AI coding has made me much more productive—but at a strange cost. I can't listen to music when I code anymore.
What AI coding has taken away is the time where you know exactly what you want to implement and have a rough mental model of how to do it—like what files to modify, what data structures to create, what to test for—and then you put some headphones on and just CODE. There was a beauty and joy to this part that I miss, a flow state you can hit with a nice linear progression, like building Legos or going on a hike.
This process has been replaced by prompt-crafting—writing a blurb that contains the same mental model and letting the AI code it up. Followed by looking at the output, and iterating on the prompt till you get what you want. This part is much more jittery—you can’t course-correct half way through, so you just evaluate outputs and often revert and try again from scratch. It also happens so fast that you have way more code to read and understand. I’ve actually made the font size in my editor smaller since I’m scanning files at a much higher level, rather than diving deep into a specific function. Anecdotally, I feel like I’m using more brain cells now than I did with pre-AI coding, or maybe just different brain cells. Whatever it may be, I can no longer listen to music during this process and get things done.
For me, apples to apples, a 4-hour session of AI coding is more cognitively intense than a 4-hour session of non-AI coding.[1] Is this cognitive load worth it? Absolutely—I feel like I can ship at a crazy velocity now, like I have a team of interns at my disposal to code up my every silly demand.
This makes me realize that programming was never the hard part—it was just another way to express what you want to exist in this world, a medium for thoughts and ideas, not too different from English. As AI models get better, the abstractions we work in will get higher and higher, and my “font size” equivalent will get smaller and smaller. But the challenge of thinking—the mental load of going from foggy thoughts to structured ideas, and then from structured ideas to clear outputs—will remain, and maybe even grow. Overall, I’m super optimistic about this future.
I do miss listening to music though…
1. I primarily work on medium to large, business-logic heavy backend codebases. For frontend code and my side projects, AI coding seems to be even more effective and actually reduces the cognitive load, winning in all dimensions.