Manus Ai and The Reality Behind the Hype
- 14/03/2025 01:17 AM
- Emma
Manus, a new AI platform that launched in preview, has been generating massive excitement, with some users calling it the most advanced AI tool they’ve ever seen. Its official Discord server quickly grew to over 138,000 members, and invitation codes are reportedly being sold for high prices on reseller platforms in China.
However, not everyone is convinced that Manus lives up to the hype.
What is Manus?
Developed by the Chinese startup Butterfly Effect, Manus is an "agentic AI" platform designed to automate complex tasks. It claims to go beyond typical AI chatbots by taking action, such as analyzing financial reports, drafting research papers, and even making purchases.
According to reports, Manus uses a combination of existing AI models, including Claude from Anthropic and Alibaba’s Qwen, rather than developing its own AI from scratch. Despite this, the company promotes Manus as a revolutionary tool capable of everything from buying real estate to building video games.
A viral video by Yichao "Peak" Ji, a research lead for Manus, suggested that the platform outperforms other AI-powered agents like OpenAI’s deep research and Operator. He claimed that Manus ranked highly on the GAIA benchmark, which evaluates an AI’s ability to browse the web, use software, and complete tasks.
Does Manus Really Work?
While some users are impressed, others report frustrating experiences with Manus. Many have encountered technical issues, errors, and incomplete tasks.
For example, AI startup co-founder Alexander Doria shared that he ran into endless loops and error messages while testing the platform. Others noted that Manus makes mistakes in factual answers, struggles to provide citations, and sometimes fails to find easily accessible information online.
A firsthand test of Manus revealed similar problems:
- When asked to order a fried chicken sandwich from a local restaurant, the AI crashed the first time. On the second attempt, it found a menu item but failed to complete the order or provide a checkout link.
- When tasked with booking a flight from New York to Japan, Manus could only generate a list of airline websites and travel search engines—some with broken links—rather than completing the booking process.
- A simple request to reserve a table at a nearby restaurant also failed.
- When asked to build a Naruto-inspired fighting game, the AI struggled for half an hour before returning an error.
The Reality Behind the Hype
If Manus still has so many issues, why has it gained so much attention?
One major factor is exclusivity. The limited availability of invites has created a sense of scarcity, driving up demand.
Chinese media outlets have also played a role in building excitement, with some publications calling Manus a major AI breakthrough. However, misinformation about its capabilities has spread online. A widely shared video falsely showed Manus performing complex tasks across multiple smartphone apps, which Ji later confirmed was not an actual demo of the platform.
Some social media influencers also compared Manus to DeepSeek, another Chinese AI company known for developing its own models. But unlike DeepSeek, Butterfly Effect has not built its own AI models and has not made its technology openly available.
What’s Next for Manus?
Manus is still in early access, and its developers acknowledge that there are issues to resolve. A spokesperson from the company said they are focused on improving the platform, scaling computing capacity, and fixing reported problems.
While the potential for agentic AI is exciting, the current version of Manus appears to be more hype than reality. For now, it serves as a reminder that AI development takes time, and not every "breakthrough" is as advanced as it seems.
