Coding Agents killed my identity. How do you feel?

People who never really had these feelings about coding will turn up to tell you that you are wrong.

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 a vocation.

I share your feelings and have been focusing my energy on learning local AI because it seems to be impossible to avoid AI entirely.

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.

Code generated with AI has no worth (i think), even if its a core piece of any system.

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?

Indeed. Vibe coding is making those of us in tech lose ourselves more and more.

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.

I mostly strongly identify with this. Been coding for ~35 years, if you count toy programs on a Spectrum. My world is being completely destroyed. And it's probably normal. I also don't understand TikTok, new Avengers, or queueing for bubble tea. That's why evolution invented death (and renewal).

One outlet in all this AI craze is finding creative use of it in mundane work. For example, recently I made SIMD-level optimizations for XLS parser that is used maybe 2 times per day. Before AI I could never waste a week or two on doing something silly like that.

I'm sure the "price reckoning" will come to us when VC will find out they didn't have infinite money, but for now we can have some level of fun even with AI.

I relate to this on a deep level. I have been writing software professionally for 18 years. Programming was the fun part of the job. Solving deep technical problems, digging through docs, etc. All that seems to be gone now. Unfortunately the business does not care about any of the above, all that matters is output. My job pays me for output therefore I have no choice but to embrace the Claude pandemic.

I wish I had better news, but really just want you to know you are not alone and I empathize with you. I do think there is a cost to pay when using AI to output all of the code. A deep mental model is impossible and this will have downstream impacts. I am hopeful this will settle into a happy middle ground where AI is used but not relied upon.

I have a CS degree and have been a SWE for 13 years but never really developed a love of coding or software in general, I just happened to learn it and be pretty good at it in 2008. But even I have lost a bit of my identity, having one of the few things I have done most days for a decade+ be commoditized is pretty brutal no matter how I feel about software. For the past few years I've been trying to see where the puck is going (I've been leaning towards moving into management in the past few years anyway) but at this point I'm just trying to hold on until I reach financial independence. I assume my viewpoint is probably square in the middle of most folks in my position.

I switched gear to learn how to train AI, design evals and explore neuroscience so we can further apply how our brains better in the new world. I enjoy coding, but it's time to adapt.

I came to this realization after I used Cursor to rewrite a 10-year old codebase (I knew every bit of the details.) and it was done in 2 days. Better, fewer technical debt and I was happy to archive the old repo.

Most senior level programmers have 2 options. (junior / early-mid level folks may want to consider re-skilling to a different field)

1. Embrace the AI and continue working and earning for as long as you can. Bank as much savings as possible. Lower your cost of living. Hopefully you've already been doing this over the past decade plus, when times were good. FIRE.

2. Don't use AI to write code. Write code yourself, as you've always done. Then, just see how long you last at work. If AI truly progresses, then long term most programmer jobs as we know them today are going to disappear. So worst case you will be let go a little earlier than your peers.

The devs using AI to write all their code are going to suffer major or complete atrophy of their coding skills. It's only been about a year of this and we're already seeing it occur. Maybe this doesn't matter long term and manual coding is truly dead, or maybe it comes full circle and we get back to programmers programming, in which case you will have a massive leg up on the NPC "programmers" who haven't written code in years because they surrendered their craft -and brain- to AI agents.

Been programming since 12 (now 44). I still remember the pure joy of writing the QBasic and Turbo Pascal programs to draw random colored pixels at random locations on a mode 13h or interface with a sound blaster.

Went through this identity crisis super strong about a year and a half ago.

I feel like I'm mostly over it now. The surface area of interesting things to learn seems ever expansive. It's fun to find the frontier and the spaces where the models still struggle, and then work together to overcome a hurdle and achieve something neither of us could have achieved individually.

I suppose that frontier too will thin and eventually vanish.

For now though I think it's best to take whatever time you need to be nostalgic, but then realize how awesome it is to get to be a programmer at this inflection point.

Sorry to dissapoint but “writing code” or “programming” per se was never a goal to begin with. It was and is just an obstacle, a mean to an end at best.

Therefore it’s not AI fault that you built your identity on sand.

The goal had always been - delivering value, solving problems.

How about building your identity around that and see what happens?

