Meta has announced its latest artificial intelligence model, named Llama 3.1 405B, which claims to stand shoulder to shoulder with renowned AI models from competitors like OpenAI and Anthropic across various tasks.
In a recent blog post, the company shared that this open-source system offers developers the ability to fully customize its applications for their specific needs. It allows them to train on new datasets and perform additional fine-tuning without restrictions imposed by intermediaries. This development marks a significant milestone in AI technology as it becomes accessible for direct use, without requiring payment or subjecting one's work to the control of Meta.
At present, access to Llama is limited within Meta's own platforms and available only in the United States. However, those utilizing this model will benefit from an additional layer of safety measures that are also open-source, ensuring the company cannot dictate how it should be used by others.
Meta’s co-founder, Mark Zuckerberg, expressed his belief in the importance of open-source AI for fostering a positive future. He emphasized the potential of this technology to enhance human productivity, creativity and quality of life, as well as contributing to advancements in various fields such as medical research. By keeping AI development open-source, Zuckerberg argues that power is distributed more evenly and safely throughout society while preventing it from being concentrated among a few corporations.
Zuckerberg acknowledges the possibility of misuse by bad actors with advanced AI capabilities but posits that widespread deployment may allow for better control over such individuals, as larger entities can counterbalance their actions. While Meta has highlighted the comparable size of Llama 3.1 405B to top competitors' systems, it remains essential to conduct fair assessments between this model and others like GPT-4 to confirm if its effectiveness matches those currently leading in AI technology.
Read next
US officers worry Meta smart glasses could covertly record them
In January, the NYPD’s counterterrorism unit warned officers about a new security threat: Ray-Ban Meta glasses. A memo directed officers to “conduct thorough inspections of all eyewear permitted within inmate cells,” cautioning that recently arrested individuals could use the devices’ small cameras and microphones to record inside police
Google's Pixel 11 sets new standard for flagship phones, review finds
Google’s Pixel 11 continues to define what a base flagship phone should deliver, with leading cameras, extended software support and nearly all the features of its pricier siblings.
The standard Pixel 11 is priced at £879 (€999/$899/A$1,499), £80 higher than its previous version, reflecting ongoing
AI Expert Father of Three Shares Essential Insights for Parents on Artificial Intelligence
A father once promised his child they would write a story together called "The Day Nobody Went to Disneyland." The plot imagined a perfect day when the entire world avoided the park, expecting crowds, leaving it completely empty for a father and child to enjoy alone. Three decades