It's becoming a craze for the whole nation to raise shrimps with OpenClaw. Why is shrimp farming not friendly to ordinary people?

Recently, if you were to ask what's the hottest thing, the nationwide craze for "shrimp farming" (i.e., the deployment fever of OpenClaw) would undoubtedly be one of the top contenders. For a while, almost all social media platforms and friend circles were abuzz with discussions about shrimp - farming experiences, which clearly shows the popularity of OpenClaw. However, in the midst of this nationwide shrimp - farming craze, why do we say that shrimp farming is not friendly to ordinary people?

1. Has Nationwide Shrimp Farming Become a Craze?

According to a report by Wall Street News, in early March, outside the Tencent headquarters in Shenzhen, Tencent engineers set up stalls in the north square of the building like they were at a fair, offering free installations of the "lobster" OpenClaw to users. The line stretched on and on. Some people were holding NAS devices, some were carrying MacBooks, and others were lugging around mini - PCs. It was very much like a gathering of geeks who used to flash Android systems a decade ago.

In fact, many large tech companies have been actively promoting their own "lobsters". Xiaomi has started the internal testing of MiclawAgent, aiming to embed AI agents into Xiaomi's "full ecosystem of smart phones, smart cars, and smart homes" so that mobile phones, cars, TVs, and home appliances can all become AI execution nodes. Even earlier, just before the Spring Festival, Baidu App supported one - click access to OpenClaw. After users completed the deployment on Baidu Smart Cloud, they could access it through the search box or the message center. Cloud providers have started "setting up stalls", and when terminal device manufacturers start integrating agents into operating systems, this "lobster" storm has kicked off the second half of the large - model era.

On GitHub, OpenClaw has exceeded 250K+ stars, becoming the software project with the most stars on GitHub. On social media, news about starting a "one - person company" with OpenClaw and making huge profits daily with it is constantly being shared, as if the wealth code for 2026 is hidden here. Different from the chatbot represented by Doubao, OpenClaw is an open - source, free AI agent framework that can run locally. Its core lies in "letting AI actually do the work".

2. Why Is "Lobster Farming" Not Friendly to Ordinary People?

In recent years, "nationwide shrimp farming" has become a phenomenon with a touch of magical realism in the tech circle and even in public opinion. However, behind this craze, we must calmly realize that although shrimp farming is popular, it is far from being a widespread technological dividend for ordinary people. Instead, it is more like a "technological adventure" with high thresholds, high costs, and high risks.

First of all, the AI agent direction represented by OpenClaw is indeed of epoch - making significance. Undoubtedly, the emergence of OpenClaw marks a new stage in large - model technology. In the past, no matter how intelligent large models were, they were essentially "answerers". You ask a question, and they give an answer; you push them, and they move. OpenClaw, on the other hand, tries to endow AI with "initiative": it can understand task goals, break down steps, call on tools, and even correct itself after failure. This leap from "dialogue intelligence" to "action intelligence" is indeed the inevitable direction of technological evolution and also indicates that AI will be deeply integrated into human work and life processes in the future. From the perspective of industrial economics, this is not only an upgrade of efficiency tools but also a reconstruction of production relations. When AI can independently complete repetitive tasks such as data sorting, email writing, schedule arrangement, and code debugging, the boundaries of human productivity will be greatly expanded.

However, the ideal is beautiful, but the reality is harsh. Currently, OpenClaw is still in the early experimental stage, and its design logic is highly skewed towards developers and tech geeks rather than end - users. Its core value lies in "programmability" and "composability", but this precisely means that ordinary users must have a certain level of technical proficiency to use it. For the vast majority of ordinary people who are not familiar with the command line, do not understand API calls, and have no concept of the operating system's underlying principles, OpenClaw is not an assistant but an insurmountable technological barrier.

Secondly, the threshold for deploying and using OpenClaw is extremely high, far beyond the technical capabilities of ordinary people. After I spent a lot of effort deploying and getting started with OpenClaw, several obvious problems made me realize that although shrimp farming seems appealing, it is not friendly to ordinary people. We must clearly recognize that there is a huge gap between the ideal and the reality. When ordinary users eagerly want to have a dedicated AI butler, reality often gives them a heavy blow.

