AI Assistant Manus Makes Breakthrough Progress, Sparking In-Depth Discussions on Security and Efficiency

Recently, the AI assistant Manus achieved breakthrough results in the GAIA benchmark test, outperforming large language models of the same tier. Manus demonstrated the ability to independently complete complex tasks, such as multinational business negotiations, which involve multiple aspects including contract clause breakdown, strategic forecasting, and proposal generation. Compared to traditional systems, Manus's advantages lie in dynamic goal decomposition, cross-modal reasoning, and memory-enhanced learning capabilities. It can break down large tasks into hundreds of executable subtasks while handling various types of data and continuously improving decision-making efficiency and reducing error rates through reinforcement learning.

The progress of Manus has once again sparked discussions in the industry about the development path of AI: Is the future heading towards a unified model of Artificial General Intelligence (AGI), or a collaborative model of Multi-Agent Systems (MAS)? This question arises from the design philosophy of Manus, which suggests two possibilities: one is the AGI path of continuously enhancing the individual intelligence level to approach human comprehensive decision-making capability; the other is the MAS path of acting as a super coordinator, directing thousands of specialized agents to work together.

This discussion actually touches on the core contradiction of AI development: how to balance efficiency and safety? As individual intelligences get closer to AGI, the risk of opacity in their decision-making processes also increases. On the other hand, while multi-agent collaboration can disperse risks, it may miss key decision-making opportunities due to communication delays.

The development of Manus also highlights the inherent security risks of AI. For example, in medical scenarios, it needs access to sensitive genomic data of patients; in financial negotiations, it may come into contact with undisclosed financial information of companies. Furthermore, AI systems may have algorithmic biases, such as giving unfair salary recommendations to specific groups during the hiring process. In terms of legal contract review, the misjudgment rate for terms in emerging industries may also be relatively high. Even more concerning is that hackers may interfere with Manus's judgment during negotiations by implanting specific audio frequencies.

These issues highlight a concerning reality: the more intelligent AI systems become, the broader their potential attack surface.

In the Web3 space, security has always been a highly discussed topic. The "impossible triangle" theory proposed by Ethereum founder Vitalik Buterin (that blockchain networks cannot simultaneously achieve security, decentralization, and scalability) has inspired the development of various cryptographic technologies:

1. Zero Trust Security Model: Emphasizes strict authentication and authorization for every access request.
2. Decentralized Identity (DID): Allows entities to obtain verifiable identity recognition without the need for centralized registration.
3. Fully Homomorphic Encryption (FHE): Allows computation on encrypted data, enabling data processing while protecting privacy.

Among these technologies, fully homomorphic encryption is considered a key technology to solve security issues in the AI era. It can play a role in the following aspects:

1. Data Layer: All information input by users (including biometric features, voice tone) is processed in an encrypted state, such that even the AI system itself cannot decrypt the original data.
2. Algorithm Level: Achieve "encrypted model training" through FHE, ensuring that even developers cannot peek into the decision-making process of AI.
3. Collaborative Level: Communication between multiple agents uses threshold encryption, so even if a single node is compromised, it will not lead to global data leakage.

In the Web3 security field, several projects have been dedicated to addressing these issues. For example, uPort is one of the earlier decentralized identity projects launched on the Ethereum mainnet; NKN has made attempts in the zero-trust security model; Mind Network is the first FHE project to go live on the mainnet and has established partnerships with several well-known institutions.

As AI technology approaches human intelligence levels, establishing a robust defense system becomes increasingly important. Fully Homomorphic Encryption (FHE) not only addresses current security issues but also lays the groundwork for the development of strong AI in the future. On the path to AGI, FHE is no longer an option but a necessary condition to ensure the safe operation of AI systems.