🚨 Warning to Web3 developers, builders, artists, and contributors. My X account, GitHub, and several social accounts were compromised after a fake Web3 recruitment process. The project involved in this incident is LumaPad — @LumaPad_pro. I am writing this thread to document the full timeline: how trust was built, how I was pushed into a fake StreamYard meeting, how my machine was infected with Trojan:Win32/NetSupportRat!AMTB, and how that led to my X, Instagram, GitHub, and important development data being compromised. This is not drama. This is an incident report and a warning to the community.

Haha I want 1 duck

A lot of developers are worried about one very specific question: AI can write code incredibly fast now, so will developer salaries still hold up? I do not think the answer is simply “developers are safe” or “developers are going away.” What is actually happening is that the market is becoming more divided. Work that mainly involves repetitive coding, simple CRUD features, basic UI changes, or clearly defined implementation tasks is becoming easier to automate and, as a result, easier to price down. At the same time, engineers who can take responsibility for real systems, make technical decisions, and fix things when they fail are still very valuable. Recent market data points in the same direction. Indeed Hiring Lab has reported stronger wage growth in occupations with high exposure to AI, including software development, compared with occupations with lower exposure. But that growth is not distributed evenly. Senior roles have benefited much more than junior roles, while the share of junior openings has fallen. PwC has also reported a significant salary premium for jobs that require AI-related skills. Data from platforms such as Levels.fyi and Dice also suggests that engineers who use AI effectively can earn more than people at the same level who work in a more traditional way. So AI is not simply increasing the salary of every developer. It is increasing the gap between developers. The lower end of the market is becoming more crowded and competitive. The upper end is becoming more expensive. That leads to another question. Why would a company still pay a software engineer a high salary instead of simply buying more AI subscriptions? Because generating code and running a real system are two very different things. AI can generate a feature in a few minutes. But once that feature reaches production, the situation becomes much more complicated. It has to work with an old database. It has to fit into an existing codebase. It can introduce race conditions. It can corrupt data. It can increase infrastructure costs. And it can fail in ways that are difficult to diagnose because the current logging is not good enough. When the system breaks at two in the morning, the company still needs someone who understands what is happening. AI can suggest a fix. But a person still has to decide whether to roll back. A person still has to decide whether a change is safe enough to ship. And when something goes wrong, a person still has to explain the situation to customers, management, finance, legal, or other teams. That responsibility is part of what companies are paying for. Another important point is that higher productivity does not automatically mean headcount falls by the same percentage. AI allows engineers to do more work. But companies often use that extra productivity to build more products, run more experiments, or let smaller teams handle workloads that previously required much larger teams. The result depends heavily on the person using the tools. A senior engineer who understands the system can use AI to move much faster. Someone who does not understand architecture, databases, or production systems can use the same AI tools and create a lot of technical debt very quickly. Both people may say that they “know how to use AI.” But the value they create is completely different. Demand for developers is therefore not disappearing. It is moving toward more difficult work. General software engineering job postings in the US have fallen significantly from the peak of the tech hiring boom. At the same time, demand related to AI, machine learning, data, and infrastructure has become much stronger. The US Bureau of Labor Statistics still expects software developer employment to grow over the coming years. But the definition of a strong software engineer is changing. A few years from now, engineers may be judged less by how much code they can write themselves. They may be judged more by whether they can design systems, evaluate AI output, integrate models into real business workflows, maintain quality, and prevent a codebase from becoming a mess after thousands of lines of automatically generated code. In Vietnam, the direction is similar. Experienced backend and full-stack engineers at strong product companies can still earn salaries ranging from tens of millions to more than one hundred million VND per month. AI engineers who can actually deploy models into production can often earn even more, especially when working remotely or for international companies. At the same time, fresher and junior developers working mainly on basic CRUD tasks face much stronger competition. And this is probably the most important change. The gap between junior and senior engineers is no longer just about years of experience. It is increasingly about how much responsibility a person can handle. In my view, the value of a developer today comes from four main things. The first is ownership. It is not enough to finish the code and move the ticket to “Done.” You need to know whether the feature actually works in the real system. The second is understanding the business problem. A valuable engineer does not only ask: “How should I build this?” They also ask: “Why are we building this?” “What metric are we trying to improve?” “What is the cost of this solution?” The third is knowing how to use AI properly. This does not mean writing more prompts. It means knowing when AI output is trustworthy, when it needs to be checked, which tasks can safely be delegated to an agent, and which decisions should remain under human control. The final part is experience with real systems. Database migrations. Authentication and permissions. Idempotency. Monitoring. Incidents. Infrastructure costs. Customer data. These are difficult to learn just by reading code. Usually, you understand them after working on real systems, fixing real failures, and sometimes after being responsible for systems that have actually broken. So I do not think the best question is: “Will AI make developers less valuable?” A better question is: “Which part of my work is becoming a commodity?” If most of your value comes from turning a clearly written requirement into a few hundred lines of code, then AI is moving very quickly into that area. But if you understand the system, understand the product, use AI well, and can take responsibility when things do not work as expected, your value is still very high. In some cases, it may be higher than before. Because a strong engineer today does not only produce code. They can multiply the productivity of an entire team. So the main idea I would leave you with is this: Do not compete with AI on the work that is becoming cheap. Become the person who decides what should be built, judges whether the output is correct, and takes responsibility for the result. Developer salaries are becoming less connected to the number of lines of code someone can produce. They are becoming more connected to the value of the decisions that person can be trusted to make.

Ever since I started using AI, one thing has really bothered me: the garbage it churns out. Pointless lines, silly phrasing, terrible grammar, and so on. I once wanted to create something to put an end to this. Fortunately, I met some like-minded people, and it’s been amazing. Adei is one of them. Together we developed something called Antislop. It can eliminate the flaws and redundancies generated by AI, but refining it further is still a long journey ahead.We’ve only just begun. Stay tuned. Congrat @AdeiWeb2