One is that the deployment is very complex, and the technical threshold keeps the vast majority of ordinary people out. For most non - tech - savvy people, the current "shrimp - farming" environment is far from being user - friendly. If you can't code or even install an operating system, don't bother with "lobster farming" using OpenClaw. Different from simply downloading an app and using it immediately, the deployment of OpenClaw involves a series of complex computer operations such as environment configuration, dependency library installation, and API interface debugging. The debugging requirements during the deployment process are extremely high. For someone without a technical background, even a single line of error code might as well be a foreign language. This technical barrier directly makes "shrimp farming" a celebration for geeks and programmers, while ordinary people can only look on enviously from the outside.

Two is that although OpenClaw seems great, it is not a commercial product, and the user experience is extremely "hardcore". Many users mistakenly think that once installed, they can assign tasks to it like hiring a cheap laborer. This is a complete misunderstanding. Currently, OpenClaw is more of a prototype of an open - source project rather than a mature commercial software. Due to the lack of sufficient user - experience optimization, the perceived user experience is not good. During the task - execution process, the AI agent may get stuck in an infinite loop, misunderstand instructions, or even get stuck because it can't call a certain tool. Users need to spend a lot of time "training" it and even need to correct its mistakes like managing an inexperienced intern. This extremely high human - maintenance cost is far from the easy "hands - off" scenario people imagine.

Three is that the cost of using OpenClaw is really high. It's a continuous "money - burning" game. This is the hidden cost that is most easily overlooked. Although there is a lot of online publicity about OpenClaw being open - source and free, this is just the surface. The code framework is open - source, but the "fuel" that supports AI thinking, Tokens, is not free. Let's not even mention the expensive Token - charging standards of OpenAI. Even the Token consumption of domestic models like DeepSeek, Kimi, MiniMax, and GLM is astronomical.

The working principle of the AI agent makes it a "chatty" one. It needs a large number of internal self - dialogues, inferences, and reflections to complete a task. I only tested it for two hours at noon and gave it a few very simple tasks, but it cost me dozens of yuan. It's like driving a sports car with extremely high fuel consumption. Although the car may be given away for free, the fuel cost can bankrupt you. For ordinary individual users, spending hundreds or thousands of yuan a month just to complete some basic tasks that could be done manually is obviously a very poor return on investment, which is enough to make ordinary people back off. Moreover, in addition to the Token cost, in order to enable OpenClaw to better complete various tasks, users also need to purchase a large number of paid applications for it to call. These additional costs, combined, are undoubtedly a heavy burden on the economic situation of ordinary people. It's like raising a pet. You not only have to buy the pet itself but also various pet supplies, food, and pay for medical expenses. For some people with limited financial resources, they simply can't afford it.

Thirdly, why is OpenClaw so popular? We must rationally view the popularity of OpenClaw. Behind the craze is not simply the spread of technology but the result of multiple factors working together. This also makes many ordinary people ignore the underlying thresholds in their blind following. The reason why OpenClaw has triggered a nationwide "shrimp - farming" craze in a short period is mainly driven by three aspects, rather than it having reached a mature stage suitable for ordinary people.

One is the guidance of technological trends. OpenClaw precisely hits people's core needs for an "AI assistant", and its future potential has attracted a large number of people to follow suit. In today's era of rapid development of artificial intelligence, people's expectations of AI have long gone beyond simple chatting and content generation. Instead, they hope that AI can truly become a "helper", taking over tedious and repetitive tasks, liberating hands, and improving efficiency. The AI agents represented by OpenClaw exactly meet this need. It makes people see the possibility of "AI working for humans". This beautiful vision of the future makes many people willing to give it a try. Even if they don't have the relevant technical skills, they hope to keep up with the technological trend through "shrimp farming" to avoid being left behind by the times. This pursuit of technological trends is the core driving force behind the popularity of OpenClaw.

