Manus' popular red and black: marketing is indeed too much, but the product is amazing

To be honest, there is no invitation code for manus so far, so if you insist that you have no say without experience, just crossing it can save you ten minutes.

I want to talk about two major points, one is about the marketing controversy about manus; the other is about manus's product tricks.

Product: No breakthrough, but breakthrough

Manus has not made a breakthrough in technology, which is perhaps the biggest consensus after the controversy. The most core verification case comes from the MetaGPT team (who does programming agent team) replicated an OpenManus in three hours.

But there is no doubt that the products manus brings to everyone are shocking again. It regards AI as the "human hand" and automates a series of processes, including a large amount of independent information collection, browser interaction and other content. These will eventually be encapsulated, and the user just needs to tell manus what he wants to do and can move a chair and watch the show directly.

If you pay attention to some dynamics in the AI ​​field like me, you will find that manus is more like a stitching monster compared to deepseek model training innovation:

1. Task sorting and knowledge base call. Most models have this capability, and from the perspective of propt engineering, disassembly and sort out tasks first will improve the final output of AI. In my daily interactions with AI, I either make a list directly or let AI sort it out for me based on my needs.

2. Information collection, sorting and analysis. That is, Deep Research, which is now basically supported one after another.

3. External tool function call. Whether it is MCP or a large number of open source tools such as browser-use, they have been incorporated into daily use scenarios.

4. Multi-agent cooperation. It should have been a year or two since Devin (I don't remember the exact time). What the metaGPT team that was replicated in the three-hour period is multi-agent cooperation in the field of programming.

Many people are scrupulous about the fact that manus has not been innovative, and start to sarcasticize when they see praise.

It is important to know that arrogance is the biggest enemy of progress.

Why do you want to ask yourself: Since they are all existing things, why does it explode when they are integrated? They are all things that are already there, why don’t you sew them?

Everyone’s perspective on innovation is still too narrow, and they always stick to technology, but they don’t know that innovation in product concepts and business models are more far-reaching.

I put an article at the end and wrote very well about the understanding of manus products (because it is a retelling team itself):

This passage is also the main reason why I wrote this article. I think the points mentioned here are worth savoring by everyone who makes products.

Let me talk about what I saw from it and what manus is worth learning:

1. Anti-interference of the intelligent workflow

A similar concept is also mentioned when making games, "flow".

That is, the user's experience of using the product is composed of behavioral processes. For example, if I use Alipay to pay the phone bill, I need to open Alipay, click "Recharge" on the homepage, select the mobile number, select the amount, confirm, pay, and receive feedback on the recharge successfully. In such a process, if there is interference from an extra-process factor, it will interrupt the overall process and cause a decline in product experience.

Although the existing AI agents are not completely excluded from user workflows as mentioned above, they do have problems of occupying workspace. Taking the AI ​​that you may commonly use to read web pages as an example, AI provides an understanding based on the current web page. You cannot change the current web page, otherwise it will be invalid. Otherwise, copy the link manually and give it to the AI ​​in an independent window. Another product I have tried recently is the same.dev, which focuses on directly copying the front-end source code, and there is a problem of interrupting the operation if you bypass the web page, occupying a working space alive, and the experience is very poor.

Of course, it is not about which one is better or worse. But the changes at the product level of manus are meaningful for us to think about the evolution of AI agent form.

From browser embedded AI to AI interactive page embedded browser, the former meets users' daily usage needs (AI is only part of the work and life process), and the latter meets the needs of intelligent agents (reducing irrelevant interference).

AI Agent products, exposed to users, only need to be an interactive page that receives input, displays the process, and emits output. The process is only allowed to be displayed and the user will not be disturbed by the process.

I remembered the problem with the browser embedded in OKX wallet in the previous article. You cannot let users actively interrupt the process to achieve their goals.

The experience of interruption of "flow" is terrible.

2. Rethink "external tools"

In the eyes of most people, only standard protocols such as MCP or packaged libraries may be considered "external tools".

In fact, the so-called tool is the stability of input and output, which can give users a black box with clear expectations.

Programming is the most certain one in AI application scenarios, countless scripts and modules can be packaged as tools.

It is no problem to compete and solve math problems from the perspective of model training; but if you directly follow math problems from the perspective of model application, it is stupid.

Why don’t you write code to solve it? You have to use a vector mapping?

manus actually pointed out the key point in the design of a general-purpose agent: don't try to directly solve all problems with AI.

AI is just a hand.

Use tools to solve the problem. Tools can be defined in advance or written temporarily.

People - Hands - Tools - Tasks.

Do you care about how many layers of tools in the middle?

3. Non-destructive use scenarios

Students who often program must have a deep understanding of the different projects, the dependency library version is messy, and as long as one of the key library versions is not matched, the project may run and report an error.

