The business that renamed itself just three years ago to go all-in on the previous tech craze of the metaverse appears more determined than ever to catch up as OpenAI and Google sprint ahead in the continuing Silicon Valley AI race.
The Wall Street Journal has recently reported that Meta is quietly working on a potent new AI model that will compete with OpenAI’s GPT-4, the most recent edition of the rival AI company’s large language model (LLM).
The specifics of the model are unknown; according to the WSJ, they could evolve. Currently, the AI is purportedly intended to “help other companies to build services that produce sophisticated text, analysis, and other output.” It’s also important to note that this isn’t Meta’s first entry into the LLM space; the company’s highly anticipated language model, LLAMA-2, was only recently launched.
But according to WSJ’s sources, Meta expects its new AI to be “several times more powerful” than LLAMA-2. This may indicate that Meta realizes it needs to substantially boost its game if it wants to keep up with its rivals because LLAMA-2 is less effective than a GPT-4 substitute.
Dust in the wind
According to the story, the model is being developed by a Meta team that Mark Zuckerberg, the company’s founder and longtime dictator, founded early this year. The team’s primary mission was to speed up the company’s endeavor to develop a generative AI system that could “produce human-like expressions.”
Please don’t hold your breath; the company reportedly intends to begin training the new AI by early 2024, and in the meantime, Meta has allegedly been busy collecting those coveted Nvidia H100 processing chips to build its AI-supporting infrastructure.
According to the report, Zuckerberg has been pushing for the company’s new AI to be made open-source. This decision has reportedly troubled Meta’s lawyers, who have allegedly voiced some legitimate worries about threats like potential copyright infringement aided by AI and the use of AI to create misinformation.
Notably, OpenAI is more than Meta’s only significant rival. While smaller companies like the heavily Google-funded Anthropic have also made a major stir, Google is also a leader in the gold rush-hungry pack. Apple is reportedly spending “millions of dollars every day” to train its AI while this is taking place.
It’s also crucial to remember that Meta is only one of several companies developing a new business model. There is much anticipation for Google’s Gemini AI, which could embarrass GPT-4. And if Meta releases something noteworthy and novel after then? Of course, everything is still in the air, but Zuck might find himself much further behind than anticipated.