Manus, an “agentic” AI platestablish that begined in pappraise last week, is generating more hype than a Taylor Swift concert.
The head of product at Hugging Face called Manus “the most amazeive AI tool I’ve ever tried.” AI policy researcher Dean Ball depictd Manus as the “most polishd computer using AI.” The official Discord server for Manus grew to over 138,000 members in fair a scant days, and ask codes for Manus are inestablishedly selling for thousands of dollars on Chinese reseller app Xianyu.
But it’s not evident the hype is fairified.
excellent https://t.co/TfeV9QZ1d0
— jack (@jack) March 9, 2025
Manus wasn’t broadened entidepend from scratch. According to inestablishs on social media, the platestablish engages a combination of existing and fine-tuned AI models, including Anthropic’s Claude and Alibaba’s Qwen, to carry out tasks such as writeing research inestablishs and analyzing financial filings.
Yet on its website, Monica — the Chinese commenceup behind Manus — gives a scant untamed examples of what the platestablish presumedly can accomplish, from buying authentic estate to programming video games.
In a viral video on X, Yichao “Peak” Ji, a research direct for Manus, implied that the platestablish was better to agentic tools such as OpenAI’s proset up research and Operator. Manus outcarry outs proset up research on a well-comprehendn benchtag for ambiguous AI helpants called GAIA, Ji claimed, which probes an AI’s ability to carry out labor by browsing the web, using gentleware, and more.
“[Manus] isn’t fair another chatbot or laborflow,” Ji shelp in the video. “It’s a finishly autonomous agent that bridges the gap between conception and execution […] We see it as the next paradigm of human-machine collaboration.”
But some punctual engagers say that Manus is no panacea.
Alexander Doria, the co-set uper of AI commenceup Pleias, shelp in a post on X that he come apassed error messages and finishless loops while testing Manus. Other X engagers pointed out that Manus produces misconsents on factual inquires and doesn’t stablely cite its labor — and frequently leave outes adviseation that’s easily set up online.
Deep Research finished in under 15 minutes. Unblessedly, Manus AI fall shorted after 50 minutes at step 18/20! 😑 It was carry outing quite well-I was watching Manus’ output & it seemed excellent. However, running the same prompt a second time is a bit frustrating as it consents too lengthy! https://t.co/bGtmOI65CP
— Derya Unutmaz, MD (@DeryaTR_) March 8, 2025
My own experience with Manus hasn’t been incredibly chooseimistic.
I asked the platestablish to deal with what seemed to me enjoy a pretty straightforward ask: order a fried chicken sandwich from a top-rated rapid food unitet in my dedwellry range. After about ten minutes, Manus crashed. On the second try, it set up a menu item that met my criteria, but Manus couldn’t finish the ordering process — or provide a verifyout join, even.
Manus analogously whiffed when I asked it to book a fweightless from NYC to Japan. Given teachions that I thought didn’t depart much room for amhugeuity (e.g. “see for a business-class fweightless, prioritizing price and changeable dates”), the best Manus could do was serve up joins to fares apass disjoinal airline websites and airfare search engines enjoy Kayak, some of which were broken.
Hoping the next scant tasks might be the charm, I tageder Manus to reserve a table for one at a restaurant wislfinisher walking distance. It fall shorted after a scant minutes. Then I asked the platestablish to produce a Naruto-backd combat game. It errored out half an hour in, which is when I determined to throw in the towel.
Honest opinion after trying Manus AI for the last 3 days, here’s the excellent and the terrible.
Good:
– The research it does on the internet and the inestablishs it produces are incredible.
– Its ability to run scripts behind the scenes to carry out tasks is amazeive.
– The set ups it…— AshutoshShriimmenseava (@ai_for_success) March 9, 2025
We’ve accomplished out to Monica for comment and will modernize this post if we hear back.
So if Manusis is droping low of its technical promises, why did it blow up? A scant factors gived, such as the exclusivity produced by a scarcity of asks.
Chinese media was speedy to tout Manus as an AI shatterthcimpolite; unveilation QQ News called it “the pride of domestic products.” Meanwhile, AI swayrs on social media spread deceiveation about Manus’ capabilities. A widely-splitd video showed a desktop program, ostensibly Manus, taking action apass multiple cleverphone apps. Ji verifyed that the video wasn’t, in fact, a demo of Manus.
Other intransmitial AI accounts on X sought to draw comparisons between Manus and Chinese AI company DeepSeek — comparisons not necessarily rooted in fact. Monica didn’t broaden in-hoengage models, unenjoy DeepSeek. And while DeepSeek made many of its technologies uncoverly engageable, Monica hasn’t — at least not quite yet.
To be unprejudiced to Monica, Manus is in punctual access. The company claims it’s laboring to scale computing capacity and mend rerents as they’re inestablished. But as the platestablish currently exists, Manus eunites to be a case of hype running ahead of technoreasonable innovation.
