Web3 people face "AI replacement anxiety": If you don't panic, you won't lose your job.

**The subtle relationship between AI and Web3 can best be described as competition and cooperation. **

From the perspective of capital, the story of AI is obviously more beautiful and practical than that of Web3, so that the arrival of ChatGPT can easily transfer the original influx of Web3 funds to the AI field, and the substitution relationship of capital is clearly visible. But in terms of the market, Web3, which has always been a hot topic, is not willing to give up the opportunity, and many projects are integrating it in the name of AI. The shell of Web3 is like a chameleon, changing from the metaverse to AI is just a matter of tweeting.

For Web3 workers, this relationship is even more complicated. They are worried about machines replacing their jobs, and they are also worried about being replaced if they don’t have machines. Their hesitations about machines come and go like a tide, but in the end they are not well-regulated. What is left in Web3 are just all kinds of absurd imitations and pseudo-concepts, and descriptions of workers are rare.

From a macro point of view, since the emergence of machines, the existential crisis surrounding machines and humans has never stopped. People can create things that are far smarter than themselves, but they will also fear these things. This is not only the embodiment of human wisdom, It is also the subconscious instinct to resist non-me races.

**And in the field of Web3, which fully embodies human greed and desire, this confrontation can only intensify. **

Will Web3 workers be replaced by AI? With this question in mind, the author interviewed 4 workers who are using AI in the circle. In different jobs, some people are pessimistic, thinking that they will replace their own jobs in the long run and undergo transformation; Regarding the intrigue of human beings; although the opinions are slightly different, the interviewees agree that AI will never have the upper hand in the struggle against complex human nature.

**It seems that in Web3, which is difficult to understand even AI, at least it is safe to use it as a worker. **

01. "AI can only guarantee the correct process" - Ivan, a small asset management asset operation

Ivan, who works in an asset management institution, can be regarded as a senior worker to some extent. His main job content is a series of work around funds, including storage, value-added and operation. Control is more similar, that is, monitoring transactions, identifying possible risk points caused by transactions, and effectively intervening. From his perspective, AI is far from enough to have an impact on his work, because it is difficult for AI to identify human problems. The following is his self-report:

My job is more pleasant to hear. I work in asset operations in an institution, and realize asset appreciation and risk control through operations. Before this job, I worked in risk control at a local bank. In terms of the nature of the work, there is actually no qualitative difference between the two. The difference is that risk control places more emphasis on monitoring and early warning, while operations also have value-added needs. .

After the emergence of ChatGPT, our boss mentioned the need to learn to use it to improve work efficiency, and also asked so-called professional questioners to give us training courses. However, since there is no mandatory requirement, as far as I know, the frequency of use is not actually used. not tall. **But this is not to say that ChatGPT is not useful or is simply a criticism, but that for our industry, it is difficult for AI to replace human work. **

To give a very simple example, many of our large B-side customers transfer millions of dollars. The risks of using AI for approval are obvious, because it is difficult for AI to understand the process and conspiracy behind the money. It may be just a sum of money on the surface, but behind it may be a complex composition of money laundering, distribution, fraud and other events. This is actually very similar to a bank. It seems to be a very automated process, but every node is reviewed one by one by the manual force of the bank's backend, and several or even a dozen departments are operating around the money.

Relatively speaking, retail customers will have a higher degree of automation. The AI in the background of the occasional witch, double flower and other wool-picking operations can automatically identify it, but it cannot be completely left to AI. Just like when we apply for a bank card, in the last step we often ask the account manager to come and verify the identity on site. Why? **Because it is difficult for AI to test people through the process. AI only needs to ensure that the process is correct, but only humans can identify human problems. **Looking at the entire financial industry, although digitalization has improved efficiency and there are various complicated systems, manual labor is still the main force. Anything related to money needs to be held accountable. This has happened many times in history. lesson.

But if you ask whether AI is completely useless, it is definitely not, otherwise the giants would not devote all their efforts to development. Taking myself as an example, Paperwork will use AI when it doesn’t want to move, and then polish and modify it. Let me also mention that assignments, contracts, etc., which everyone thinks can be replaced, are not that simple to get. These are templates that have been reviewed and approved by the legal and financial people one by one. So you see, people are the key. the most important.

