Today I saw the installation of crayfish in Tencent Building, which gave me some insights. Recently, many large factories have been crazily asking non-technical employees on the front line to install crayfish, and there are even 500 on-site installation services available online. Everyone is desperately searching for usage scenarios, demanding implementation, and proving that this thing is so important that it cannot be missed. The whole process gives me a strong sense of the folding of cyber technology. It's interesting to see that someone who can't even pretend to be a lobster, how could they use a lobster. Going one step further, without even establishing basic usage, it is even more difficult to first create a complete scenario, produce results, and demonstrate value. There are two things stacked together behind this. One is illusion. Many bosses have watched too many video clips and have been repeatedly bombarded with exaggerated narratives and omnipotent cases, which can really create an illusion that this thing can do anything, accept anywhere, and anyone should install it. Once installed, there should be immediate output. The other is anxiety, and everyone is afraid of missing out on this wave, so they start using administrative actions to promote it, using collective anxiety to replace real needs. So you will see a strong contrast. The slogan on one side is very loud, as if everyone is about to enter the era of AI native. On the other hand, a large number of people cannot even articulate what they truly have to entrust to it. This contrast will only become stronger and more absurd in the future. Because tools never generate value through installation, they only generate value through task density, unclear processes, and visible results. There are no continuous tasks, no SOPs, no conditions for online completion, no clear input and output, and even the strongest thing is just an icon. It won't automatically grow scenes just because it's installed. So I have always believed that lobster is not suitable for everyone. It is very suitable for commanders, very suitable for one person companies, and also very suitable for people who always have things to do in their minds, can break down work into steps, and can complete many things online. Especially if you have used skills and tools, and know the boundaries of AI's capabilities, being able to string processes together, set up scenarios, and complete tasks step by step, this kind of situation will be very suitable. For example, to me, this scene is very natural. Especially when there is a lot of work to be done, but I happen to be away from home or at work, carrying my phone outside, or inconvenient to use the computer, I will ask my two nanobots to check my open source product issues, generate technical solutions, and then the other one to review and submit, all in one go. It's really convenient for me to do things elegantly on the way to work in the morning. But for someone who doesn't usually have much work to do outside, or even someone who doesn't even want to turn on their computer when they get home, how can they force themselves to do things. Eating fun makes it very comfortable. No scene means no scene, there's really no need to worry. I think the easiest thing to magnify this wave is not the ability gap, but the scene gap. People with scenes tend to use them more smoothly, run faster, and eventually have a few more clones. People without scenes can easily wander back and forth through concepts, tutorials, cases, and videos, and in the end, nothing changes except for installing a few more software. The biggest problem for many people today is not that they haven't installed lobster, but that they have installed a certain tool as if they have entered the AI era. In fact, the real watershed has always been in task understanding, process design, and result judgment. Do you have any ongoing problems to solve, can you break them down and hand them over to the system, and can you judge whether the results are correct? These are the factors that determine whether you can truly gain value from AI. So there's no need for anxiety. When there is no scene, there is little point in pretending to be a lobster. If you really want to experience where this generation of AI is strong, why not spend $20 to package a Claude Code, or even more interestingly, package a ChatGPT member, use GPT 5.4 to help you handle something that you really find difficult, generate solutions, promote execution, and experience this simple, efficient, and direct problem-solving process? It's much better than just putting on a lobster. Lobster is suitable for people with scenarios, suitable for commanders, suitable for one person companies, and suitable for those who can standardize, streamline, and complete processes step by step. It is certainly strong, but it does not prove itself strong by being installed, it proves itself by completing the work for you. Many people are pretending to be crayfish today, but what really needs to be understood first is one sentence: What problem do I have that deserves to be solved by AI. This matter may be more important than pretending to be anything.