GuShiio.com In the alternating heat wave of capital and cold wave of technology, how can China AI companies maintain their original intention of innovation? Perhaps the answer lies in the simple words of DeepSeek founder Liang Wenfeng that there are no shortcuts in technology, in the late night when every engineer debuts models, and in each line of refactoring code. When the short-term hot searches and highlights dissipate, what really remains must be those companies that engrave innovation into their genes.
“Science and Technology Innovation Board Daily”, March 7 (commentator Tian Ye)An AI agent application called Manus suddenly exploded in the technology circle recently. Its internal beta invitation code was speculated to a high price of nearly 100,000 yuan on second-hand platforms, and A-share agent concept stocks collectively went up and down. Behind this phenomenal communication lies the public’s expectation and anxiety for AI technology breakthroughs: When the prediction of the “first year of AI Agent” encounters commercial reality, how should we view this technological carnival?
丨 Carnival and questioning: The double-sided mirror of AI innovation
Manus’s popularity stems from its positioning as “the world’s first universal AI agent.” In the demonstration video, it can independently complete complex tasks such as resume screening, stock analysis, and travel planning, and even generate PPT, code and other results. In the GAIA benchmark test, its performance surpassed the OpenAI peer model, which led some people to call it and DeepSeek the “twin stars” of China AI. However, controversy ensued: the invitation code mechanism was accused of “hunger marketing”, the technical principle was questioned as “shell integration”, and some even believed that it was just an “advanced plug-in” that did not break through the existing framework.
This division of public opinion reflects the deep contradictions in the AI field. On the one hand, the public is eager to see real “killer applications” that bring AI from the laboratory into daily life; on the other hand, the industry’s criteria for judging technological innovation have not yet been unified, and the boundaries between “disruptive breakthroughs” and “incremental optimization” are blurred. Manus’s explosion is just like a magnifying glass, reflecting society’s complex mentality of both expectations and vigilance for AI innovation breakthroughs.
丨 Technological deconstruction: Integration capabilities are also productive forces
Manus’s core technology lies in the Multiple Agent System and asynchronous processing in the cloud. It automates the entire process of tasks through division of labor and cooperation between planning agents, execution agents, and verification agents, and calls tools such as browsers and code editors in the virtual environment. This “software robot” model essentially transforms the logical reasoning capabilities of large models into productivity. Its value lies in lowering the threshold of the tool chain and allowing users to complete complex operations without programming.
However, technological integration is not simply “take-away”. The Manus team achieved the leap from “conversational AI” to “collaborative AI” by optimizing task disassembly algorithms, improving tool invocation efficiency, and enhancing multi-round interactive experiences. Although its underlying model may rely on third-party technology, interface design, task flow planning, user feedback mechanisms, etc. all require deep engineering capabilities. As some experts said: “Integration capabilities are themselves an important part of technological innovation.”
丨 Innovation inspiration: Avoid “technological romanticism” and reject “commercial short-sightedness”
The Manus controversy is essentially a microcosm of the development path of the AI industry. Currently, global AI has entered a “application deepening period”, and technological dividends have spread from large models to vertical scenarios. This requires both basic model breakthroughs like DeepSeek and application innovations like Manus, both of which are indispensable. But the industry needs to be wary of two tendencies:
First, avoid “technical romanticism”. Technological breakthroughs are not castles in the air and need to be built on a solid foundation of research and development. Manus’s success lies in accurately capturing user pain points, but the limitations of its predefined processes cannot be ignored. Real AI Agents should have stronger autonomous decision-making capabilities rather than relying solely on a predetermined tool chain.
Second, reject “commercial shortsightedness”. The invitation code mechanism may be a helpless move by the technical team, but it objectively exacerbates market anxiety. AI innovation takes time to settle, and companies should balance business rhythm and technological depth to avoid falling into the vicious cycle of “concept hyping-rapid monetization”.
The controversy and discussions triggered by Manus provide a precious mirror for the industry. The enthusiasm and controversy caused by the invitation code will eventually ebb, but the thoughts left should not dissipate: In the alternating heat wave of capital and the cold wave of technology, how can China AI companies maintain their innovative intentions? Perhaps the answer lies in the simple words of DeepSeek founder Liang Wenfeng that there are no shortcuts in technology, in the late night when every engineer debuts models, and in each line of refactoring code. When the short hot searches and highlights fade away, what really remains must be those companies that engrave innovation into their genes.
