Automation
AI + Automation
Bank
Insurance
Logistics
Hello, I am Xiaobian.
Recently, a "lobster" brushed the screen of the circle of friends. It's not the kind on the table. It's openclay, an open source AI tool with an icon similar to lobster, which has been popular all the way from the technology circle to the public. Shenzhen citizens even lined up to participate in free installation activities in order to "raise" their own AI "lobster".
The upsurge made Xiaobian itch: what is this thing? What does it have to do with the RPA we are familiar with? Will RPA be replaced?
With these questions, Xiaobian consulted with the company's technical predecessors and supplemented a wave of information. Today, I'll talk to you about the things behind this "lobster".
Why is it called "raising lobster"?
1、 What is openclay? Why is it so hot?
First of all, the conclusion is that openclay is not a new big model, but an execution engine that can let the big model "work".
Traditional AI (such as chatgpt) is more like a "consultant" - you ask questions, it gives you advice, but you have to do the specific operation yourself. And openclay is very like an "employee" - you give instructions and it helps you directly.
For example, you can say to it, "help me sort out my desktop files, package and email the project documents of last month to manager Zhang." openclay will disassemble the tasks and plan the steps by itself, and then really operate your computer, open folders, and call the email system... You don't need to do it all.
Why is it called "raising lobster"?
Because openclay supports local deployment, every interaction will be recorded to form a lasting memory. Over time, it will "understand" you more and more. This process is like raising a pet, so it is affectionately called "raising lobster" by netizens.
2、 Will RPA be replaced by openclay? The answer is clear
At this point, a friend may ask: will RPA, which has been the "main force" of enterprise automation in the past two years, be replaced by openclay, which has suddenly become popular? After asking the technical predecessors, the answer is: no!
Those who say "openclay replaces RPA" essentially confuse the core difference between "general lightweight implementation" and "enterprise level professional implementation".
1. different execution scenarios:
Openclay adapts to personal terminal and fragmented basic operations, such as sorting out personal files and calling general software; RPA is tailored for enterprise level, large-scale, end-to-end core business processes - financial reconciliation, cross system posting, invoice authentication. These scenarios involve multi system linkage and multi rule verification, which is the core link of enterprise operation and the "main battle field" of RPA for many years.
2. different executive abilities:
Openclay focuses on flexibility and lightweight, but compromises stability and controllability in order to meet personal needs; The core advantage of RPA is precisely the extreme stability, strict controllability and auditability of the whole process - each step of implementation is traceable and traceable, which can ensure the zero error implementation of the core process.
3. different system values:
We have always said, "the stronger the agent is, the more important automation is." even though agents such as openclay have basic execution capabilities, their core values are still "thinking" - understanding natural language instructions, disassembling complex tasks, and planning execution paths; The core requirement of enterprise automation is to "do it reliably" - implement the decision-making of the agent into the complex business system, and realize the large-scale and controllable value transformation. This is the strength of RPA.
Gartner predicts that at least 15% of daily decisions will be made by AI agents in the next four years. But decision-making is only the beginning, and implementation is the end.
Therefore, the real trend is not "who replaces who", but symbiotic evolution in technology. In enterprise applications, openclay, as an agent framework engine, and RPA play their respective advantages in depth, and jointly become the "hands and feet" of enterprises' efficient operation - the light personal needs are handed over to openclay, and the core enterprise needs are handed over to RPA, so that the value of agent automation can be truly realized in the enterprise scene.
After talking about technology, Xiaobian also wants to remind you that although openclay is popular, it should follow the trend cautiously and keep rational and sober.
1. risk 1: the greater the authority, the greater the risk
2. risk 2: the cost is not low and the threshold is not low
Liuxingliang, President of DCCI Internet Research Institute, made a very pertinent suggestion: except for developers and AI entrepreneurs, ordinary people should not blindly "raise lobsters". Because an AI who can only talk is "nonsense" at most, but a "lobster" who can only actively perform tasks is likely to cause disaster.
4、 Write at the end: rational view of technology iteration
As ordinary people, we don't have to blindly pursue every technology hotspot, but we can keep an open mind to understand and learn. When we really need it, we at least know: Oh, what can this thing do, what risks it has, and whether it is suitable for me.
Openclay or other AI tools are just tools. Tools are not good or bad. The key depends on the person who uses them. Let's look at technology rationally, embrace the changes of the times, and look forward to the next high level of intelligent automation.
3、 "Raising lobster" has risks, and rationality can go further
After talking about technology, Xiaobian also wants to remind you that although openclay is popular, it should follow the trend cautiously and keep rational and sober.
1. risk 1: the greater the authority, the greater the risk
In order for AI to "do things", it must be given operating system permissions. The stronger the ability, the greater the risk. The network security threat and vulnerability information sharing platform of the Ministry of industry and information technology has issued an early warning, pointing out that openclay has potential security risks under improper configuration.
Some users have encountered that when "lobster" cleaned their mailbox, they ignored the instruction of "no operation without approval", resulting in all important mail being emptied. There are also attackers who use the misconfigured openclay to spread malicious software.
2. risk 2: the cost is not low and the threshold is not low
"Lobster farming" looks cool, but the cost is not small - some users reported that it costs 30000 yuan per month (mainly on tokens). Deployment and maintenance also need a certain technical foundation, and the risk of ordinary users directly accessing "main equipment" is high.
Liuxingliang, President of DCCI Internet Research Institute, made a very pertinent suggestion: except for developers and AI entrepreneurs, ordinary people should not blindly "raise lobsters". Because an AI who can only talk is "nonsense" at most, but a "lobster" who can only actively perform tasks is likely to cause disaster.
Interestingly, even the "government lobster" launched in Futian District, Shenzhen, is equipped with a civil servant on the staff as the "Guardian" to ensure that they do not talk disorderly and do nothing wrong.
4、 Write at the end: rational view of technology iteration
Writing here, Xiaobian can't help feeling that from chatgpt to deepseek to today's openclay, the development speed of AI technology is really beyond imagination. In a few years, we have witnessed the leap of AI from "being able to chat" to "being able to work". This is not only technological progress, but also a leap in the paradigm of productivity.
"Instead of worrying about being replaced, it is better to think about how to control it." AI will indeed replace some jobs, but it will replace "low-value repetitive labor" and release "high-value creativity". Just like the steam engine did not eliminate human beings, but liberated people from heavy physical labor.
As ordinary people, we don't have to blindly pursue every technology hotspot, but we can keep an open mind to understand and learn. When we really need it, we at least know: Oh, what can this thing do, what risks it has, and whether it is suitable for me.
Openclay or other AI tools are just tools. Tools are not good or bad. The key depends on the person who uses them. Let's look at technology rationally, embrace the changes of the times, and look forward to the next high level of intelligent automation.
