The amount of training data that might exist in the world has been exhausted, since everyone has already scanned the Internet. It is also starting to suffer from more and more restrictions, especially from large content creators. Even OpenAI, Google, and Microsoft have difficulty obtaining a larger volume of data. Meta, with 2 billion daily users creating very human interactions, will be able to search for text, conversations, slang and things that are otherwise hard to find
Alex Wintsky
If you think you’re working for free, you’re right. Even your relentless messages can lend Zuckerberg a helping hand.
Let’s say you issue a command, the AI makes a mistake and you provide a feedback: “You shouldn’t have done that.” This data can be used to make the system more efficient. There is a relationship in which the company benefits from interaction guided by the use of artificial intelligence
João Victor Arcegas, Coordinator of ITS-Rio (Institute of Technology and Society)
However, it is possible to ask Meta not to use your data to train the AI. However, don’t think it will be easy. it is necessary Fill out this form to get.
It is worth noting that Meta IA is powered by Llhama 3.2, the main language model for Meta. In addition to fighting on equal terms with other companies, it has the advantage of adopting an open source method. In other words, while Google and OpenAI charge for access to their models, Meta provides it for free.
More Stories
What ChatGPT knows about you is scary
The return of NFT? Champions Tactics is released by Ubisoft
Happy Meal will contain exclusive Mario Kart 8 collectible toys! See details