Who are you? Who am I?

To get into software engineering, one had to hold two particular viewpoints toward the craft. You either enjoyed it for the money prospects, or you enjoyed it for the problem solving and building software with your own intellect and your own hands. It's no surprise that the joy of programming for most has been dissolved since large language models have been introduced, by and large.

Through striving, pain, loss, and gain, you can find real satisfaction in the works your hands created. If you're involved in the trades—maybe woodworking, leather crafting, glass blowing, or painting—there is real pride in the effort required to create what is in front of you, a manifestation of your thoughts and ideas.

But now we step back and prompt an AI for the creation we would like to see. We have disconnected ourselves from the process of creating something, so we no longer have to endure the trials and tribulations along the way.

Instead, we stare at our terminal and ask Claude to investigate a bug, praying it is correct because we know we won’t read the code. We become more comfortable letting AI create what it thinks is the right outcome to implement a solution to a problem.

I find myself down this route. Sometimes, depending on how much I let AI take the wheel, for personal projects I don't mind saying, "Claude, build this thing, make no mistakes."

But there is no universe where I am fully connected to the code it creates and to my product. I become a glorified product owner and QA tester. I don't have an emotional connection to what I've written or to the code in the code base. I can no longer tell you where a bug might be when it arises; instead, I rely on Claude to do the debugging and leverage more tools such as Sentry. Because I didn't go through the process of creating, I merely read whatever Claude or another AI outputted. I therefore cannot tell you what might be the cause of a bug or how to fix it.

While it's true that you can build faster than ever now, and achieve more than what was previously possible, is this a good thing? Now we don't have to go through the turmoil of building an idea to fruition. We can sit back and prompt our agent, spinning off more agents to spin off even more agents to run CI, bug‑fix, and test. We can offload all of our critical thinking, and instead of flexing our cognitive process to reason about a problem, we hand it off to multiple agents. We've become disconnected.

I think, while it is wildly beneficial that AI does exist and offers a vast array of benefits, we are losing pieces of ourselves in the process. We are trading a part of us for an outcome. I think it would actually be a disservice not to embrace AI, not to use it. It's just undeniable how fast you can move as a developer with the right tools.

Hard times create strong men, strong men create easy times, easy times create weak men, and weak men create bad times. The cycle continues. I think this is no different here, although maybe a little less complicated, a little less circular—perhaps just binary. But identities are being shredded.

Who we are as software developers is being eroded to a degree. The fun of software development is being dissolved unless you intentionally set aside time to program with your own two hands, like a traditional coder of past years. For many of us, our identity has been wrapped up in what we can produce with our intellect.

Now that’s gone, we must either deliberately keep our skills sharp and program something with our own hands, avoiding the allure of handing off to an agent, or succumb to it and become one with our agents.

I for one, am not so strong as to resist the allure of using agents.

I think I've probably gone a little bit circular in this post, but it's more just that there's a feeling I can't quite put my finger on. There's a greater concept that is fuzzy around the edges, which I feel to be true, that is hard to put into words right now: things still evolve. We see the impacts of large language models not only on industries but on us as people. “I could get two more features out by then. I'll go to bed in 30 minutes—just one more feature.” It really starts frying your internals, at least. It hits me at different points.

So I think, who are we? Who am I? I used to identify with being a software engineer, and don't get me wrong—I still do. I am a software engineer. LLM or not, I can still produce good code and scalable features. But my identity now has to separate from that, and I don't know what that looks like.

I guess it depends on your philosophical view. Now we can't find our identities and what we do anymore, not that we should, but we have to look elsewhere. Can you say that you're a software engineer with pride? I guess that would depend on what you do and who you're doing it for.

It used to be a little more simple. “Yeah, I'm a damn good engineer. Look at this linked list traversal I did with my eyes closed." Isn't it efficient? It saves both space and time.” I think we're further away from that kind of pride.

I think it'll be interesting to see where we end up. What will the future look like if all of this goes to hell in a handbag? Will we return to our roots as programmers using open weight models? Will we reject AI at a societal level and program with our own two hands? I doubt that, but that would be fun.

I think, in the end, I'm just me. My identity isn't in my work or what I produce; it's a culmination of my beliefs, intents, morality, hobbies, and the like. I am me.