Mark Zuckerberg, the founder of Meta, recently revealed his ambitious goal. He wants to create a super-smart AI called Artificial General Intelligence (AGI). He wants to join the ranking of tech giants like OpenAI and Google in the race for superhuman intelligence.
Zuckerberg plans to merge Meta’s two AI research divisions, Fundamental AI Research (FAIR) and the Generative AI team, to accelerate progress towards Artificial general intelligence (AGI). This move is part of Meta’s broader mission to create open-source AGI that benefits everyone.
Zuckerberg acknowledges that it’s a bit difficult to say exactly what super-smart AI is or when it’ll happen. He thinks it’s more like a slow and steady progress rather than a big sudden change.
The Company’s recent focus is on developing large language models like Llama 2 and the upcoming Llama 3 that will have code-generating capabilities. This shows that Meta is really serious about making AI better.
To support these endeavors, Meta is expanding its AI teams and infrastructure. By the end of the year, the company aims to own over 340,000 Nvidia H100 GPUs, a key component for building generative AI. Zuckerberg emphasizes that this computing power scale may surpass that of any other individual company.
The issue of who controls AGI remains a hot topic. People are talking a lot about who gets to decide how super-smart AI works. Zuckerberg leans towards an open-source approach. He prefers the idea of sharing and making it open for everyone, which is different from other companies like OpenAI that keep things more private. He thinks being open is fair and gives everyone a chance. But in the end, the choice to share or not at Meta is up to Zuckerberg.
Despite this AI-centric shift, Zuckerberg clarifies that Meta remains committed to the metaverse and Reality Labs initiatives. He envisions AI playing a pivotal role in creating virtual worlds with intelligent characters that interact with users across Meta’s social apps.
Big tech companies are really trying hard to find the best people to work on super-smart AI. Mark Zuckerberg says it’s super tough because lots of companies want smart researchers and engineers. These smart people can get paid in huge amounts, like over $1 million every year. Meta is spending a lot of money on AI and supercomputers to show how serious they are about being the best in this AI field.
In the broader context, Zuckerberg’s AGI push comes amidst Meta’s recent emphasis on the metaverse. While some may perceive it as a pivot, Zuckerberg asserts that Meta remains steadfast in its commitment to Reality Labs and the metaverse, with ongoing investments exceeding $15 billion annually.
Mark Zuckerberg is really into making super-smart AI, and he’s doing this while also focusing a lot on the “metaverse.” Some people might think he’s changing direction, but Zuckerberg says Meta is still very committed to “Reality Labs” and the metaverse. They’re spending more than $15 billion every year to show they’re serious about it.
Meta wants to change how we use really smart AI in our lives. They believe that in the future, AI will be a big part of our everyday routine.