In today’s technology landscape, the launch of Manus, an “agentic” AI platform, has ignited a wave of enthusiasm that rivals the frenzy surrounding a blockbuster concert. The fervor is palpable, with the head of product at Hugging Face labeling it “the most impressive AI tool I’ve ever tried,” while AI policy researcher Dean Ball touted it as “the most sophisticated computer using AI.” Social media’s response has been a whirlwind, as evidenced by the official Discord channel for Manus skyrocketing to over 138,000 members within mere days. To add a surreal twist, invite codes for Manus have garnered astronomical prices on Chinese reseller platforms like Xianyu. However, amid all this glitter, a critical examination sheds light on the possibility that the excitement may not be fully warranted.
Behind the Curtain: An Incomplete Masterpiece
Digging into Manus’s technological makeup reveals a blend of established AI models, utilizing components like Anthropic’s Claude and Alibaba’s Qwen. This intricacy raises a compelling question: how much originality can we truly attribute to Manus? While the platform claims it can perform an impressive array of tasks—from drafting research reports to real estate purchases—its execution sometimes appears half-baked. The company behind Manus, the Chinese startup Monica, seems to flaunt spectacular promises that may outstrip reality. A viral X video featuring Yichao “Peak” Ji suggested that Manus surpasses other agentic tools, claiming a stellar performance on the GAIA benchmark for general AI assistants. Yet, claims of a revolutionary “autonomous agent” seem to lack substantial backing when evaluated against user experiences.
The Human Experience: Expectations vs. Reality
User feedback is where the illusion begins to crack. One notable voice, Alexander Doria—the co-founder of AI startup Pleias—recounted encounters with error messages and frustrating loops while experimenting with Manus. Such red flags signal deep root problems in the platform’s reliability. Amid this chaos, other users have noted that the AI struggles with factual accuracy and often fails to appropriately cite sources. A particularly telling instance from user Derya Unutmaz highlights Manus’s inability to finish tasks that require sustained algorithmic prowess, ultimately crashing after an extended attempt. These glitch-ridden experiences contrast sharply with the platform’s grandiose claims of almost superhuman capabilities.
Where the Bottlenecks Lie
My own foray into the realm of Manus felt more like a tech letdown than a breakthrough. Tasked with something as uncomplicated as ordering a fried chicken sandwich, the AI crashed before completing the task. Subsequent attempts led to frustrating failures when Manus could not even provide a checkout link, let alone execute an order. Even something as crucial as booking a flight turned into a series of incorrect links or unfinished transactions, showcasing a significant disparity between what Manus purportedly can do and what it delivers.
This situational analysis points to a glaring issue in the AI’s task execution. A marvelous concept of “autonomy” has translated into a tool that struggles with straightforward tasks while also generating data and reports that, though impressive, fail to compensate for its inefficiencies.
The Mirage of Scarcity and Influencer Misrepresentation
The phenomena fueling Manus’s meteoric rise may have less to do with its actual performance and more with psychological manipulation through scarcity. By providing limited access, the company crafted an exclusive aura that naturally attracts attention. Chinese media outlets have readily embraced Manus as a groundbreaking innovation, but this echoes a more strategic narrative rather than an objective assessment of capabilities. Social media influencers seem enthusiastic yet occasionally misled, often comparing Manus to more established entities like DeepSeek without firm evidence to back such claims.
Whereas DeepSeek has been more transparent in sharing its technology with the public, Monica has held its innovations closer to chest. This lack of transparency raises ethical concerns in a rapidly evolving digital ecosystem, where truth is often overshadowed by lucrative marketing campaigns.
Tinkering with Failure: A Call for Accountability
Even as early access allows for user feedback, the frequent issues trigger concern about the trajectory of such a promising tool. The expectations surrounding Manus are stunning, but the disparity between anticipation and reality cannot be overlooked. Whatever the future holds for Manus AI, it is clear that the developers at Monica must address foundational issues if they wish to maintain their relevance and credibility in an industry that thrives on genuine progress and transparent delivery of promises. The tantalizing vision of a fully autonomous AI agent now hinges on the balance of hope and the stark accountability of user experiences—a balancing act that will ultimately determine its success or failure in the competitive AI landscape.