Instagram, Facebook, WhatsApp and more roof company Meta, that advanced artificial intelligence models can have disastrous consequences knows.
Meta, who recently shared a document, said that advanced artificial intelligence models could have disastrous results and stop the development of artificial intelligence systems that it deems very risky. Risks within itself “High risk” And “Critical risk” The company, which divides it into two, states that risky systems can help cyber, chemical and biological attacks. Artificial intelligence systems are already used for malicious purposes, More advanced systems reveal great threats for everyone. Companies such as OpenAI, Microsoft or Meta need to take very strict security measures. Normal users Artificial intelligence models that allow him to reach more information in seconds, It poses an unprecedented risk for humanity.
You may be interested in
According to the statement made by Mark Zuckerberg Meta, By the end of the year, artificial intelligence -oriented will have 1.3 million GPUs. META, which currently has very large server systems, will continue its investments in this field in 2025 without slowing down. Stating that Meta will spend between 60 billion and 80 billion dollars for investment in different fields in 2025, CEO Mark Zuckerberg1.3 million GPUs aimed at reaching at the end of the year (Most of them will come from Nvidia) with productive artificial systems and work will add to a great extent. Last week, Microsoft came up in 2025 that he planned to spend $ 80 billion for new artificial intelligence -oriented data centers.
Meta, in the last months of the Llama 3.3 70b artificial intelligence model had made a sound. In a statement about the new model, the following were mentioned: “Llama 3.3 70B offers the performance of our 405B model, but our newest model is easier and more costly to run. 70B, which benefits from the latest developments in post -training techniques, including online optimizations, is a more accessible option for the entire open source community by significantly increasing performance at a low cost. ”