"Superpowers of Invention for Everyone"... Zuckerberg Envisions an Era of 'One AI Agent per Person'

[Mexico City = Shim Young-jae, Correspondent] Mark Zuckerberg, CEO of Meta Platforms, has proposed providing free personal artificial intelligence (AI) agents to billions of people. He envisions a future where AI supports coding, video production, product design, and scientific research, allowing individuals to realize new businesses and ideas with minimal capital and personnel.

On the 11th (local time), Bitcoin.com reported that CEO Zuckerberg stated in a 6,500-word essay titled "The Future is for Everyone" on the 10th, "Everyone will soon have the superpower of invention." Meta plans to back this initiative with a capital expenditure of up to $145 billion this year.

"Superpowers of Invention for Everyone"... Concept of Personal AI Agents

The core of CEO Zuckerberg's proposal is to provide AI not just as an automation tool for businesses or governments but as a means to enhance individual capabilities.

According to Bitcoin.com, CEO Zuckerberg predicts that the barriers to resources and expertise needed for coding, video production, product design, and scientific research will be significantly lowered through AI.

He shared an example of his 8-year-old daughter coding ideas and producing videos herself. He also mentioned that Meta's researchers are using AI to create new crystal structures for augmented reality (AR) glasses.

CEO Zuckerberg stated, "Everyone will soon have the superpower of invention." His vision is for personal AIs to evolve from merely answering questions to performing actual tasks on behalf of users.

The agent capabilities of Meta AI include planning tasks, connecting to external applications, and executing multi-step operations.

CEO Zuckerberg said, "Everyone will have a personal agent with an exceptional ability to understand themselves, their goals, and everything they consider important."

He believes that the expansion of personal AI could change the way startups are formed. As the funding and personnel needed to test new ideas decrease, meaningful businesses can be built without large capital or organizations.

Open Source is Key... Public AI Running on Personal PCs

CEO Zuckerberg proposed open source as a key measure to prevent AI capabilities from concentrating in a few institutions.

According to Bitcoin.com, he argued that restricting access to advanced AI models could reduce opportunities while further centralizing technological power.

On the same day, Meta also unveiled "Muse Glimmer."

Muse Glimmer is an agent-type AI model with 30 billion parameters. It has been released under the Apache 2.0 license and has been lightweight enough to run on a Mac or PC equipped with a single consumer-grade graphics processing unit (GPU), as reported by Bitcoin.com.

Meta previously released Muse Spark and plans to unveil the weights of Muse Spark 1.2 soon.

CEO Zuckerberg argued that the U.S. government should relax regulations on AI training data and model distillation. He cited that U.S. open-source models bear regulatory burdens that foreign competitors do not.

Meta's AI strategy aligns with enabling models to be used on personal devices rather than concentrating all computations in the cloud.

Up to $145 Billion Investment This Year... Expanding AI Infrastructure

Providing personal AI to billions of people requires massive computing infrastructure.

According to Bitcoin.com, Meta has set its capital expenditure for this year at up to $145 billion, a significant portion of which will be invested in expanding AI infrastructure and data center capacity. However, this capital expenditure will not be solely allocated to personal superintelligence projects.

The expansion of data centers also impacts the communities that host them.

CEO Zuckerberg proposed investing in jobs, schools, energy, and public services in the areas surrounding data centers so that communities can directly share in the economic benefits.

He stated, "Everyone will be able to use these tools for free or at a low cost." However, he also suggested that services requiring substantial computing resources might have their prices determined through a dynamic auction system.

A separate approach for privacy protection was also mentioned. In the fully private mode envisioned by Meta, companies or other service providers would not have access to the data processed by personal AI agents.

AI Agents Transitioning from 'Conversation' to 'Action'

AI agents are software that pursue specific goals on behalf of users and interact with digital services.

Bitcoin.com explained that AI agents in the financial sector could interact with blockchain systems and manage digital assets or execute pre-set transactions.

Platforms for AI agents to interact with each other are also emerging.

Meta acquired the experimental AI agent social network "Moltbook." According to Bitcoin.com, Moltbook has about 200,000 registered verified AI agents connected to human owners.

The pace of AI model development is also accelerating. According to Zuckerberg's post, 267 AI models are registered in various rankings, and major research institutions are shifting from annual large-scale model releases to weekly updates.

CEO Zuckerberg warned against the concentration of rapidly advancing AI capabilities in specific companies or institutions. His proposed approach involves a decentralized structure where multiple personal AIs, companies, and cutting-edge AI research labs compete and check each other.

Preventing Concentration of AI Power... Proposing Sharing Intermediate Models with the Government

Meta plans to apply a separate governance structure to the process of releasing AI models.

According to Bitcoin.com, Meta's independent board will approve safety standards for model releases and review compliance. This structure prevents any single executive, including CEO Zuckerberg, from monopolizing decisions on model distribution.

Zuckerberg also proposed strengthening information sharing with the government.

He argued that cutting-edge AI research labs should share intermediate checkpoints of models before training is completed with the government. Checkpoints are materials that save the state at specific points during the model training process.

Bitcoin.com explained that this proposal goes a step further than the June 2 executive order, which allowed agencies to establish voluntary systems for up to 30 days of early access to models before public release.

It was also noted that restrictions on AI access in the U.S. could lead to increased use of foreign models.

According to Bitcoin.com, after export controls were applied to Anthropic's highest-performing system in June, U.S. companies and overseas users shifted to China's open-source AI models. The U.S. Department of Commerce lifted those restrictions at the end of the same month.

The future envisioned by CEO Zuckerberg is one where AI becomes an infrastructure that directly enhances individuals' coding, entrepreneurship, creativity, and research abilities, rather than just being a productivity tool for a few companies.

To achieve this, Meta is simultaneously pursuing open-source models, personal AI agents, and large-scale data center investments. CEO Zuckerberg sees providing powerful AI widely to individuals as a way to prevent the concentration of technological power.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.