02. “What’s more important is what AI can do in the future”—Iris Content Media Editor

I work in an editorial position in a content media, and it is generally considered one of the positions that are most likely to be replaced by AI. In general, AI, such as GPT, is relatively common in our daily work. Especially when the industry is in recession and everyone's originality is decreasing day by day, we will use AI to complete tasks such as translation and news capture, which can greatly improve efficiency.

In the original direction, according to the content requirements of the manuscript, the role of AI is also different. For word processing, GPT's ability is still very strong, so the basic outline of the draft is basically the same. However, due to the nature of the industry , it still has limitations in understanding many professional terms or in adapting to people's reading habits, so it requires manual adjustment from multiple angles. The correlation between in-depth manuscripts and AI is relatively weak. In-depth manuscripts are more about personal field visits and investigations, which are difficult to replace with AI. Even if AI is fed with data to produce manuscripts, the effect will be poor.

The performance of slightly lower versions of GPT is even more unsatisfactory. A colleague used GPT3.5 to write a manuscript. Since there is no data after 2021, the latest data needs to be split and fed in by the colleague for him to learn first. After trying for a whole day, he finally published a manuscript, and most of it was correct. There are even some nonsense contents, and there are even random hallucinations. In the end, it will either be drastically changed or become a scrap. Of course, this is also due to our personal lack of questioning ability.

Going back to your question, do you think AI will replace humans? **I think it will definitely happen to a certain extent. Maybe what it can do now seems limited and everyone disagrees with it, but more importantly, what can it do in the future? No one can have a more powerful learning ability than AI, and no one can understand the composition of the black box in which AI emerges. This means that through continuous deep learning and analysis, AI will emerge with more intelligent capabilities at a certain stage. **Creative jobs with repetitive nature may be replaced earlier, such as programmers, editors and other cultural industry practitioners. This type of work is creative, but its essence is also easy to learn and repetitive. As long as there is text to In terms of learning, AI will definitely run faster than humans, so humans need to release more of their imagination.

**Looking back at every technological advancement, there will always be jobs replaced. This is the tool revolution brought about by the technological revolution, and it is irreversible. The Luddite movement can only be reduced to a footnote in history in the end. **Currently, since the bear market, the company has laid off 40% of its employees, and the content editing department has even fewer people. Maybe this has something to do with AI?

Personally, I will try my best to find the more core part of the work, the part that is difficult for AI to replace. For example, I will pay more attention to the opportunity to go out and the accumulation of contacts than before. No matter what, it is difficult for AI to touch link.

03. "Legal and ethical risks cannot be avoided" - Vivian Encryption Lawyer

**In fact, this question can be enlarged. In traditional fields, can AI replace lawyers? I personally think it is not possible, for the simple reason that legal work is dealing with people, and the legal and ethical risks involved in it are unavoidable. **

For example, in litigation, lawyers have the responsibility and obligation to keep the information provided by the parties confidential, and this is often not defined by the good or bad of the bystanders outside the court, but by the parties involved in the case. As defined by the Center, unless it involves acts endangering national security, lawyers should keep their information confidential. The public nature of AI is destined to make it difficult to do this. The information you give him can even be sent to other inquiring parties at will, which obviously does not conform to the professional ethics of lawyers.

Similarly, in non-litigation aspects such as legal consultation, AI lawyers will also behave more inhumanely. It is difficult for them to perceive the real needs or hidden meanings of the parties. In many civil lawsuits, the parties seek round-the-clock consultation. The most important thing is how to protect their own interests, even if they are already the party at fault. But it is difficult for AI to understand this type of demand. If you ask how to reasonably avoid risks, it may ask you to tell everything, or even ask you to surrender. Therefore, at present, AI can only handle some inquiries where the problem is clearly understood and the answer is already established. , and do some non-private desk work, the limitations are obvious. **

It is also worth noting that AI will have machine hallucinations, which is also a taboo for legal service workers. Some colleagues fell into a trap some time ago. Lawyers from Levidow & Oberman Law Firm in the United States submitted documents written with AI assistance in a dispute. As a result, the judge discovered that some of the precedents did not actually exist. In the end, the firm was fined for providing false information to the court. A fine of $5,000 was imposed. False information is a violation of the Lawyer Law, and it is also a rule that is expressly prohibited by the law, but until now, the illusion of AI has not been resolved.

