AI Is Eroding Pure Code Skills; Engineers Need Psychology and Product Sense Now
Translation: What Should Software Engineers Do Now?
I've heard people say that to stay competitive, software engineers (SWE) need to keep learning coding and various tech stacks. I don't agree.
Every time a new model is released, the value of these purely technical skills shrinks a bit. Every day I feel more like a product manager managing AI, and the need for deep technical knowledge is decreasing. The "technical moat" I thought was solid has long dried up. Now, writing a user story to the model as a prompt works even better than a rigorous technical solution. The skills required for the job are becoming less "technical" and more "human."
Sometimes, seeing AI solving the hard problems I used to do, like developing SaaS software, makes me feel a sense of loss. In the past, I really enjoyed the satisfaction of polishing an application and watching it become a reality step by step. Watching the skills I worked hard to learn become almost worthless is indeed disheartening. Building SaaS applications was once a high-barrier problem, which is why I loved it. But now it's no longer a hard problem, so I must shift to areas where real problems still exist. I need to think bigger.
One of my favorite books is The Beginning of Infinity (by physicist David Deutsch), and there's an idea that impressed me deeply: human innovation has no end — "as long as we don't violate the laws of nature, as long as we have the right knowledge, anything is possible." Every time we create something new, thousands of new possibilities are born. The invention of computers eliminated many old jobs, but it ultimately created far more new ones.
AI is a remarkable innovation that gives us the ability to do more, and it demands a new way of thinking. In my old way of thinking, building a SaaS app was already a big deal, but now it's not like that at all. Today, developing a complex enterprise-grade software is even easier than building a simple website five years ago. "Doing big things" means thinking on a larger scale — national, global, or even cosmic. Imagine what you could do if you had a team of 100 software engineers. What about 1,000? Now we can have that level of productivity at a very low cost, so we must never limit ourselves to just building a few apps.
As engineers, we also need to broaden the boundaries of our abilities. We can't just focus on technical skills; we need to step outside the tech circle and understand the entire business loop, including product, marketing, and sales.
To stay irreplaceable in the current environment, my approach is to learn things that AI can't help with: product positioning, creativity, how to delight humans, psychology, marketing, and so on. These are "human problems" that AI can never solve independently, because there is no standard answer and no ready-made test set to train on. AI is extremely good at pattern matching, but it can never truly understand what it's like to be human. And that is exactly where the irreplaceable value of us humans lies, and it's the direction I'm doubling down on.