I have been asking this question a lot lately: is it really an ego issue? Why using AI tools give me this uncomfortable feeling, even when I like the result?

First of all, I am not anti-AI, and I am actively using it today for my work and even for side projects. I am not saying it is not useful, and I am not going into the ethical or environmental issues of AI usage. It is not in the scope of this rant.

I am talking about personal feelings and career goals.

A viral post from Andrej Karpathy, one of the founding members of OpenAI, highlights how using AI coding agents has transformed his work while admiting that it does “hurt the ego a bit”:

It hurts the ego a bit but the power to operate over software in large “code actions” is just too net useful, especially once you adapt to it, configure it, learn to use it, and wrap your head around what it can and cannot do.

I favorited his original post on X because I found it insightful, and since then I cannot stop thinking about it - is it really an ego issue what I’m feeling?

I never took issue in receiving contributions or feedback on my code, and I always thought that collaboration is key for software projects, especially open source ones. I never felt it hurted my ego, so why would a non-human contribution hurt my ego?

I think the issue goes a bit deeper than that. I used to feel proud of my code and my engineering work. I would look at a semi-finished project (well, all coding projects are semi-finished, as software is never complete) and think “wow. I built all this!” And that was enough to make me happy, even if the project was not perfect (it never was) and the TODO list was big. Even if it took me several weeks to get there. I could see the repercussions. People would be genuinely interested when the project was useful to them.

Now, code is cheap. It doesn’t matter how the code looks, and people just don’t care anymore, because they can “build a replica with Claude Code in less than one day”, getting whatever engagement might come from yet another copy of something that could have been good if people would just collaborate more. At the same time, open source maintainers are already overwhelmed by low-quality, poorly designed pull requests made entirely with AI tools where folks don’t even understand what the changes do to the project, ignoring future repercussions and based on naive assumptions.

Yeah, I built a cool app with Copilot in under a month, but I just don’t feel like open sourcing it. I understand most of the code, but I don’t feel proud of it; is this what we’ll be doing in the next 5 years? Do we really need all this velocity, instead of the slower but steady pace of well-crafted, thoroughly designed software?

And most importantly: who is gonna review all this shit?

Review fatigue is real, and it sucks. As much as you want to oversee all steps along the way, there is no way you can ship at that speed if you thoroughly review everything the LLM does in your codebase. Been there, done that. It’s just not possible. You start with good intentions and sharp eyes, but at some point you’ll just let go.

Is this the future we want after years, decades of study and practice improving our coding skills? I understand that “the genie is out of the bottle” and there’s no way it’s coming back, but the environment that is being shaped up is just not sustainable in the long term! With less people willing to share open source and less people learning in public (see the decline in Stack Overflow), soon there will be no more “free lunch” for LLMs to get trained, and the overall quality and security of AI-generated code tends to get much worse.

But yeah, let’s call it an ego issue 🤙🏻