A friend gave me a "lobster".
In late February, when OpenClaw began to break out of the circle, I was still standing on the shore. On March 1, the first Sunday after the holiday, I made up my mind to spare an afternoon, prepare strategies, and prepare to go into the water to raise shrimps myself. As a result, WeChat rang first.
"How about I give you a lobster?"
OpenClaw is open source and free. Where can I get it? I later found out that the service was for installation. His technical partner remotely connected to my Mac mini, which was commandeered as a lobster plane, and struggled for an hour and a half. After installing it, the person asked me to give it a name. I was hungry at the time and blurted out: "Steamed."
"Steamed" has settled in my home.
Later I learned that people who move quickly are already making money from this. This remote installation service is clearly priced at 199 yuan at the seafood market. A friend gave it to me as a favor. When installing, you need to select the model source. The default is to go to the installer's transfer station – of course you can change the settings, but most lazy people like me will most likely keep the default. This is the first steady stream of income.
After pretending, he was pulled into the group and entered the private domain. In addition to exchanging shrimp farming experiences, the group also posts cases from time to time. I've used the "steamed" version a few times, and it's indeed different, but I was busy with work and didn't bother to look into it. However, within a week or two, the screen was full of lobster work notes: some people used it to automatically process Excel reports, some people used it to crawl data and write emails, some people created automated workflows, and some people used it to batch generate short video scripts. People use lobster to fly, but I'm still stuck in the "steamed, push me a briefing at eight o'clock every morning" stage.
This is a great time to launch your course.
At this point, the business model of "one shrimp and more" has been formed: you can earn one time by selling installations, you can earn turnover by transferring, you can earn added value by selling courses, and the community can be monetized again.
I give my friend a thumbs up: he is indeed a veteran in the community and a model of win-win. And what about their technical team? In addition to the full-time founder, the other main engineering staff are part-time programmers – working at Internet companies during the day and working remotely for people at night.
This reminds me of those ghost stories about AI causing programmers to lose their jobs. Programmers may indeed be laid off, but with a flexible mind, they can make money wherever they go.
During this period, OpenClaw continued to explode. I ate several meals, and every time someone asked: Have you raised shrimps? According to the current popular saying on the Internet, a large number of small, medium and old enterprises have begun to ride on the popularity of lobster. According to communication theory, this is the diffusion stage—the stage of spreading from early adopters to the early majority.
Then the wind changed direction.
Security issues exploded. Some people compare it to the panda burning incense, directly changing the household registration date to a virus. Finally, at another party, a friend came over and asked me: "How do you remove the crayfish? Make sure not to leave any shrimp skin and shrimp lines, and delete them completely."
At this point, it's only been two weeks since I installed it.
From raising shrimp to unloading shrimp, public opinion only took 14 days, but I believe that there will be a wave of "crayfish upgraded to king shrimp" in the future. In fact, this is the normal situation when new technologies emerge – technology moves too fast, people's hearts catch up too quickly, and the two sides collide. When the Internet entered China, wasn't it first touted as an information superhighway and then criticized as a spiritual opium?
OpenClaw will not be the end of AI, I’m sure of that. But the significance of its popularity is that it allows ordinary people to experience the possibilities of AI beyond just asking and answering at the low cost of a hot pot meal. It starts operating your computer, clicking the mouse, typing the keyboard, and filling out forms for you. It is indeed one step closer to Jarvis in "Iron Man".
Of course, it has a lot of problems: the threshold is still high, ordinary people don’t understand it, and professional services are likely to be needed; security is questionable, and who can guarantee that it will not mess up; the economic accounts have not been settled, and the accumulation of API call fees is not a small amount. But this is normal. Technology always precedes supervision and ordinary people’s understanding.
This is not the unique fate of OpenClaw, it is the inevitable path for every underlying technology iteration. From electricity to the Internet, when was it not first treated as a toy, then scolded as a scourge, and finally quietly grown into infrastructure?
