Workflow
360安全云智能体
icon
Search documents
周鸿祎:大模型缺“手和脚”,进化到智能体是必然
"智能体是大模型的'手和脚',可以直接干活,大模型进化到智能体是必然的",在8月6日开幕的第十三 届互联网安全大会(ISC.AI 2025)期间,360创始人周鸿祎在大会演讲中提出,AI发展如果停留在大模 型阶段,只能成为玩具,而非生产力工具。 周鸿祎称,智能体才是解决之道,"大模型进化到智能体是必然的。"它能够理解目标、规划任务、调用 工具、具备记忆,实现从需求到结果的完整交付。其核心在于使用工具的能力和利用大模型进行复杂任 务规划的能力。 谈及智能体演进的路径,周鸿祎分析,L1级别的智能体属于"玩具级"智能体,主要充当聊天助手;L2 低代码工作流智能体,已经从"玩具"进化为"工具",但必须由人类来设置流程,AI来执行任务,人再通 过操作工具提高生产效率;L3推理型智能体,已能实现AI自主规划完成任务,但仍会因缺乏协同规划 能力而陷入瓶颈;L4级智能体则发展为多智能体蜂群,多个专家智能体可像搭积木一样灵活"拉群组 队"、多层嵌套、分工协作。 转自:北京日报客户端 在网络安全领域,智能体对网络安全正逐渐形成颠覆性影响。周鸿祎表示,当前企业面临双重挑战,一 方面,安全运营专家稀缺且培养周期长;另一方面,"智能体 ...
晚报 | 8月7日主题前瞻
Xuan Gu Bao· 2025-08-06 14:14
Robotics - Unitree Technology has launched a new quadruped robot, Unitree A2, weighing approximately 37 kilograms with a range of 20 kilometers and a maximum speed of 5 meters per second [1] - Major technology companies like Tesla, Huawei, and Figure AI are investing in humanoid robotics, which is expected to accelerate industry advancements and commercialization [1] - The global market for humanoid robots is projected to exceed $150 billion by 2035, driven by policy support, technological maturity, and increasing demand [1] Computing Power - Huawei announced the full open-source of its Ascend hardware and CANN, enabling users to develop custom applications [2] - The national integrated computing power network has completed the release of nine technical documents, marking a transition from planning to application [2] - The computing power industry is experiencing high growth, with potential valuation increases due to ongoing demand and supply chain dynamics [2] Autonomous Driving - Tesla is advancing its Full Self-Driving (FSD) technology, with a new model expected to be released by the end of next month, featuring ten times the parameters of the current version [3] - The FSD system has shown strong performance in complex driving scenarios, and its deployment in the U.S. is planned for this year [3] - The introduction of FSD in the domestic market is anticipated to accelerate the development of the intelligent driving industry in China [3] Nano-Imprinting - China has successfully developed its first PL-SR series inkjet stepper nano-imprinting equipment, breaking foreign monopolies in high-end semiconductor manufacturing [4] - The new equipment supports nano-imprinting lithography processes with linewidths smaller than 10nm, surpassing similar products from international competitors [4] - This technology is expected to reduce equipment investment costs by 60% compared to traditional EUV lithography [4] Environmental Protection - A new round of global negotiations on plastic pollution is underway, aiming to establish a legally binding international agreement [5] - The global waste plastic market is projected to exceed $137 billion by 2025, with China being the largest producer of plastic [5] - The recycling and regeneration market in China is expected to surpass 400 billion yuan by 2030, driven by chemical recycling and biodegradable materials [5] Macro and Industry News - The National Development and Reform Commission and the National Energy Administration have issued basic rules for electricity market measurement and settlement, marking a significant step in market construction [6] - The Ministry of Transport aims to complete the construction of 300,000 kilometers of new and renovated rural roads by 2027 [7] - Shanghai's government has released a development plan for the embodied intelligence industry, targeting a core industry scale of 50 billion yuan [8]
周鸿祎:现阶段智能体竞争的唯一护城河是执行力
Tai Mei Ti A P P· 2025-08-06 11:42
Core Insights - The rapid evolution of AI agents leads to a very short product lead time, with companies needing to focus on execution and adaptability to stay competitive [2] - The concept of "Swarm L4" categorizes AI agents into five levels, with increasing complexity and application value as the level rises [3] - Single AI agents face significant limitations in task execution, while multi-agent swarm collaboration shows a high success rate and efficiency in completing complex tasks [5] Group 1: AI Agent Development - The competitive edge in the AI agent industry lies in the ability to quickly iterate and update products, rather than just launching them [2] - The "Swarm L4" framework indicates that higher-level agents can handle more complex projects, enhancing their task processing capabilities [3] Group 2: Multi-Agent Collaboration - Multi-agent systems can execute up to 1000 steps with a success rate of 95.4%, showcasing their effectiveness in complex task execution [5] - Challenges in multi-agent collaboration include task allocation and communication costs, but the benefits outweigh these difficulties [5] Group 3: Human-Machine Collaboration - The "human-in-the-loop" principle emphasizes the importance of user oversight in AI operations, allowing for decision-making and risk reduction [6] - The unpredictability of AI outputs necessitates a collaborative approach where humans guide AI execution, enhancing overall efficiency [6] Group 4: Specialized vs. General AI Agents - Specialized AI agents focusing on single domains are more effective than general-purpose agents, which struggle to excel in multiple areas [7][8] - General AI agents are suitable for repetitive tasks, while specialized agents provide more precise and efficient services for creative tasks [8] Group 5: Cybersecurity Challenges - The rise of AI agents introduces new cybersecurity threats, with the emergence of "super hackers" capable of automating attacks using AI [9] - Companies are encouraged to deploy security AI agents to counteract these threats, acting as digital counterparts to human security experts [9][10] Group 6: 360's AI Initiatives - 360 is advancing its entire product line towards AI integration, with the "AI Factory" enabling customized security AI agents for various scenarios [10] - Data shows that security AI agents significantly outperform traditional human services in threat detection and operational efficiency [10]