Generative AI is advancing at a rapid pace, with new platforms claiming to revolutionize the way we interact with technology. One such platform, Manus, launched in preview last week, generating an immense amount of buzz. With AI enthusiasts calling it “the most impressive AI tool ever tried” and invite codes selling for thousands on reseller platforms, expectations were sky-high. But does Manus live up to the hype?
The Rise of Manus AI
Manus, developed by Chinese company The Butterfly Effect, claims to be an “agentic” AI—designed to execute complex tasks autonomously, from drafting research papers to purchasing real estate. The platform allegedly outperforms competitors in benchmarks assessing AI’s ability to browse the web, utilize software, and complete multi-step processes.
Its rapid rise to prominence has been fueled by social media and Chinese media, which have dubbed Manus a breakthrough in AI technology. However, deeper investigation suggests the platform may not be as revolutionary as claimed.
Behind the Curtain: Manus’ Technical Reality
Despite its ambitious claims, Manus does not appear to be an entirely original creation. Reports suggest it relies on existing AI models, including Anthropic’s Claude and Alibaba’s Qwen, rather than a proprietary model developed from scratch. While fine-tuning existing models can improve performance, Manus’ reliance on external technology raises questions about its true uniqueness.
Moreover, a viral video implied that Manus could autonomously control various smartphone apps, leading to exaggerated claims about its capabilities. However, the company later confirmed that the video was misleading and did not accurately depict Manus’ current abilities.
User Experience: High Expectations, Disappointing Results
Early testers report significant limitations. Users have encountered frequent crashes, failed requests, and unreliable execution of tasks.
- Failed Automation: Requests to book flights or order food often result in incomplete transactions, with Manus struggling to finalize simple processes like online purchases.
- Inconsistent Performance: While Manus can identify flight prices, it fails to complete bookings or provide valid links.
- Error Messages & Loops: Several testers noted that the platform enters endless loops or errors out when asked to perform multi-step tasks.
AI startup founder Alexander Doria described his experience as frustrating, encountering multiple system failures. Others have pointed out that Manus sometimes provides incorrect information and fails to properly cite sources.
The Manus Hype Cycle
So, why is Manus still generating so much excitement? A major factor appears to be the exclusivity created by invite-only access, making the platform appear more desirable.
Additionally, some AI influencers have exaggerated its capabilities, with comparisons drawn to DeepSeek, a Chinese AI company that has developed its own foundation models. However, unlike DeepSeek, Manus does not appear to have built any core AI technologies in-house.
Company Response and the Road Ahead
In response to criticism, a Manus spokesperson acknowledged the platform’s current shortcomings, stating that the closed beta is focused on stress-testing the system and identifying issues. The company also cited efforts to scale computing power and enhance reliability.
While Manus is still in early development, its current state suggests that it is more of a work-in-progress than a revolutionary AI breakthrough. As the technology evolves, it remains to be seen whether Manus can transition from hype to genuinely useful innovation.