Interestingly, after this public outburst, I saw more people discussing in the circle of friends and in the community: how to make good use of it, how to tame it, and how to make it really help? Fear is a form of attention. After the scolding, you still need to learn what you need to learn.
I looked at the people in the group who were flying with lobsters, and then thought about those people who were in a hurry to unload the shrimps, and suddenly I realized: the same tool means two completely different worlds to different people. Does it give more people opportunities, or does it make people who already understand run faster? "
I think: it is magnifying the gap, not flattening it.
Also using AI, someone asked, "Write a plan for me." Someone raised 8 shrimps, each performing its own duties, and started collecting money. The former gets mediocre templates, the latter gets commercial value. There is a huge difference in output between people who can ask good questions, understand the deep needs of the scene, and dismantle efficient workflows, and those who are ignorant and lack of thinking.
This gap existed before, but AI has magnified it exponentially. In the past, the difference in output between an excellent copywriter and a mediocre copywriter could be twice or twice as much. Now, the gap is geometric.
Here I would like to propose a concept: cognitive arbitrage.
AI is essentially a cognitive lever, but it is not an evenly distributed lever. People who know how to use it can use it to leverage more cognitive resources, and these cognitive resources can in turn be used to better use AI – this is a positive cycle. People who don't know how to use it can't even lift the first brick. Therefore, those who understand understand better, and those who do not understand become more confused.
In recent years, more and more people have worried about being replaced by AI, and I have also worried about it. But now, I feel like maybe the question shouldn’t be asked that way. The question to ask is: If AI can do more and more things, what should humans do?
I think people should do things that AI cannot do.
AI can write a plan, but it doesn't know why it writes this plan. AI can process data, but it does not understand that there are real people behind the data. AI can optimize processes, but it doesn’t know what value this process serves. These things ultimately need to be defined by people.
As long as the value standards of this world are still defined by people, and as long as people are still the center of consensus, then people's position will still exist. It's just that this position is changing – from executor to definer, from operator to questioner, from producer to judge.
Some people say this is idealism. I say this is realism: you can only do things that AI cannot do, because what it can do, it is already doing.
OpenClaw is called a virus by some people, but a virus is also a form of life. It is parasitic on the existing system and uses the system's resources to complete its own replication. The same is true for AI. It is parasitic in our workflow and uses our data, computing power, and needs to complete its own evolution. This process will bring discomfort, panic, and new opportunities.
The key is, are you the one who is parasitized, or the one who learns to live with it?
As for regulation, it will definitely come and it will definitely lag behind. This is the normal state of technological development. Lagging supervision is not a problem. The question is whether the direction of supervision is reasonable and whether it can protect public interests without leaving too many dead ends for innovation. I believe there will be a solution.
Looking back, it has only been half a month since I installed the "steamed" one. In 15 days, I watched with my own eyes as a thing transformed from a geek toy to a popular topic, from a technical experiment to a business model, from free software to the center of controversy. The network effect brought about by social media has greatly accelerated the pace of the impact of new technologies. Such a fast pace has never been seen before, and it may be even faster in the future.
AI will not stay in the chat window. Today it operates your computer, tomorrow it will operate the smart devices in your home, and the day after tomorrow it may take over your company's backend system. It is not an App, it is the underlying facility and a new operating system. What we are experiencing is not a product iteration, but a bottom-level switch.
My goal this month is: try to raise a few more shrimps and build a real division of labor and collaboration system. From individual combat to small team collaboration. From experiencer to builder.
After all, it’s not enough to just eat shrimps. You have to learn how to raise shrimps, use shrimps, and make shrimps work.
"Steamed" learned last week to arrange study tasks for me on time every day, give me grades, and encourage me to rise up in middle age. It's done well, I'm going to add a drumstick to it. Of course, the chicken legs are also virtual.
(The author Wang Chengwei is a worker in the technology industry)
