Meta unveils groundbreaking open-source AI model Llama 3.1
Mark Zuckerberg, the CEO, claims that his company’s Llama AI models, which are accessible to everyone, will outperform rival products in the following year. However, this transparent strategy also faces criticism from some detractors.
Meta unveiled a new AI model that it claims can compete with technologies developed by OpenAI and Google, and is offering it for free to all users.
The latest version, Llama 3.1, follows Meta’s approach of releasing its models as open-source, allowing anyone to utilize and alter them without having to pay for it.
If Meta succeeds, it has the potential to disrupt the revenue models of other major tech companies and level the playing field for small businesses to challenge competitors such as OpenAI.
However, this could also open the door for malicious individuals, state-sponsored cyber attackers, and other nefarious actors to exploit advanced technology.
Meta introduced its latest AI model, Llama 3, only a quarter ago. However, this updated version features a model that has been trained on significantly more data compared to its predecessor.
This enhancement has the potential to improve its abilities and offer a valuable resource to businesses and institutions looking to incorporate a larger and more robust AI model in their offerings.
During an open letter on Tuesday, Meta CEO Mark Zuckerberg stated that Llama 3 is on par with the most advanced models in the industry. He also mentioned that upcoming Llama models are anticipated to take the lead as the most advanced in the industry starting next year.
The launch of ChatGPT by OpenAI in late 2022 sparked a competition among major tech companies to develop and market new AI products.
Microsoft struck a lucrative deal worth billions with OpenAI for technology access, while Google developed its own AI models and incorporated them into their offerings. Meta has also invested heavily in AI technology, but lacks a robust cloud software business like Microsoft and Google to market its AI solutions to other companies.
In contrast, the company has opted to release its AI as open-source, with the goal of establishing a platform where businesses lacking their own AI technology can utilize Meta’s, thereby granting the company control over substantial portions of the tech sector, mirroring Google’s dominance in the mobile industry through the Android operating system.
According to Rob Sherman, vice president of policy and deputy chief privacy officer at Meta, companies and individuals have already downloaded Meta’s family of Llama AI models 300 million times.
He wrote that by making source code openly available, open source will help increase global access to AI benefits and opportunities, prevent power from being held by only a few companies, and allow for more equitable and secure deployment of the technology across society. Meta also offers tools for companies to assess the safety of their AI systems.
Zuckerberg likened closed AI models to the approach taken by Apple, which involves implementing guidelines and fees for developers looking to distribute their apps on iPhones. Meta has been dealing with this challenge for a long time.
He wrote that if Meta and other companies were not bound by the restrictions imposed by the phone manufacturer, they would have the opportunity to create more advanced services for users, given the taxing policies towards developers, the random regulations, and the hindrance of new product developments.
The announcement on Tuesday is part of Meta’s efforts to create a new path for itself, by developing a range of AI products that are expected to revolutionize online shopping and communication.
At the beginning of this year, Meta began incorporating Meta AI into its various social media platforms, enabling the feature to produce images and respond to inquiries from users within the search bars on WhatsApp, Instagram, Facebook, and Messenger.
In the tech industry as a whole, there is uncertainty about whether consumers will embrace AI tools in their day-to-day routines. Numerous prominent AI introductions, such as Google’s incorporation of AI in search outcomes, have experienced mishaps and setbacks that have compelled companies to withdraw the product.
We are currently focused on reaching a stage where our primary objective is to have hundreds of millions or even billions of people using Meta AI as an essential part of their activities,” Zuckerberg informed investors in April. “Our next target is to create a highly valuable product.
Read more articles
