Yesterday, Meta CEO Mark Zuckerberg published a letter titled "Personal Superintelligence," revealing Meta's grand vision in the field of artificial intelligence - to build a "Personal Superintelligence" and hinting that the company's future strategy in AI may trend toward being closed-source.
Zuckerberg pointed out in the letter that although recent AI systems have shown signs of self-improvement, this process is still slow, and the task of developing superintelligence is imminent. He admitted that it is still unknown where superintelligence will lead humanity, but he himself is optimistic, believing that superintelligence has the potential to open a new era of "personal empowerment," allowing everyone to have greater autonomy to change the world.
Unlike some industry views that advocate centralizing superintelligence, Zuckerberg emphasized that Meta's goal is to make superintelligence accessible to everyone, enabling each person to harness this power.
Notably, Zuckerberg emphasized in the letter that the benefits of superintelligence should be shared as widely as possible with the world, while also paying attention to potential new security risks. He specifically mentioned: "We need to take a strict approach to reduce these risks and be cautious about the content we choose to open source." This statement has been interpreted by the outside world as an indication that Meta may adopt a more conservative closed-source strategy for its core AI technologies in the future.