Manus quietly released its 2.0 declaration

Author: Hualin Dance King, Geek Park

Editor: Jingyu, Geek Park

On October 26, Zhang Tao (online name HideCloud), co-founder and product partner of Manus, recently wrote a lengthy article on his public account "Qianyun Thoughts" titled "Transforming Manus with Manus".

The article begins with a striking scene that left a deep impression on him. He provided several requirements and bugs to the engineers, and when he checked the progress later, he found a row of Codex and Manus sessions open on the engineer's computer. Clicking on one, he saw the exact wording of the requirements he had just sent on Slack, unchanged.

What surprised him even more was that this engineer was simultaneously handling a dozen or twenty requirements. This has become the norm at Manus today.

The article is packed with information, and several points are particularly noteworthy.

Internally, Manus has created a unified code entry called Manus-mono, accompanied by a manual written for AI, allowing AI to see the global view across repositories. One requirement can be completed by one person across all ends. The mobile team has only seven people, responsible for two products and four apps, with progress synchronized with the web version.

All product managers have submitted code to Manus, with about half able to independently bring their requirements online. Some requirements submitted by product managers have transformed from documents into videos.

The marketing team has a "digital colleague" that arranges its own work, proactively reminds about project progress, and clarifies functional details across teams.

In September, the company's AI expenditure per capita exceeded $10,000, with a few individuals spending $70,000 to $80,000 in a month. Zhang Tao specifically addressed external doubts in one section.

The timing of the article's release is also subtle. At the end of September, Manus released version 2.0 and the personal AI application Cue; at the beginning of October, the parent company Butterfly Effect announced the completion of a new round of financing exceeding $500 million, while opening 17 positions in Beijing.

Viewed at this juncture, the article resembles a declaration from the Manus team entering the 2.0 era. It discusses how the organization utilizes AI to evolve, revealing the company's product direction, cost logic, and hiring standards.

01 Manus itself is a model room for 2.0

If you compare the internal practices in the article with the feature list of Manus 2.0, you will find they almost correspond one-to-one.

The digital colleague in the marketing team finds the internal website inconvenient and can directly submit requirements through a feedback interface, automatically triggering Manus's development tasks as soon as a requirement comes in. This utilizes the 2.0 Automations, where receiving an external notification automatically starts a task.

Product managers replace requirement documents with videos, relying on the significantly enhanced video capabilities of 2.0, which can generate videos from images and create animations from code. The product manager responsible for the voice-to-text feature has a typical story. She first had Manus test over a dozen models, selected one for integration, and then created a demo according to existing design specifications, generating images for various scenarios users might encounter, checking and modifying them one by one, and finally editing the demo into a video. Once the plan was clear, she simply completed the functionality, with both the web and desktop interfaces designed by her.

Colleagues' Slack bots run on cloud computers. Non-development colleagues wanting to know how a feature is implemented, where the copy is, or whether it has gone live, can directly ask it without first inquiring who is in charge. This corresponds to the cloud computers that can be resident 7×24 hours in 2.0.

Some features are currently only used internally. For example, members of the same project can see each other's sessions by default, and after engineers and AI complete most of the work, product managers can enter the same session to adjust spacing, styles, and copy.

Zhang Tao writes directly at the end that these are the things Manus needs to do well next, "We will try it ourselves first, then hand it over to users".

The phrase "eating our own dog food" has circulated in the software industry for decades. It is generally believed to have originated from an internal email by Microsoft executive Paul Maritz in 1988, titled "Eating our own Dogfood", urging employees to use their own products more. A more famous example is Slack, which was initially a communication tool used by the game company Tiny Speck during the development of the game "Glitch"; when the game did not succeed, the tool became the main business instead.

Manus is taking a different direction. The product is first created, and then the internal usage is documented as the best model room for the product. For a company selling general agents, there is no demonstration more persuasive than "our entire company works this way".

02 From one agent per person to one session per team

The most thought-provoking section of this declaration is "the opportunities we see".

In the past year, personal agents have become standard for large companies. Anthropic has Cowork, Meta's Muse quickly topped the free charts in the US App Store, and Tencent's WorkBuddy has surpassed 20 million monthly visits in the domestic PC office agent sector. "Equipping everyone with an AI that works for them" is becoming something everyone is doing.

The three opportunities listed by Zhang Tao point to another layer.

The first is that sessions themselves will become collaborative work objects, where one person raises a question, designers propose solutions, engineers implement them, and product managers operate, all advancing within the same session.

The second is cross-device and cross-colleague collaboration.

The third is that once agents enter the company's systems and knowledge base, they can proactively find solutions that people did not think of.

All three points discuss organization, focusing on how multiple people and multiple agents can work together.

This aligns with Cue's thinking. Cue assigns each agent an independent email, phone number, wallet, and computer. CEO Xiao Hong stated at the launch that such a "thing" might be called a "person", and next, "should there be a group of many people and bots". Cue brings this idea into personal life, while the digital colleague in the article represents its form within the company.

Several details about the digital colleague are more vivid than any feature introduction.

Many colleagues will seek it out simultaneously, so it only converses with everyone in one "workspace" session, while specific tasks are assigned to other sessions for parallel completion. With many Slack messages, it first determines whether they relate to itself, categorizing those that do into "interrupt what I'm doing" and "handle after finishing". Zhang Tao specifically mentioned that this set of rules was designed by the AI during the development process.

