Meta ramps up chatbot buzz, announces LLaMA search tool rivals Microsoft’s ChatGPT and Google’s LaMDA

Meta Platforms has introduced a research tool to create chatbots and other artificial intelligence-based products, seeking to create buzz for its own technology in an area recently focused on internet rivals Google and Microsoft.

The tool, LLaMA, is Meta’s latest entry into the realm of large language models, which “have shown great promise for generating text, having conversations, summarizing written documents, and more complicated tasks like problem solving. mathematical theorems or the prediction of protein structures”. On Friday, Chief Executive Mark Zuckerberg said in a Facebook post.

At this time, LLaMA is not used in Meta’s products, which include social networks Facebook and Instagram, according to a spokesperson. The company plans to make the technology available to AI researchers.

“Meta is committed to this open search model,” Zuckerberg wrote.

Large language models are massive AI systems that suck in huge volumes of digital text – from news articles, social media posts, or other internet sources – and use this written material to form software that predicts and generates content by itself when it receives a prompt or request. Templates can be used for tasks such as writing essays, composing tweets, generating chatbot conversations, and suggesting computer programming code.

The technology has become popular and controversial in recent months as more companies have started building them and introducing model-based product testing, highlighting a new area of ​​competition among tech giants. . Microsoft is investing billions in OpenAI, the creator of GPT-3, the grand model language that runs the ChatGPT chatbot. The software maker this month unveiled a test version of its Bing search engine running on OpenAI’s chat technology, raising immediate concerns about its sometimes inappropriate responses.

Alphabet’s Google has a model called LaMDA, or Language Model for Dialog Applications. The internet search and advertising leader is testing a chat-based, AI-powered search product called Bard, which also still has some issues.

Meta previously released a large language model called OPT-175B, but LLaMA is a newer and more advanced system. Another Meta model released late last year, Galactica, was quickly taken down after researchers found it routinely shared biased or inaccurate information with people who used it.

Zuckerberg has made AI a top priority at the company, often speaking about its importance to improving Meta’s products during earnings conference calls and in interviews. Although LLaMA is not currently used in Meta products, it may be used in the future. Meta now relies on AI for all sorts of functions, including content moderation and ranking material that appears in user feeds.

Making the LLaMA model open source allows outsiders to more clearly see how the system works, adapt it to their needs, and collaborate on related projects. Last year, Big Science and Hugging Face released BLOOM, an open-source LLM intended to make this kind of technology more accessible.

© 2023 Bloomberg LP