In this context AI is nothing but good news, with proper attitude and skill it helps achieving the goal better, faster.

[flagged]

I remember when my dad brought home our first computer, a Gateway 2000. A new world opened up for me. Bits, bytes, and hacking away at HTML and C++ until the wee morning hours just before my parents would wake up. I was a programmer, and this was my identity.

That computer is where I downloaded my first porn and where I met my first girlfriend on IRC. We would role play almost every night about a cyberpunk future where we both became sexualized variants of the characters in Snowcrash. I never talked to her in real life or on the phone, but it felt like we were physically dating.

Eventually, we saw the advent of MySpace, Facebook, and all this other fucking bullshit that nobody should give a fuck about. I don’t post pictures of myself anymore because I was bullied on these platforms for my appearance. I’m a 400-pound man that is five feet tall. One of my hands also stopped growing when I was three years old. They would call me “that fat midget with the small hand.” Very clever. Fuck these stupid ass websites and the people on them.

Fast forward to 2022. I’m a lead software engineer at a mom-and-pop shop that grew rapidly. I have six software engineers under me. Teaching them things about programming annd exploring the wonders of logic and reason. Telling them to rid themselves of superstition and religion. They saw me as a god, and I felt like one, too. I had honed my skills at typing code into my computer and making it run. Every day I felt like John Wick.

Fast forward to four months ago. I’m reading trash 1,000-line pull requests and pulling my hair out by how awful my engineers have become. I tell them I don’t care how efficient AI is. They don’t understand and appreciate the craft. They all look at me like I’m crazy. I tell them that we are going to stop using AI and going back to the old ways. The sloppification of my profession will no longer be allowed.

Fast forward to a month later. I’ve grown to be 450 pounds. The mom and pop of the shop tell me that people have been complaining that they can’t use AI to code, and two people complained about how I smell. (I remember the days where this was a sign of genius and something that was occasionally understood. Sometimes I have better things to do than shower.) I tell the mom and pop to shove their AI in their ass. They fire me.

Fast-forward to today. I’m writing my own operating system from scratch to compete with Apple and Windows. It is called WindmillOS, and it can only be used by people who care about hand-crafted software. I have a contact at Acer who says there is a chance it will be used as the OS on all systems moving forward because they agree it is the right thing to do. I have also exchanged emails with RMS, and he agrees the OS may be groundbreaking.

The moral of this is that everyone else is delusional about AI. Software is an expression of my soul and is an art. It is not something that can be generated by a computer. The world will soon see that human-made code is the only code that will work in the long run. Keep the faith.

[flagged]

Personally I feel a bit sad and annoyed sometimes, lots of folks I know that are working in some non technical roles e.g. lawyers in public departments or doctors in medecine come to me to show me their apps and tools and data science scripts they've written, and I have to admit their work looks totally respectable and they get it published on the app store. It can be a bit demoralizing as I feel like IT/software doesn't have any protected spaces for us engineers, whereas e.g. my friends in medecine or law feel fine as even though AI might come up with diagnoses or law expertise it won't be legally allowed to replace their function as there's a ton of red tape to ensue only certified lawyers and doctors can practice law or medecine. There's nothing like that in software engineering (at least in most fields) so in principle every "hack" can let their vibe coded data science results, software or script loose on the world. To a degree it feels quite disprespectful even. But it has always been like that I guess.

I've said it multiple times: I'm in a position where using AI is completely optional. And in a place where there's a strong divide between those that do and those of us that don't, the result is apparent: ship fast and have no clue what to do when shit hits the fan, or ship slow(relatively speaking if you have 3 major updates/week) and be able to handle the inevitable shitstorm.

Overall, I am not concerned about the slopification beyond the fact that cybersecurity is dead. Your only way out of it is to do what I've been doing and cut all ropes to cloud providers and and self-host everything, but that's a different topic.

I've had to scroll through tons of slop and the harsh truth that sloppers/clankers/ai evangelists cry over is the fact that you end up with a ton of unmanageable, over-engineered spaghetti soup that is held together by hope and dreams.

The constant data leaks, security incidents, downtime are not becoming more frequent by pure coincidence. The endless todo apps are not gaining traction because there is no demand for them at all and is the only reason why security incidents are not more frequent.

