A Meta it wasn’t enough to challenge Twitter to its own game with Threads, now it aims to launch the artificial intelligence LLaMa commercially. But what is LLaMa? What is it about? And how does it differ from OpenAI’s ChatGPT and Google Bard?
Put it this way: In the world of artificial intelligence, one issue divides developers: make the code secret or share it freely as “open-source”? While OpenAI has kept data from projects like GPT-4 confidential, Meta has decided to make LLaMa available to researchers and academics. Now, according to the Financial Times, Mark Zuckerberg’s company is about to launch a new commercial version of its chatbot, a model that companies can customize to their liking.
What do we know about LLaMa, Meta’s artificial intelligence
The first signs of an important step in the world of AI were taken a few days ago, when Yann LeCun, vice president and chief AI scientist at Meta he had declared that “in the coming months, but perhaps even in the coming weeks, the competitive landscape of AI will be completely revolutionized”. Words that sound more like a warning to competitors than an announcement to users. The fact is that Meta’s plans seem to be quite clear: an open-source model that can be customized according to the needs of companies. Or even depending on the needs of the individual user, why not.
Open-source models have both advantages and disadvantages. Giving a large audience the different functionalities of an AI can encourage model learning, thanks to the increase in the volume of data it receives. However not everyone who uses the code has noble goalswhich can pose risks when it comes to a technology that involves many people in and out of the tech industry.
Access to LLaMa’s business model should be free at launch, but that may change later. According to the Financial Times, sources close to the company have reported that the company could set a tariff to allow companies to customize the chatbot.