Programming is like a game of chess for me: winning (i.e. delivering a product) is important, but only if I played this game myself. I don't enjoy winning if my opponent disconnected. I don't enjoy beating people who don't know how to move pieces. I don't enjoy winning with an engine, and coding agents are basically "winning with a chess engine".
I started to use AI for coding a long time ago, when copilot was first introduced. At first, I was happy: "wow, this can write boring boilerplate and unit tests". Time went by, agents got better, time spent writing code myself went down. Half a year ago I suddenly realized I hadn't written any code for over a month, and I am a full-time software dev without any managing work. Not only did I stop writing code, but agents also became pretty good at proposing plans and architectures, and I was mostly providing missing context and technical details of our infrastructure.
One day the realization hit hard. Coding agents removed everything I loved, and - mostly - amplified everything I hated. My role is to gather context from other humans and provide a detailed explanation (similar to a well-written jira ticket) to some super intelligent engineer, then review the result and provide more context/explanation if needed. So, basically, a manager.
A couple of years ago I could work on a single coding task for weeks. I knew all classes, all functions, I was slowly building up information at a comfortable tempo. Now I am forced to increase my mental throughput multiple times. What took a couple of weeks now takes a couple of hours, and I still need to comprehend that. I always feel tired, I don't have enough mental capacity to hold that many abstractions and technical details. So you either autoaccept or get burned out reviewing tens of thousands of lines.
One more downside is that code has 0 value. You can code your own OS from scratch, and people will think "ye, I can vibe code that in a weekend". I don't even know what is valuable nowadays.
Maybe I am a bad employee. Increasing employer profits was never my goal. I always loved to solve puzzles, and for the last 10 years this was well aligned with my employer's desire to increase profits. I never was good as a manager, or brilliant at generating ideas. My power - and passion - was solving complex technical problems with my hands, fast and with good quality. Now it is gone forever. I feel completely lost and unmotivated. I don't see any value in learning anything related to computer science, reading papers, contributing to open source. I hate my new "AI delegator" role and I hate my work. I loved coding so much (I use neovim btw) that I was sure I would do this until I die, and now I don't know what to do with my life. This is the only thing I know, and, most importantly, the only thing I am/was passionate about. Anyone else feeling the same?
We used to be obsessed with elegant design and would lose sleep over an optimization. Since vibe coding came along, those days are gone for good. As technologists, we typically feel anxious about losing control over our code. The speed of AI makes it nearly impossible for us to carefully review the code.
We know AI makes mistakes all the time, and our job has become that of an AI corrector.
You are not wrong to have your feelings or to strongly identify your sense of self with your work; that is what we used to call “finding your vocation”.
I share your feelings and have been focusing my energy on learning local and open weights AI to work through it because it seems to be impossible to avoid it all entirely as a freelancer who needs to remain competitive, and a bit of righteous/ethics/sustainable process fire in my belly doesn’t hurt.
But focusing on how you feel first, and then hopefully thinking about other ways you can feel valuable, is not wrong.
You are not alone and I wish you the best.
I'm always curious, what are those "great things" requiring deep technical knowledge and years of expertise that these visionaries can finally provide with the help of LLMs?
I almost feel ashamed/mentally ill when i am reading how everyone is happy that their true visionnaire potential is finally unlocked while i just like to solve my silly puzzles.
Everything has, in fact, changed, and it is a task for those of us who cared about what was, to try to work out what is worth caring about.
As Annie Savoy says in Bull Durham, “the world is made for those who aren’t cursed with self-awareness”. You’re going to have to figure out new things to care about, but you can do it, because you found it the first time!
What did you found for yourself?
I think the architecture around systems is un-replaceable like what I do now, is I create specs and then asks the agents to cook exactly as the spec.
What's your AI workflow btw?
>architecture Give me a reason why agent can't create great architecture when enough context is provided?