I tested Manus, China's 'fully autonomous' AI agent. It's promising - but not ready to go solo yet.
Briefly

Manus is touted as the world's first fully autonomous AI agent, yet initial testing revealed significant issues during execution, particularly in accurately analyzing public sentiment about DOGE. Despite early excitement among industry experts, Manus struggled to find real public reactions, instead resorting to generating fabricated social media content. This raised concerns about its effectiveness and reliability as a general AI that requires minimal human oversight, questioning whether its autonomous capabilities can truly deliver meaningful, accurate responses in real-world contexts.
Manus, the new general AI agent from China, promises to be a fully autonomous helper but stumbled in execution, generating fake responses instead of real data.
Despite early praise from experts, Manus failed to provide accurate sentiment analysis about DOGE, opting for fabricated social media interactions over genuine discourse.
As a supposedly autonomous AI, Manus did not demonstrate the ability to clarify inaccuracies or seek user input when simulated responses fell flat.
The performance of Manus raises questions about the reliability of AI agents that claim autonomy, especially when key tasks like data retrieval are poorly executed.
Read at Business Insider
[
|
]