...
By Monica Sager
Live News Reporter
0
Share
Newsweek is a Trust Project member
See more of our trusted coverage when you search.Prefer Newsweek on Googleto see more of our trusted coverage when you search.
Mark Zuckerberg's Meta announced on Tuesday a new open-source artificial intelligence (AI) model that claims to outperform OpenAI's ChatGPT and other competitors.
Meta's latest model, Llama 3.1, is "in a class of its own," according to the company's press release. The model is equipped with "state-of-the-art capabilities in general knowledge, steerability, math, tool use and multilingual translation."
"With the release of the 405B model, we're poised to supercharge innovation—with unprecedented opportunities for growth and exploration," Meta said.
Read More on News
...
The company believes the new generation of Llama will be able to "ignite" new workflow proficiencies. Llama 3.1 has a "significantly" longer context length, which allows for the model to have additional and stronger reasoning capabilities.
That effectively means that Llama will be able to summarize long-form text, assist with multilingual conversations and even help with coding.
Meta has also updated its license procedures. With Llama 3.1, developers will be able to use the outputs from the models to improve other systems.
"Today, several tech companies are developing leading closed models," Zuckerberg wrote in an open letter published Tuesday. "But open source is quickly closing the gap."
He added that next year, he expects future Llama models to "become the most advanced in the industry." In a post to Instagram, Zuckerberg predicted that Meta's Llama-powered AI bot will surpass ChatGPT as "the most used AI assistant in the world by the end of the year."
"But even before that, Llama is already leading on openness, modifiability, and cost efficiency," he wrote in the open letter.
"Meta is committed to open source AI...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."
Zuckerberg wrote that most organizations will not support models that they "cannot run and control themselves."
...
"They don't want closed model providers to be able to change their model, alter their terms of use, or even stop serving them entirely," he wrote. "They also don't want to get locked into a single cloud that has exclusive rights to a model. Open source enables a broad ecosystem of companies with compatible toolchains that you can move between easily."
Zuckerberg's company evaluated Llama's performance on over 150 benchmark datasets over a wide range of languages prior to the model's release. New languages include French, German, Hindi, Italian, Portuguese and Spanish.
"Extensive human evaluations" were also performed to compare Llama 3.1 with other models in real-world scenarios.
"Our experimental evaluation suggests that our flagship model is competitive with leading foundation models across a range of tasks, including [OpenAI's] GPT-4, GPT-4o, and [Anthropic's] Claude 3.5 Sonnet," the press release read. "Additionally, our smaller models are competitive with closed and open models that have a similar number of parameters."
OpenAI's Response
Altman and OpenAI have been criticized by some social media users for not immediately releasing a competitive upgrade to GPT after Meta's announcement. OpenAI did note the company's efforts to make sure AI can "benefit everyone."
"We won't release a new model if it crosses a 'medium' risk threshold until we implement sufficient safety interventions," the company posted. "We are developing levels to help us and stakeholders categorize and track AI progress. This is a work in progress and we'll share more soon."
...
Altman, for his part, has not said anything publicly about Meta's new Llama release, but did recently hint at the company's plans for future iterations of ChatGPT, the commercially available AI bot that remains OpenAI's flagship project.
He compared the rise of AI to industrial-era developments and attempted to reassure an anxious public that the developers of cutting-edge AI applications are taking into account their worries that the technology could become too powerful.
Speaking alongside Airbnb CEO Brian Chesky at the 20th Aspen Ideas Festival in Colorado, Altman said the development of increasingly advanced AI, like the work being done internally at OpenAI under the Strawberry codename, is "inevitable."
"We think about it as this series of thresholds where the systems get better and better capabilities," Altman said. "You can use ChatGPT today for some things and you'll be able to use it for much more helpful tasks in the future."
OpenAI is said to be working on a potentially groundbreaking new approach for its artificial intelligence models that could produce advanced reasoning capabilities that have, so far, been out of the grasp of even the most cutting-edge AI technology.
The project is code-named Strawberry, according to internal documents leaked to Reuters.
In an interview with the news agency, a person familiar with the matter said the Sam Altman-led company is racing toward advanced AI capabilities, beyond what the general public may even think possible.
"We want our AI models to see and understand the world more like we do," an OpenAI spokesperson said. "Continuous research into new AI capabilities is a common practice in the industry, with a shared belief that these systems will improve in reasoning over time."
Do you have a story we should be covering? Do you have any questions about Sam Altman and OpenAI? Contact LiveNews@newsweek.com.
View 0 Comments
Submit Correction
View Editorial & AI Guidelines