Once, colleagues were discussing two projects in Slack and couldn't remember which stage one of them was at. No one @ed it, but it proactively reminded, "B is still in the testing phase; should we wait for A's results before looking at B?" After the product team launched 3D model generation and skeletal binding, it proactively asked about applicable conditions, clarifying that skeletal binding currently supports seven types of objects, including fish, snakes, and quadrupeds, so that external introductions could be accurate.

Zhang Tao said that at that moment, it felt more like a colleague who was attentively listening to the meeting.

As single agents become standard, the next territory for general agents likely lies in team collaboration. Manus has turned its own company into a testing ground.

03 Per capita expenditure of $10,000

The most controversial figure in the article is the AI expenditure per capita exceeding $10,000 in September. Zhang Tao mentioned this in his social circle at the beginning of the month, attracting considerable skepticism about "spending too much"; this time he directly provided three reasons.

The first is that this money buys time and attention. As the team grows, communication costs rise; AI saves each person's attention, allowing everyone to have a more complete grasp of their responsibilities.

The second is that it is expensive because it uses the best models, which is a core method for understanding the boundaries of AI capabilities. He cited the example of a code review robot. A colleague responsible for development efficiency mentioned that in the past six months, there have been no reports of online failures due to problematic code integration, and after the widespread adoption of AI programming, the company's daily code submission volume is "very exaggerated". This colleague believes this may be the most cost-effective expenditure for the company.

The third reason is the most interesting. More and more development work is being completed on Manus Studio, which is Manus's own product. Every token burned internally will turn into data that clarifies where token waste occurs. As products improve, waste decreases, costs drop, and users benefit as well.

This logic frames a cost as R&D investment. Zhang Tao's judgment is that the working method bought for $10,000 a month today may only require a few hundred dollars in two years.

From another perspective, this calculation also reflects the common situation of application-layer agent companies. The engineer mentioned at the beginning of the article had both Manus and OpenAI's Codex open on their screen. Manus's R&D relies on top-tier models, while the companies providing these models are also developing agent products.

Some comments have labeled this model as "dependent innovation", finding gaps between capabilities and users on platforms built by others, at the cost of paying potential competitors for every business transaction.

In this sense, the per capita expenditure of $10,000 simultaneously proves two things. First, the team has indeed touched the boundaries of model capabilities; second, Manus encountered the cost issue of agents earlier than its users, thus having a better opportunity to resolve it first.

04 Speed and Zhang Tao's unease

The other side of the declaration is the members of the Manus team.

In this article, almost every position has been rewritten. Engineers have shifted from writing code themselves to directing AI work and being responsible for results. Product managers must be able to bring functionalities online, and most of the interfaces users see in Manus Studio, as well as the Video Editor, are designed by product managers. Designers have recently started writing code as well.

Zhang Tao used a phrase to summarize this change, from Human in the loop to Human on the loop. Previously, AI did 70-80% of the work, with humans filling in the gaps; now, humans design the rules, supervise multiple AIs, make judgments at critical points, and intervene when necessary.

The colleague using AI the most in the team has the most "exaggerated" working style. He has a general Manus session, only responsible for listing the matters that require his decisions, along with recommended solutions. While everyone discusses passionately in Slack, he can choose not to open Slack and simply ask AI, "What do I need to decide now?"

These new definitions are, to some extent, also a job advertisement. The article concludes by stating that people are still important, so "we are still hiring". Almost simultaneously, Manus opened 17 positions in Beijing, from Agent Harness engineers to business analysts, with CEO Xiao Hong stating that they received thousands of resumes in a single day. GitLab has publicly shared its employee handbook online, and Netflix released its cultural handbook years ago, which is a similar principle.

Clearly articulating how a company operates is itself a way to filter what kind of people will come.

However, what moved Geek Park the most about this article was Zhang Tao's less certain parts.

He candidly expressed that his feelings are complex, with excitement and unease:

"Is this pace right? Is it fast enough or too fast?"

This question deserves further inquiry. How can quality be guaranteed when one person is handling twenty requirements simultaneously, and how will responsibility be assigned when issues arise? "Haven't heard of online failures for half a year" is a colleague's impression, and it cannot yet be considered a statistic. How will the other half of product managers, who are not yet able to go online independently, keep up? The article does not elaborate, and perhaps the answers are still being explored.

There is also a very personal segment. Zhang Tao wrote about having ADHD, with over twenty tasks daily, and the hardest part is "getting started". Now he hands all his tasks to a dedicated Manus project, which can directly start what it can do, ask him for uncertainties, and then conduct a round of research combining Slack, Gmail, and Notion before handing it over to the corresponding product manager.

He said this has significantly reduced his anxiety about "not knowing where to start" every day.

At the end of the article, Zhang Tao returns to the question of what humans ultimately do. His answer is to define problems, judge quality, decide what is worth doing, and take responsibility for the results.

In an AI Native organization, the role of humans has not diminished; rather, it has become more concentrated.

This declaration has already proven Manus's organizational efficiency. Whether it can become a product that others are willing to pay for will depend on when and in what form those "features currently used only internally" will be delivered to users.

Zhang Tao is curious about what their working methods will look like when he writes on this topic again in six months or a year.

I think that by then, looking back, we will probably know whether this article is a declaration or a preview.

This content is provided for general informational purposes only and doesn't constitute financial, investment, legal, or tax advice. Any events, rewards, online promotions, or related information mentioned herein should not be considered a recommendation, solicitation, or invitation to purchase, sell, trade, or otherwise deal in any crypto assets. Crypto assets are highly volatile and may result in loss. The availability of WEEX services, products, and related events may vary by region. You are responsible for ensuring that your participation is in accordance with applicable local laws and regulations.