Two is the promotion by large - model companies. OpenClaw truly offers the possibility of profit for large - model companies, which represents a qualitative change in the business model. This is the core logic from a financial perspective. In the past, people mostly used large models for free chatting. This state where it was almost impossible to find an effective business model made large - model companies continuous losers. However, the operating mechanism of AI agents requires high - frequency API calls, and these calls are not free; in fact, they are quite expensive. Once widely applied, large - model companies can shift from "selling models" to "selling services" and "selling traffic". OpenClaw has become the perfect tool for large - model monetization. Once this business model is proven successful, the promoters behind it spare no expense in marketing and promotion.

A piece of data can illustrate the point well. According to a query by "China Entrepreneur" on the global leading AI model API aggregation platform OpenRouter: throughout February, with a weekly - based statistic, "Crescent Moon" and MiniMax took turns "topping the list". From February 9th to February 22nd, MiniMax's M2.5, "Crescent Moon"'s Kimi K2.5, Zhipu's GLM - 5, and DeepSeek's V3.2 occupied four of the top five positions. From February 9th to February 28th, the call volume of MiniMax M2.5 ranked first for three consecutive weeks. A European development studio that has deployed OpenClaw said online that 80% of its daily inference engine work is completed by Kimi K2.5, and only 20% of complex tasks are assigned to Claude. The daily cost of using K2.5 is about $5 to $10, and if all tasks were assigned to Claude, the monthly cost would be between $800 and $1500.

Three is the "water - seller effect". A group of people have taken advantage of the "shrimp - farming" craze to make profits by providing services such as installation and teaching, further amplifying the influence of the craze. Historically, in every popular industry craze, the first ones to make money are often not the "gold - diggers" but the "water - sellers" who provide services to them. The "shrimp - farming" craze of OpenClaw is no exception. As the nationwide enthusiasm for "shrimp farming" has risen, many people have seen business opportunities. To address the deployment difficulties of ordinary people, they offer on - site installation and remote debugging services; to help non - tech - savvy people use it, they offer paid tutorials and one - on - one teaching services. These services have relatively low thresholds but can charge high fees, and the demand is strong, allowing many people to make quick profits. The active participation of these "water - sellers", through various forms of publicity and services, has further amplified the "shrimp - farming" craze and made more ordinary people mistakenly think that "shrimp farming" is easy and profitable, thus blindly following the trend.

Fourthly, is "shrimp farming" with OpenClaw meaningful? The answer is definitely yes. Criticizing the craze does not mean denying the technology. In specific scenarios and for specific groups of people, OpenClaw has shown amazing potential.

If you are a professional technical person and have a large amount of repetitive work in your daily job, then OpenClaw is undoubtedly a very effective helper. For this group of people, the technical threshold does not exist, and the debugging process can even be a source of fun. Moreover, compared with their high labor costs, the Token cost can be ignored. They can use OpenClaw to automatically process data, write scripts, and test programs, achieving exponential efficiency improvements.

If you regularly record your work comprehensively and even create a large number of work documents for training, then you are also a good fit. OpenClaw will have many opportunities to learn your thinking logic and workflow. For expert - level users with a structured knowledge base, the AI agent can quickly replicate their work patterns and become a real digital clone. This group of people has the ability and resources to "tame" AI and are the real beneficiaries.

However, if you are an ordinary non - tech - savvy person, OpenClaw is not friendly to you at least for now. You not only need to spend a large amount of money to buy Tokens but also need to purchase a large number of paid applications for OpenClaw to call (such as API interface fees and professional software licensing fees). Such costs are beyond the affordability of an ordinary person. More importantly, you also need to invest a huge amount of time to learn how to communicate with the machine and deal with various unexpected situations it brings.

It can be said that the nationwide "shrimp - farming" craze is essentially driven by people's urgent longing for an intelligent future and commercial interests under the wave of technological change. However, technological progress never happens overnight, and it will not benefit all people equally. The AI agents represented by OpenClaw are undoubtedly the future direction, but currently, it still has a distinct elite attribute. The dual barriers of technical and cost thresholds make it difficult for ordinary people to truly participate. Therefore, although "shrimp farming" is popular, it is not suitable for everyone.

This article is from the WeChat official account "Jianghan's Vision and Observation" (ID: jianghanview), written by Jianghan's Vision and Observation, and is