Meta CEO Mark Zuckerberg announced the company is merging its two AI research teams, FAIR and GenAI, to pursue artificial general intelligence. In a video posted on Instagram and Threads, he stated the combined team will work toward building AGI and making it open-source for public access.
Artificial general intelligence, which would match human cognitive abilities across all tasks, does not yet exist. Meta's announcement outlines a long-term goal without a specific timeline.
To support this effort, Meta is building massive computing infrastructure including 350,000 Nvidia H100 GPUs, with total computing power reaching 600,000 H100 equivalents by year's end for training Llama 3. The company also continues developing AI applications for its Ray-Ban smart glasses, which Zuckerberg described as an ideal form factor for AI assistants that can see and hear the world alongside users.
The announcement comes amid industry debate about AGI timelines and safety. OpenAI CEO Sam Altman recently discussed managing AGI risks at the World Economic Forum in Davos, while Meta's chief AI scientist Yann LeCun has stated AGI may not arrive for at least five years.
Meta's commitment to open-sourcing AGI technology distinguishes it from competitors like OpenAI and Google, though this approach continues to spark debate about responsible AI development and potential security risks.
Post updated to improve title and body errors.
Read next: EU Asks Tech Giants For More Data Regarding Access Given To Researchers For Risk Studies
