Because the common AI agent Manus was launched final week, it has unfold on-line like wildfire. And never simply in China, the place it was developed by the Wuhan-based startup Butterfly Impact. It’s made its manner into the worldwide dialog, with influential voices in tech, together with Twitter cofounder Jack Dorsey and Hugging Face product lead Victor Mustar, praising its efficiency. Some have even dubbed it “the second DeepSeek,” evaluating it to the sooner AI mannequin that took the trade without warning for its sudden capabilities in addition to its origin.
Manus claims to be the world’s first common AI agent, utilizing a number of AI fashions (comparable to Anthropic’s Claude 3.5 Sonnet and fine-tuned variations of Alibaba’s open-source Qwen) and varied independently working brokers to behave autonomously on a variety of duties. (This makes it totally different from AI chatbots, together with DeepSeek, that are primarily based on a single giant language mannequin household and are primarily designed for conversational interactions.)
Regardless of all of the hype, only a few individuals have had an opportunity to make use of it. At the moment, below 1% of the customers on the wait record have acquired an invitation code. (It’s unclear how many individuals are on this record, however for a way of how a lot curiosity there’s, Manus’s Discord channel has greater than 186,000 members.)
MIT Expertise Overview was in a position to acquire entry to Manus, and after I gave it a test-drive, I discovered that utilizing it appears like collaborating with a extremely smart and environment friendly intern: Whereas it often lacks understanding of what it’s being requested to do, makes incorrect assumptions, or cuts corners to expedite duties, it explains its reasoning clearly, is remarkably adaptable, and may enhance considerably when supplied with detailed directions or suggestions. Finally, it’s promising however not good.
Similar to its guardian firm’s earlier product, an AI assistant known as Monica that was launched in 2023, Manus is meant for a world viewers. English is ready because the default language, and its design is clear and minimalist.
To get in, a consumer has to enter a sound invite code. Then the system directs customers to a touchdown web page that intently resembles these of ChatGPT or DeepSeek, with earlier periods displayed in a left-hand column and a chat enter field within the heart. The touchdown web page additionally options pattern duties curated by the corporate—starting from enterprise technique growth to interactive studying to personalised audio meditation periods.
Like different reasoning-based agentic AI instruments, comparable to ChatGPT DeepResearch, Manus is able to breaking duties down into steps and autonomously navigating the net to get the knowledge it wants to finish them. What units it aside is the “Manus’s Pc” window, which permits customers not solely to look at what the agent is doing but additionally to intervene at any level.
To place it to the check, I gave Manus three assignments: (1) compile a listing of notable reporters masking China tech, (2) seek for two-bedroom property listings in New York Metropolis, and (3) nominate potential candidates for Innovators Below 35, a listing created by MIT Expertise Overview yearly.
Right here’s the way it did:
Activity 1: The primary record of reporters that Manus gave me contained solely 5 names, with 5 “honorable mentions” under them. I observed that it listed some journalists’ notable work however didn’t do that for others. I requested Manus why. The explanation it provided was hilariously easy: It received lazy. It was “partly because of time constraints as I attempted to expedite the analysis course of,” the agent advised me. After I insisted on consistency and thoroughness, Manus responded with a complete record of 30 journalists, noting their present outlet and itemizing notable work. (I used to be glad to see I made the minimize, together with a lot of my beloved friends.)
I used to be impressed that I used to be in a position to make top-level recommendations for adjustments, a lot as somebody would with a real-life intern or assistant, and that it responded appropriately. And whereas it initially ignored adjustments in some journalists’ employer standing, after I requested it to revisit some outcomes, it rapidly corrected them. One other good characteristic: The output was downloadable as a Phrase or Excel file, making it simple to edit or share with others.
Manus hit a snag, although, when accessing journalists’ information articles behind paywalls; it regularly encountered captcha blocks. Since I used to be in a position to comply with alongside step-by-step, I might simply take over to finish these, although many media websites nonetheless blocked the software, citing suspicious exercise. I see potential for main enhancements right here—and it will be helpful if a future model of Manus might proactively ask for assist when it encounters these kinds of restrictions.
