing efficiency and getting access to advanced capabilities.

IBM has also said that the release of Meta’s Llama is just the start, as it will also be launching additional software into its watsonx platform. Among these additions will be AI tuning studios, fact sheets, and other generative AI models.

Llama 2, launched last July, is being distributed by Microsoft Azure and is a direct competitor of OpenAI’s ChatGPT and Google’s Bard. It aims to create a large-scale language model that can learn from any text on the Internet but without human supervision. It is reportedly based on a novel architecture that utilises a combination of self-attention and Locality-Sensitive Hashing (LSH) to process long sequences of text.