As software engineers, our identity was never "the person who can write code" - it was "the person who can solve problems with software." For 50+ years, writing code has been the unavoidable tax we paid to materialize ideas into running systems. We got so good at paying that tax that we mistook it for the actual value. Folks are unsure if AI fundamentally changes engineering. Watch what happens next. The engineer who feels this loss most acutely is often the one positioned to create the most impact. Because now the constraint shifts from "can I implement this?" to "what should exist?". Taste matters. Product intuition. System design. Understanding what users actually need. Knowing which of seven possible approaches will scale. These don't get commoditized when the syntax does. The three-person startup building four products instead of one? They still need someone who knows which four products are worth building. The enterprise team trying seven approaches? Someone has to recognize which one actually solved the problem. Your identity evolves from craftsperson to architect, from implementer to multiplier. The question isn't "who am I now?" - it's "what can I build now that was impossible before?" #ai #programming #softwareengineering
This week I had the total opposite epiphany as I saw AI agents crash and burn while writing the most basic of tests. They couldn’t do it. I tried the most basic to the most powerful of models and the tests were just not good. They were over engineered or under engineered. Tests are the powerful tool that we have to benchmark if the model has written good code or not without first inspecting that code. So when I looked at the code, it’s was a chaotic mess. While I celebrate the rapid developments that we are having in the artificial intelligence space, we are still far from a model that actually performs critical thinking. The tests lacked intuition and common sense. To your point, the value of an engineer has never come from the act of typing code, but from the act of thinking critically on how to solve a problem and then translating the solution from analysis, conversations, and discussions into code. AI is still far from that. And, even when AI agents get to the point of “thinking critically”, humanity will have newer and larger problems to think about.
The engineer of the future is the person in the middle. You connect the tech implementation with real world requirements and act as the bridge for people who need digital solutions. All your hard skills and your tech understanding will still be important, but you will have to shift your skill set towards soft and human skills, let's call them skills beyond code.
I relate to the feeling, but it's been said here plenty before: Being a software engineer is not about writing code. It's about solving problems. For me, AI has 10x'd what I've been able to accomplish because I'm constantly asking the question, "how can this be better?" and I now have a tool to make the ideas come to life where before I'd barely have the time. AI is just a new tool for a software engineer. The fundamental skills that allow you to get the value of AI that a vibe coder cannot are what makes you an engineer.
Perhaps the identity of a software engineer was never truly about how fast we write code, but about how we understand problems, ask the right questions, make trade-offs, and take responsibility for the systems we build. AI can write code, but it does not live with the consequences of technical decisions, nor does it understand the human context behind the problems. The craft is changing, the tools are changing, but the core value of engineers is not disappearing — it is shifting.
Exactly Addy ... The writing was never the most important part of telling a story. It was just the means to the end. The end of software engineering is making things that matter. And if AI makes us do it faster and better, then that's a great collab. Love how you talk about materializing ideas by asking what should exist. AI unshackles us from "what do I know how to do" and moves us to "what should exist" as you say.
The real challenge isn't "AI coding." It's leveraging AI to optimize the *entire* SDLC for existing teams and legacy stacks, not just greenfield. That's where tangible leverage is missed.
In another words developers has to go up in the value chain to flourish
I hear you, and personally feel very empowered by AI, but I do feel a real sense of loss, if not for myself then for others. Many programmers happened across it as something creative and fun to do. A thing they could excel at. They felt it "click" and made it into prideful work they could do every day. It was a mixture of artisanal craft and grunt work through myriad bugs and strange puzzles. Sometimes art more than science. Making programmers into 'value creators' and 'impact ideators' is a bit gut-wrenching. Perhaps the problem is that there was never a single archetype, but many. There were people who built software because they wanted to create value and manifest ideas, ... and they'll be fine I'm sure..., and then there are those who just wanted to build. They felt it.