This is also the necessity of virtual environments in programming. npm is based on project dependency installation, python creates virtual environments, docker containers, etc. In my understanding, these are all to ensure the independent customization of the environment.

This may be the consensus of all product-level agents directly toC at this stage: do not invade the user's local environment and use the cloud.

There are bolt.new, mgx.dev, etc., and they all choose to run directly in the cloud, write and debug. However, these are all programming agents, and general agents lack comparison.

In sharp contrast to manus's approach, there is another product, Highlight. After you download and install, a floating window will be formed on the desktop, and there are some AI-combined operations based on the application of the current workspace on the floating window.

It seems to make people shine?

For example, if I don’t know how to crawl, I will adjust to the browser page and let Highlight crawl for me?

I personally experienced that the practice of highlight has interfered with the original workflow because I had to open the process and let highlight operate. But in fact, anyone will switch pages back and forth at work, and it is impossible to wait for the AI ​​operation to be completed specifically. In addition, if the AI ​​uses my browser to crawl, will it affect my future access?

The local scene was destroyed.

These points actually seem to be a feeling that is not difficult to imagine. But I think it is still worth experiencing how to conduct a series of designs in a framework.

Finally, let me say something about the manus ceiling, my personal expectations.

My expectations are to control expectations – AI can’t do everything for us. Even if smart homes introduce AI modules in the future, they become "tools" that can be used by the "human hand"; even if more and more "tools" appear on the desktop can allow us to control production-level software by only natural language; many times, humans must also conduct process control and tests, because AI's understanding of the world is based on a huge black box, and they will have "illusions".

In any case, since we believe that human reality is greater than AI’s reality, human beings need to test the production of AI. And the more you hand it over to AI, the more content you need to test. This will eventually balance above a boundary.

This boundary is the upper limit of general-purpose Agent products.

Marketing: Don't be afraid of controversy, afraid of not being

controversy

Looking at this wave of manus's out-of-the-box is not so much a spontaneous controversy in the circle, but rather a pair of "hands of God" are guiding them in secret.

These hands of God must have come from the official. Wouldn't it be great to lead your hands with your hands?

First, I asked Grok to summarize the major marketing events of manus during this period:

It can be clearly seen that the official core slogan is "the world's first universal AI Agent".

This is a very controversial statement.

A. For those outside the circle, this statement is very eye-catching;

B. For my daily attention but not in the circle, you can see at a glance that I was playing a text trick: I have introduced it above the product level, it is a stitching monster, which is not the first one in any case; but if the word "daily" is added, there is indeed no such consumer-grade product that says it wants to handle general-purpose tasks, which has caused a lot of public opinion. In addition, the slogan is an exaggeration tendency, so it is indeed understandable to say this;

C. As for people in the circle, I think most of them may be more angry. After all, the results of my research were sewn in, or the fact that this thing in my eyes had no technical content, but he stole the limelight.

——Confrontation of positions emerged.

If there is a stance, there will be debate, and public opinion will eventually continue to ferment, bringing a leverage effect to manus' propaganda.

Don't look at the appearance, look at the results.

The result is that manus has gained global attention. This marketing is cost-effective.

And the invitation code mechanism.

While the attention is full, the invitation code is strictly limited, which is based on cost considerations on the one hand, and it also covers up product shortcomings on the other hand. After all, the "general type" is boasting about Haikou. Once it is opened, it will be immediately blown up by various bugs and feedback from the A group of people. That will not be an opposition, but one-sidedness, and it will probably be a mess. From this perspective, the invitation code mechanism is actually similar to an early test. You can put as many codes as you fix the dev bugs, so that seed users can help improve the product first.

There is also the plan for hunger marketing. Hunger marketing is essentially a competition for attention. All those who use invitation codes will inevitably be scolded by those who "love but cannot be given", which is normal.

What happened later on was the X account freezing storm, technical jailbreak, and open source, which were regarded by many as the consequence of excessive marketing.

Regarding this, my personal opinion is: the AI ​​circle is still too literary. It is recommended to take a walk in the crypto circle and learn some shameless spirit.

The manus team is actually fine. After all, the previous product, Monica, has made steady profits, but many small teams are still struggling on the edge of food and clothing. At this time, do you tell me not to over-market? If high marketing results can be achieved at low cost, I would like to ask why not?

Can face support the R&D team to continue doing it? Can face provide enough funds for innovation?

Face is not worth a lot. In this era of entertainment until death and information explosion, attention is valuable.

The cruelest and most training part of the crypto circle is that it is too close to the money itself, so that all you can see are the most real human nature and the bloodiest routine (not the bloody routine, but you are bloody).

You must have principles in life, but you must not misunderstand and learn to accept some reality that is not the principles of yours.

Otherwise, you will die miserably.

References

Complete review: How was Manus born? | Geek Park

Behind Manus’s popularity, how can Agentic AI products build a lasting competitive advantage?