🇨🇳 Manus: Why Is Everyone Talking About the New AI Agent From China?
Manus is the first fully autonomous AI system developed in China. Many experts are calling it a breakthrough in the field of AI agents. However, the real-world experience of some early testers raises doubts about how revolutionary this innovation truly is.
Key Features of Manus:
➡️ Modular Architecture: Manus utilizes multiple models to perform different tasks (including Claude and Qwen). By the end of the year, the developers plan to open source some of its tools.
➡️ Performance: According to the developers, Manus outperforms OpenAI's research agent Deep Research in specialized benchmarks like GAIA.
➡️ Full Autonomy: Manus independently determines how to solve tasks and doesn't require user feedback. It can also accept commands during its operation—for example, you can ask it to change the format of its output results mid-task.
"This is a fully autonomous agent that bridges the gap between concept and execution. We see it as the next paradigm in human-machine interaction," says Yichao "Peak" Ji, co-founder and chief researcher at Manus AI.
⚙️ Where Manus Can Be Used
The developers claim that Manus can handle complex tasks requiring human-like thoughtfulness, data analysis, and decision-making among multiple options. For instance, Manus could plan a trip by booking tickets and preparing comprehensive travel guides.
It can also be used in education, as a business advisor, or as an analyst.
🤔 But What's the Reality?
Currently, Manus is in beta testing. Access to the platform is granted via special invitations, which has fueled significant hype around what some on X are calling the new DeepSeek. Invitation codes are even being sold on marketplaces, with prices ranging from $1 to $1,000.
However, many users have been disappointed with the agent. For example, a TechCrunch journalist tried to use Manus to order a burger delivery and book a flight to Japan. For the first task, after 10 minutes of processing, Manus returned an error. When searching for flights, it merely provided links to popular ticketing websites.
Other users on X report that Manus makes factual errors and often omits easily accessible information found online.
More on the topic:
⏩ Perplexity Deep Research: An AI agent for search — now available in our bot
⏩ OpenAI's Product Director Kevin Weil: On robots, AI agents, and accelerating progress
