Towards the end of 2021, we saw social media giant Facebook rebrand as Meta to emphasize its focus on the concept of Metaverse and transform itself into a tech-heavy enterprise. It wouldn’t be wrong to say that after the time that has passed, Metaverse has not gone beyond a dream, at least for now. For this ideal, virtual reality hardware is far from practical and expensive, and there is an insurmountable quality problem on the software side.
In summary, Metaverse, which did not add any innovation to the lives of users, could not get rid of abstraction. In this direction, Mark Zuckerberg, the CEO of Meta, which lost 13.7 billion dollars in 2022 despite its investments, announced the artificial intelligence supported language model .
The company, which does not want to be left behind from the ChatGPT competition that Google’s Bard and Microsoft’s attracted to its side , used the artificial intelligence system LLaMA (Large Language Model Meta AI) to create text, chat, summarize a written content and solve mathematical theorems or predict protein structures. mentioned that it shows promise in more complex tasks.
According to the blog post on the subject, LLaMA is trained through publicly available articles on platforms such as C4, Wikipedia, Stack Exchange, and ArXiv. Meta also frequently cites the advantage that LLaMA requires less computing power compared to other major language models.
LLaMA will not be integrated into any of its platforms, including Facebook and Instagram, as it is currently only available to researchers for non-commercial purposes. It seems that the company will continue to appear in different branches of technology until the environment suitable for Metaverse is formed.