Meta, the company led by CEO Mark Zuckerberg, is actively developing artificial general intelligence (AGI) and has indicated that it may release it as open source software. AGI is a form of AI that is capable of understanding, learning, and applying knowledge in a way that is similar to human intelligence. Zuckerberg has stated that the long-term vision for Meta is to build AGI, open source it responsibly, and make it widely available for everyone’s benefit.
Meta is bringing together two of its AI research teams, FAIR (Facebook AI Research) and GenAI, to work towards this goal. The company is also working on training its next-generation model, Llama 3, and is building a massive compute infrastructure to support future AI models. This includes plans to have 350,000 H100 GPUs by the end of the year, which would amount to almost 600,000 H100 equivalents of compute.
Zuckerberg has expressed that AI and the metaverse are closely linked, and he envisions smart glasses as a primary way people will interact with AI and the metaverse. He has also mentioned that Meta’s approach to AGI will be to open source it as long as it is safe and responsible to do so, although he has not committed to a definitive plan for open sourcing any potential AGI developed by Meta.
The move towards open sourcing AGI aligns with Meta’s and Zuckerberg’s advocacy for keeping AI technology open-source, which has sparked debate within the tech industry. This stance is somewhat in contrast to more secretive rivals and has led to the formation of the AI Alliance, a group launched by Meta and IBM to promote an open-source vision of AI.
Zuckerberg’s announcement follows a trend of tech leaders downplaying the dangers of AGI, and it comes at a time when the tech industry is hotly debating the control and accessibility of AGI technology. The decision to open source AGI is ultimately Zuckerberg’s, given his voting control over Meta’s stock.
In summary, Meta is investing in the development of AGI with the intention of potentially open sourcing it, which would be a significant contribution to the AI ecosystem and could influence the future direction of AI development and accessibility.