Meta has announced the release of its latest artificial intelligence model, Llama 3.1 405B, claiming it to be the first open-source system capable of competing with offerings from rivals such as OpenAI and Anthropic.
In a blog post, the company asserts that Llama 3.1 405B
demonstrates competitive performance across a variety of tasks when compared to
its competitors.
If accurate, it would signify that for the initial time, one
of the most potent AI models globally is accessible without a middleman
charging for entry – or overseeing the purpose of its technology.
"Developers have the ability to completely tailor the models to suit their
requirements and uses, train on fresh datasets, and carry out further
fine-tuning," Meta stated. "This allows the wider developer community
and the world to truly grasp the potential of generative AI.
Developers can fully tailor it for their uses and operate in
any setting... all without disclosing data to Meta."
Users utilizing Llama on Meta’s proprietary applications,
currently accessible exclusively within the United States, will benefit from
enhanced security measures, as asserted by the organization.
These measures are also open-source, and the organization
lacks the authority to compel external entities to implement them in their
respective applications of the model.
“I believe that open source is necessary for a positive AI
future,” Mark Zuckerberg, Meta’s co-founder, wrote in a blogpost. “AI has more
potential than any other modern technology to increase human productivity,
creativity, and quality of life – and to accelerate economic growth while
unlocking progress in medical and scientific research. Open source will ensure
that more people around the world have access to the benefits and opportunities
of AI, that power isn’t concentrated in the hands of a small number of
companies, and that the technology can be deployed more evenly and safely
across society.”
Zuckerberg conceded that “bad actors may be able to use the
intelligence of AI models to fabricate entirely new harms”. But, he argued, “I
think it will be better to live in a world where AI is widely deployed so that
larger actors can check the power of smaller bad actors.”
To date, Meta has been self-assessing the capabilities of
its model. While the model’s substantial size places it on par with the most
extensive systems offered by competitors, until independent parties conduct
impartial evaluations comparing Llama 3.1 405B to its peers, such as GPT-4o, it
remains uncertain whether its sheer size will translate into effectiveness
comparable to the current industry leaders.
At present, only users located within the United States have the ability to conveniently access the AI model. It has been reported that Meta made the decision to refrain from launching in the European Union due to apprehensions regarding the regulatory framework established by the bloc concerning AI and data protection. Consequently, it appears that Meta has opted to limit the launch to a single country during the initial phase.
However,
given the open-source nature of Llama 3.1 405B, it is anticipated that users
outside the European Union will likely gain access to the system through
alternative channels in the near future.