Develop a Superintendent. This is the objective of companies that specialize in artificial intelligence, particularly OpenAi and goal. It is a system that would be smarter than humans. A scenario that gives more science fiction than reality at the moment.
But that did not prevent Meta CEO, Mark Zuckerberg, to share his vision about the Superintendent, on the occasion of the publication of the second quarter. “Now the superintelligence is seen,” he said in a memorandum, ensuring that he will represent “a new era for humanity.”
“A new era of personal empowerment”
Mark Zuckerberg insists on a particular point with respect to this future form of AI: What Meta plans to develop will be personal.
“I am extremely optimistic about the ability of superintelligence to help humanity to accelerate their progress. But most importantly, it has the potential to open a new era of personal empowerment, where everyone will have more means to improve the world in the direction they want,” said the billionaire.
This personal superintendent would have an even more significant impact on the life of each individual, helping everyone to achieve their goals or even create what they want to see in the world. According to Mark Zuckerberg, he can even help a person “to be a better friend for her loved ones and become the person who aspires to be.”
When you wanted to develop and make available to all this AI, the Meta CEO intends to distinguish itself from other companies that, according to him, seeks to focus on the empowerment of work. He also believes that the personal superintendent will be, with much, the most useful, to be able to understand the objectives of people, but also know them deeply.
Beyond this new form of AI, Mark Zuckerberg also believes that personal devices, such as glasses, will eventually become the main computer devices used by humans. Because they can understand our environment to see what we see, listening to what we hear and interact with us throughout the day.
Open source
Aware that such superintelligence presents risks, the Meta CEO reveals that these systems may not be open source. “We must be rigorous to relieve these risks and be attentive to what we chose to do open source,” he said.
An important change, an activist goal for this approach that consists of offering AI models that can be used for free, downloaded and personalized in the AI field from the beginning. The flame family is, for example, open source, where Openai and other competitors offer closed models.
This announcement may surprise, but Mark Zuckerberg had already addressed this point in April 2024. “We are obviously very favorable for open source (…) If, however, at one point, the skills of the tool evolve qualitatively and that we believe that it is not responsible for making it open source, we will not do it,” he said in a vector, and added that it was “very difficult to plan.”
Source: BFM TV
