Butterfly Effect’s co-founder Yichao “Peak” Ji revealed in a launch video that the startup has been quietly developing its AI digital assistant Manus over the past year.
A new Chinese artificial intelligence agent, claimed to operate independently from humans, has sparked intense reactions among experts—ranging from concern to disappointment.
“We view it as the next paradigm of human-machine collaboration, and potentially a glimpse into AGI,” he said, referring to artificial general intelligence that aims to replicate human-like thinking.
Manus launched in an exclusive invitation-only phase last week, with access remaining highly limited.
Early reviews circulating on social media present starkly contrasting impressions:
“After gaining access, I can confirm… Manus stands as the most remarkable AI tool I’ve ever experienced,” posted Victor Mustar, Head of Product Design at Hugging Face, on X. “Its agentic capabilities are truly revolutionary, pushing the boundaries of possibility.”
However, critical feedback highlights several shortcomings:
- Struggles with basic functions like flight bookings
- Frequent error messages
- System loops that fail to resolve
- Cloud-based processing raising data security concerns
This launch comes amid intense debate about China’s growing AI leadership, particularly since DeepSeek’s emergence in January demonstrated Chinese models could rival OpenAI and Google’s offerings at substantially lower operational costs.
The current AI landscape has seen major players like Anthropic and OpenAI incorporate task-specific agent capabilities since late last year. Butterfly Effect positions Manus as uniquely capable of complex operations like:
✓ New York real estate transactions
✓ Podcast production
Yet TechCrunch’s Kyle Wiggers reported disappointing results during testing when requesting basic services:
- Failed sandwich orders
- Unsuccessful flight searches to Japan
The mixed reception underscores both the promise and growing pains of this new generation of autonomous AI agents.
China’s rapid AI advancements despite U.S. restrictions on exporting cutting-edge computer chips are causing concern in Silicon Valley.
The release of AI agents onto the internet without strict regulations raises fears of mishaps or abuses—such as stock market chaos caused by AI agents making factual errors.
Corpora.ai CEO Mel Morris did not view Manus as a “revolutionary leap” from existing AI models but warned that its ability to access remote computer servers poses potential risks to data confidentiality.