In Web3, the above limitations will be further magnified. Web3 is a very chaotic and rapidly changing industry. Although domestic and foreign laws have been promulgated and improved in recent years, virtual currency is still a new thing in the civil and commercial field. Judging from the examples of court judgments in various places, the understanding is not unified. In a field where new concepts emerge frequently, using AI to act as a lawyer will only bring more complex problems, such as stablecoins. Regulations in various regions are very different. In Hong Kong, this field is still a vacuum zone, but Singapore and Europe have already introduced regulations. Regarding the regulatory plan, it is difficult to get a targeted reply when you ask AI.

In addition, lawyers engaged in the virtual currency industry are highly qualified. In addition to having a deep understanding of the industry, they must also have strong financial services supervision and securities legal capabilities. They must act as a connector between the court and the parties and reduce information gaps. It is not easy to do this with noise, so I think that AI can assist in desk research and other text-based work to a certain extent, but it is far from being able to replace it.

04. "Elimination is just a natural social law" - Leo Developer

In fact, this problem had already appeared in Alpha Go a few years ago, and many automatic programming software appeared one after another. This time, due to the amazing performance of ChatGPT, it caused a panic in the circle.

If this issue was placed before last year, I would say categorically that it will not be replaced, but now, I may think that some programmers are at risk of being replaced. Judging from the advantages of AI, AI can already deal with slightly difficult technical problems such as arrays/strings and dynamic programming. At the same time, it has certain capabilities in common problems such as repeated code generation, documentation and annotations, and testing.

In my daily work, I also use Github Copilot. I am not alone. GitHub conducted a survey before. Among 500 developers, 92% of the respondents were using AI coding tools to complete their work and other tasks. project. Using AI tools is not just for other reasons, but because it can actually improve efficiency. Our work is essentially the translation of machine language and human language. There are many repetitive specific tasks. On these tasks, the chain call performance adopted by AI is relatively good.

**But for developers, the core ability is programming logic, and the most difficult node is building requirements, not just simple programming. In this field, AI does not yet have complete engineering capabilities. **

In actual software engineering, code is written according to customized requirements. Within this complex relationship, the objective laws of the functionality between modules, technical background, and the product itself are difficult for AI to discover through the learned database. The reason is that this type of data is usually confidential, which means that AI is difficult to compare with humans in terms of business abstraction, modeling and architecture. In addition, AI also has issues such as code security and intellectual property rights.

On the other hand, compared with the traditional Internet, Web3 has decentralized differences in the front-end and back-end architecture. The most intuitive thing is that after the smart contract code is released, developers cannot simply patch and update it, while Web3 is more about money. With high sensitivity, the option of leaving people alone will bring about many practical problems. In the industry, you can also see that there are many local dog projects running away, and anonymity often makes people feel unsafe.

It is often joked in the industry that when there is no problem, go to the centralization, and if there is a problem, you still have to go to the centralized organization, so in any case, people themselves are very important in this process. But from the perspective of the programmer industry, it is only a natural social law to eliminate programmers who will not use AI or low-level programmers who only use limited tools such as CRUD in the future. **

05. Conclusion

In a performance contest with machines, humans, made of flesh and blood, are often at a disadvantage. Therefore, not only Web3, but also in many aspects, it is not nonsense to say that machines will replace humans.

But precisely because of the existence of fragile human nature, the complex network formed by the human group and the spiritual links in it are difficult for machines to get involved, and this also feeds back on humans.

In the future, it may be more important to protect and cherish one's own humanity, release one's creativity, and not become a walking dead in the reinforced concrete urban forest, and regrettably become the nourishment of AI in the end.