Top Engineers Start to Worry: In the Age of AI, How Should We “Change Careers”?
Hello everyone, I’m your financial journalist and economist. Today’s topic might be even more poignant and realistic than the question of “whether AI will replace humans.”
Recently, Liu Shengyu, a top engineer at DeepSeek, published a lengthy essay titled “I Have to Bury My Talent in Yesterday.” He is a brilliant student from Peking University’s Turing Class, who led his team to win top honors in world-class supercomputing competitions, and he even wrote the core algorithms for DeepSeek’s latest model, v4.1. Yet, this expert at the forefront of AI technology admits, “I’m not worried about losing my job; I’m worried about having to change my career.”
This is not just the anxiety of one individual; it’s an earthquake shaking the entire tech industry. Today, we’ll break down the logic behind this news in simple terms to understand what we ordinary people, especially those in the tech sector, should do in the face of the AI revolution.
---
Core Summary: From “Writing Code” to “Managing AI,” Identity is Being Redefined
In one sentence:
The speed of AI’s development has far exceeded expectations, and even the most fundamental and complex tasks, such as optimizing algorithms (once considered too difficult for AI to handle), are now at risk of being automated. Top engineers are realizing that the skills they once took pride in are rapidly losing their value. The new core competitiveness is no longer about “how well I can write code,” but about “whether I can guide AI to create perfect systems.”
Key Signals:
1. Lowering Barriers: Tasks that once required deep knowledge of hardware and algorithms are now being automated by AI.
2. Changing Roles: Engineers are shifting from code writers to managers who review, design, and oversee AI systems.
3. Industry Consensus: 75% of new code at Google is generated by AI, and the Ministry of Industry and Information Technology has confirmed a significant shift in the skills required for software professionals.
---
In-Depth Analysis: Understanding This “Technological Revolution” from Five Dimensions
1. Why Even the Most Complex Tasks Are Being Taken Over by AI?
Previously, people thought AI could handle basic tasks like writing web pages or scripts, but more complex areas like chip and GPU performance optimization seemed beyond its capabilities, as these required human understanding of hardware and mathematical models. However, Liu Shengyu’s experience shows this is no longer true. He works on algorithms, which act as the “interpreters” and “accelerators” between AI models and chips. In the past, this involved manual, painstaking code optimization. Now, AI can read CUDA code, analyze instructions, and optimize performance on its own. Liu’s says AI thinks hundreds of times faster and can process tasks in parallel. This means that AI is quickly taking over fields we thought were beyond its reach. For professionals, this means their skills are becoming more general, reducing their uniqueness.
2. “Unemployment” vs. “Career Change”: Why Do Top Talents Choose to Act Proactively?
Liu defines it clearly: He won’t “lose his job” in the traditional sense, but he must “change his career.”
- Unemployment: Losing a job means having no income or place to go.
- Career Change: Your current role becomes obsolete or less important, and you need to adapt your work methods or even your mindset.
His anxiety stems from realizing that individuals can’t slow down the pace of technological advancement. Just as horsemen worried about cars, programmers now worry about AI. Instead of resisting change, he aims to be the one who drives the change. This proactive attitude is essential for future professionals.
3. Data Doesn’t Lie: AI Programming Has Become the Main Force
Many still think of AI programming as a supplementary task, but the data shows a dramatic shift:
- Google’s Data: By October 2024, 25% of code was generated by AI; by next autumn, it’ll exceed 50%; now, 75% of new code is AI-generated.
- Industry Surveys: 72% of developers use AI tools daily, and 42% of code is created or significantly assisted by AI.
This indicates that coding is becoming as basic as typing—no longer a high-paying skill.
Google CEO Sundar Pichai says engineers’ roles have shifted from writing code to reviewing it. Your value lies in assessing the quality and reliability of AI-generated code.
4. Beware of “Code Chaos”: Are Engineering Skills More Important in the AI Era?
Contrary to what you might think, engineering skills are still crucial. Liu and Microsoft CTO Kevin Scott emphasize that AI-generated code doesn’t equate to completed software engineering. If someone without architectural or logical knowledge uses AI to write code, the result can be chaotic and flawed. Engineering skills, such as system design and problem-solving, are more important because AI handles the “how,” while humans define the “what” and “why.”
5. New Job Directions for the Future: From Pure Technology to Multidisciplinary Experts
The Ministry of Industry and Information Technology suggests that future high-paying roles will require more than just coding skills:
- Industry-Specific Experts: Those who understand both technology and their field (e.g., AI algorithms and medical processes).
- Human-AI Collaborators: Who can coordinate multiple AI systems effectively.
- Field Engineers: Who can solve real problems on-site, requiring strong communication and industry knowledge.
Liu’s metaphor of an “AI-driven vehicle” illustrates this: You don’t need to write every line of code, but you need to know where to direct the AI and how to ensure it works safely.
---
Insights for Ordinary People: How Should We Respond?
Although the story focuses on top engineers, this transformation affects all knowledge workers:
1. Give Up the Illusion of Technical Fortresses: Don’t rely on a single skill for a lifetime; technology evolves rapidly.
2. Develop Meta-Skills:
- Questioning: Can you ask AI meaningful, logical questions?
- Judgment: Can you evaluate AI’s output quickly?
- System Thinking: Can you design solutions from a holistic perspective?
3. Embrace Human-AI Collaboration: Use AI as a tool; focus on setting directions, making decisions, and ensuring quality.
4. Deepen Industry Knowledge: Technical skills are replaceable, but combining them with industry knowledge makes you more valuable.
In Conclusion:
Liu’s words, “I have to bury my talent in yesterday,” symbolize a necessary awakening. In the AI era, stability comes from continuous evolution, not stagnation. We shouldn’t fear AI; we should fear staying behind as it evolves. The future belongs to those who can harness AI and understand human needs. This career change has already begun.