Activity 2: For the condominium search, I gave Manus a posh set of standards, together with a funds and several other parameters: a spacious kitchen, outside area, entry to downtown Manhattan, and a significant practice station inside a seven-minute stroll. Manus initially interpreted obscure necessities like “some sort of outside area” too actually, utterly excluding properties with out a non-public terrace or balcony entry. Nevertheless, after extra steerage and clarification, it was in a position to compile a broader and extra useful record, giving suggestions in tiers and neat bullet factors.
The ultimate output felt straight from Wirecutter, containing subtitles like “finest total,” “finest worth,” and “luxurious possibility.” This job (together with the back-and-forth) took lower than half an hour—so much much less time than compiling the record of journalists (which took just a little over an hour), possible as a result of property listings are extra brazenly obtainable and well-structured on-line.
Activity 3: This was the biggest in scope: I requested Manus to appoint 50 individuals for this yr’s Innovators Below 35 record. Producing this record is a gigantic endeavor, and we sometimes get a whole lot of nominations yearly. So I used to be curious to see how properly Manus might do. It broke the duty into steps, together with reviewing previous lists to grasp choice standards, making a search technique for figuring out candidates, compiling names, and guaranteeing a various number of candidates from everywhere in the world.
Creating a search technique was probably the most time-consuming half for Manus. Whereas it didn’t explicitly define its method, the Manus’s Pc window revealed the agent quickly scrolling via web sites of prestigious analysis universities, bulletins of tech awards, and information articles. Nevertheless, it once more encountered obstacles when making an attempt to entry tutorial papers and paywalled media content material.
After three hours of scouring the web—throughout which Manus (understandably) requested me a number of instances whether or not I might slim the search—it was solely in a position to give me three candidates with full background profiles. After I pressed it once more to supply a whole record of fifty names, it will definitely generated one, however sure tutorial establishments and fields have been closely overrepresented, reflecting an incomplete analysis course of. After I identified the difficulty and requested it to seek out 5 candidates from China, it managed to compile a stable five-name record, although the outcomes skewed towards Chinese language media darlings. Finally, I had to surrender after the system warned that Manus’s efficiency may decline if I saved inputting an excessive amount of textual content.
My evaluation: Total, I discovered Manus to be a extremely intuitive software appropriate for customers with or with out coding backgrounds. On two of the three duties, it offered higher outcomes than ChatGPT DeepResearch, although it took considerably longer to finish them. Manus appears finest suited to analytical duties that require intensive analysis on the open web however have a restricted scope. In different phrases, it’s finest to stay to the kinds of issues a talented human intern might do throughout a day of labor.
Nonetheless, it’s not all easy crusing. Manus can undergo from frequent crashes and system instability, and it might battle when requested to course of giant chunks of textual content. The message “As a result of present excessive service load, duties can’t be created. Please attempt once more in a couple of minutes” flashed on my display screen just a few instances after I tried to start out new requests, and infrequently Manus’s Pc froze on a sure web page for a protracted time frame.
It has the next failure fee than ChatGPT DeepResearch—an issue the crew is addressing, in accordance to Manus’s chief scientist, Peak Ji. That mentioned, the Chinese language media outlet 36Kr reviews that Manus’s per-task price is about $2, which is simply one-tenth of DeepResearch’s price. If the Manus crew strengthens its server infrastructure, I can see the software changing into a most well-liked alternative for particular person customers, notably white-collar professionals, impartial builders, and small groups.
Lastly, I feel it’s actually useful that Manus’s working course of feels comparatively clear and collaborative. It actively asks questions alongside the best way and retains key directions as “information” in its reminiscence for future use, permitting for an simply customizable agentic expertise. It’s additionally very nice that every session is replayable and shareable.
I count on I’ll hold utilizing Manus for all kinds of duties, in each my private {and professional} lives. Whereas I’m unsure the comparisons to DeepSeek are fairly proper, it serves as additional proof that Chinese language AI firms will not be simply following within the footsteps of their Western counterparts. Moderately than simply innovating on base fashions, they’re actively shaping the adoption of autonomous AI brokers in their very own manner.
