Whilst different firms are interested by conversational AI, it seems like Meta has some loftier ambitions in thoughts. The social media large is reportedly running on synthetic common intelligence (AGI).
In an Instagram Reel printed these days, Meta CEO Mark Zuckerberg published that his corporate is lately running on AGI. This actual form of AI is steadily described as human-level AI. Two of the corporate’s AI analysis groups — FAIR and GenAI — were introduced in combination to increase the AGI and make it open supply.
The social put up in query options Zuckerberg explaining Meta’s function in a video with a caption that reads:
Some updates on our AI efforts. Our longer term imaginative and prescient is to construct common intelligence, open supply it responsibly, and make it extensively to be had so everybody can get advantages. We’re bringing our two primary AI analysis efforts (FAIR and GenAI) nearer in combination to give a boost to this. We’re lately coaching our next-gen type Llama 3, and we’re construction huge compute infrastructure to give a boost to our long term roadmap, together with 350k H100s via the tip of this 12 months — and total nearly 600k H100s equivalents of compute for those who come with different GPUs. Additionally actually enthusiastic about our development construction new AI-centric computing units like Ray Ban Meta sensible glasses. Quite a bit extra to come back quickly.
This information comes after Meta’s leader scientist, Yann LeCun, expressed skepticism in December that AI superintelligence would arrive anytime quickly. The French pc scientist stated he believes that present AI methods are nonetheless many years clear of attaining some semblance of sentience. LeCun additionally added that we’re much more likely to get cat or dog-level AI intelligence years sooner than we reach human-level AI intelligence.