Meta CEO Mark Zuckerberg has overseen the development and open source release of the company’s new LLaMA 2 model, demonstrating intent to boost accessibility and innovation in the space. Speaking about the current AI state of play, he told New Statesman: “We’re not close to super-intelligence”, a notion he says is “pretty clear to anyone working on these projects”.
Speaking to employees at Meta’s Menlo Park headquarters, Zuckerberg, according to Axios, stated: “In the last year, we’ve seen some really incredible breakthroughs — qualitative breakthroughs — on generative AI and that gives us the opportunity to now go take that technology, push it forward, and build it into every single one of our products.”
Going forward, the Meta chief discussed during a podcast with MIT research scientist Lex Fridman a departure from a “one-size-fits-all” approach and a shift to customisable experiences – which further explains the decision to make LLaMA 2 open source. In addition to LLaMA, Zuckerberg revealed plans for further model development during the podcast, in an aim to stay tuned in to bespoke and evolving customer requirements.