Meta Pushes Open-source In New Llama Model & Updates Meta AI

In the artificial intelligence (AI) world, AI companies are trying their best to outdo rivals and offer the most powerful tools and platforms.

These platforms could be used to generate outputs or help developers build software applications powered by AI or GenAI (generative artificial intelligence).

While many AI companies are developing closed models, Meta is looking to offer its models as open-source products.

The shift to open-source AI model development is one that Meta has taken personally and is reflected in the release of its latest Llama generative AI model. 

So, let’s explore what Meta’s new model brings to the AI table and what the company said about its move to push open-source models!
 

What Did Meta Say About Its Newest Llama 3 AI Model Release?

 

• Through a news release written by Mark Zuckerberg, Meta’s CEO and Founder, announced the release of Meta’s Llama 3.1 405B, its first frontier-level open-source AI model, as well as new and improved Llama 3.1 70B and 8B models.

• The release also reiterated Zuckerberg’s commitment and belief that AI will develop with the help of open-source models.

• Llama 405B comes with a “significantly” better cost to performance when compared to close models. Additionally, as an open-source model, it will be “the best choice for fine-tuning and distilling smaller models.”

• Llama’s 8B model is a lightweight, ultra-fast model that users can run anywhere, while 70B is a cost-effective, high performant model for various use cases and 405B brings in the widest variety of use cases.

• The model also possesses advanced capabilities in answering complex questions like math and coding.

• Furthermore, Meta revealed that it was working with a wide range of companies to grow a broader ecosystem.

• Developers, CEOs and government officials who have spoken with Zuckerberg have often voiced their requirement of open-source models, saying they would like to train, fine-tune and distill their own models.

• Open-source models would also allow them to protect their data, not get locked into a closed vendor, control their own destiny through a model that’s efficient and affordable to run, as well as invest in the ecosystem that’s going to be the standard for the long term.

• “I believe that open-source is necessary for a positive AI future. 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.”

 

What Did Mark Zuckerberg Say About The Move?

 

• “Open-source AI Is the Path Forward” is what the news release written by Meta CEO and Founder Mark Zuckerberg was titled.

• Zuckerberg touched upon the history of software development, beginning with how major tech companies would develop their own closed versions of Unix and any other approach seemed unlikely.

• However, soon after, the open-source Linux gained popularity as it allowed developers to modify its code, which came across as more affordable and eventually “more advanced, more secure and had a broader ecosystem.” Today, it’s the industry’s standard foundation.

• “I believe that AI will develop in a similar way. Today, several tech companies are developing leading closed models. But open-source is quickly closing the gap. Last year, Llama 2 was only comparable to an older generation of models behind the frontier.”

• “This year, Llama 3 is competitive with the most advanced models and leading in some areas. Starting next year, we expect future Llama models to become the most advanced in the industry. But even before that, Llama is already leading on openness, modifiability, and cost efficiency.”

 

What New Updates Did Meta AI Get?

 

• Through a news release published on its website, Meta revealed a new set of updates made to Meta AI, its assistant.

• For one, the company announced it was expanding access to Meta AI and introducing new features.

• Meta AI will now be available in 22 countries, which includes the addition of Argentina, Chile, Colombia, Ecuador, Mexico, Peru and Cameroon.

• Users can interact with the AI assistant across WhatsApp, Instagram, Messenger and Facebook using newly added languages such as French, German, Hindi, Hindi-Romanized Script, Italian, Portuguese and Spanish, with more to come.

• Meta AI is also available on Ray-Ban Meta smart glasses.

• It will also come to Meta Quest VR headsets next month (in the United States and Canada in experimental mode), replacing the current Voice Commands. This will allow users to control their headset hands-free, get answers to questions, access real-time information, check the weather and more.

• Meta AI can also be used with Vision in Passthrough.

• Users can leverage Meta’s new 405B model to get solve complex math problems, perform tasks and get answers to questions as well as receive higher-quality computer code.

Do you think Meta’s dream of having an open-source ecosystem will encourage other AI companies to follow suit and offer their models as open-source?

Do you think this move will help Meta become the most popular AI software development platform provider?

Let us know in the comments below!