In the cutting-edge world of artificial intelligence (AI), rivalry is the name of the game, and Meta is stepping up to challenge OpenAI. According to a Wall Street Journal report, Meta, the parent company of Facebook and Instagram, is developing a new AI model aimed at dwarfing OpenAI’s GPT-4 model in power and capability.
Meta’s ambitious project is said to be “several times” more potent than its Llama 2 model, launched earlier this year. Unlike many corporate technology tools, Meta aims for the new AI model to be open-source. This strategy will allow businesses and entrepreneurs the freedom to build AI tools capable of producing high-level text and analysis.
https://t.co/lRFezdftLv #ZeroHedge #Science #Tech
— Brian Hamel (@hamel1776) September 11, 2023
“The company has also been building the data centers necessary to create such a high-level system while acquiring more of Nvidia’s H100 semiconductor chips, the most powerful and coveted chips currently available on the market,” the Wall Street Journal detailed.
While OpenAI has not disclosed its GPT-4 model parameters, estimates suggest they number somewhere around a massive 1.5 trillion. In comparison, Llama was trained on just 70 billion parameters. The yet-unnamed Meta model will likely appear after Google’s upcoming large language model (LLM) Gemini, aiming to seize a crucial market share in a burgeoning AI sector.
The fact that Meta is leaning into this venture on its infrastructure is particularly telling. Although Microsoft, a primary backer of OpenAI, has collaborated with Meta for its Llama 2 model, sources indicate no such team-up is in the cards for the new project. This move likely speaks to Meta’s growing self-reliance and aspirations to become a formidable player in high-stakes AI.
Meta is building an AI model to rival OpenAI’s most powerful system #Metaverse #AI pic.twitter.com/honS593i0M
— Crypto Chirp™ 📢 (@Chirp_Crypto) September 11, 2023
This development is symbolic of the broader AI arms race unfolding globally. Governments and corporations alike are scrambling to master and control high-level AI systems. The U.K. government has committed $130 million to develop AI systems. China’s booming AI sector has released over 70 new AI models since the country’s latest AI legislation was enacted.
While there’s much to be optimistic about, it’s essential not to overlook the concerns that come with rapidly advancing AI technology. A recent study by Stanford University and UC Berkeley researchers flagged decreasing accuracy rates in OpenAI’s GPT-4 model. As these AI systems become increasingly embedded in our daily lives, reliability will be crucial.
Mark Zuckerberg’s comments during a July earnings call point to Meta’s big bets on AI: “I’ve said on a number of these calls that the two technological waves that we’re riding are AI in the near term and the metaverse over the longer term.”
Meta’s new AI model not only poses a direct challenge to OpenAI and other tech giants like Google and Microsoft but also highlights the company’s strategic shift toward becoming a dominant force in the AI sector. If successful, this could alter the landscape of AI and offer a more competitive, decentralized ecosystem for businesses and innovators. Time will tell if Meta can make good on its lofty ambitions.