The real hard to swallow pill is that most people in software engineering were not in software engineering because of their passion or desire to be there: it was the relatively-higher pay and nothing more. AI didn't destroy that - the masks are finally down. What is sad is that more often than not, businesses are obsessed with delivering features as if there's an infinite number of clients that require them and an infinite amount of clients that would jump on board the second you release them. Sooooo. Throw a plane full of cash at anthropic/google/openai/etc. And hope for the best with a pinky-promise IOU. For those that are/were passionate about software and are not as fortunate as me and forced to slop their way though the work day, this is no doubt heartbreaking. Software is a lot like Lego - you have some basic tools, you know what you are trying to achieve and you have to fight limitations. Putting the pieces together is the most rewarding part of the process. For those people, slop machines took the most rewarding part of the job.

There's also the fact that slop is currently heavily subsidized by investors so most people are paying a symbolic fee. The real question I have is what would happen when the real prices hit? I suspect a lot of companies dying if they don't have a financial backup strategy/enough cash to invest in people figuring out what the endless slop does or simply having to shut down operations. Only time will tell.

As an older programmer being able to not waste my mental space with details and the churn of newer frameworks allows me to explore projects I wanted to build for years but lack the time and energy I had in my 20s.

The goal was being able to make things and I've judged my abilities on how much faster I can do things year after year. This speeds things up on one side but creates this backlog of things that need testing.

I still love programming but I always preferred the language and style I used 30 years ago. I've been forced to adapt to newer ways, frameworks and different methods to keep up with the times. Using AI is the new react/angularjs frontend pattern of 2015. All of these new patterns were less satisfying then what we had before.

For me I'm doing things I couldn't find time before and polishing areas I could but weren't high quality enough.

Paying for AI is a struggle mentally and trying to find a good balance is part of the newer challenge. A codex subscription gets used up too quickly or I can't find enough tasks to use up my subscription and both states cause mental burden. At work trying to maximum output while keeping within a reasonable token usage level is a struggle. Local models using cpu can do a lot for free but need to be carefully planned to limit scope. Using free opencode models works but code quality can fall apart so carefully balancing that can be hard. The code quality of top end models looks awful but can be better if you want to spend time on that aspect and trying to balance that with token usage is hard.

My biggest wonder is if job interviews are testing that you can hand code still but forcing you to use llms when hired.

For people feeling like this. There is a huge amount of work to do to make these agents work well.

It’s a new and different challenge.

> Increasing employer profits was never my goal. I always loved to solve puzzles

This is the core reason. It hit me hard too. What worked is accepting the change, and then tuning to the new world, new challenges, and new identity. It's the only option.

> How do you feel?

My experience is all encompassing. AI has attacked and destroyed almost every aspect of my previous identity over the last 2 years, as well as all of my achievements. My life today as a result is very different. Almost all of the things I used to do for fun or to pass time I no longer do.

I documented this in another comment recently so won't rehash but if you're interested, https://news.ycombinator.com/item?id=49334614

For me it's all of the writing and little projects I created which are now basically useless that hurts the most. I can try to find other ways to be useful going forward but I'll never get back decades of lost time. I have so much pointless knowledge polluting my head.

I'm a massive AI doomer though and fairly sure RSI is nearing or maybe here. Given this the next few years are likely to make any current concerns about identity seem so small and silly that I'm trying to focus on the things that matter and trying to live in the moment.

I'm basically living like I've been diagnosed with terminal cancer – just trying to absorb as much of life as I can and not worrying too much about the future which has been taken from me.

I have been shadowing this thread for a while. My past engagement on hackernews have largely towards GenAI, programming & what this industry looks like.

First, it's good to know I am not alone in feeling that a certain world, no longer exists. This is the world where when given a problem, you would spend maybe 3-4 hours scoping it out, doing some research, maybe drawing up some diagrams and then jumping into the code. Coding itself was a necessary part of this process, where it was meditative. You were in this zone between deep and shallow work. It's odd to "reminisce" about coding as if it was such a beautiful thing, in reality it was hard. Required a lot of elbow grease, but it was fulfilling (I wrote 10k LoC over 2 weeks, it isn't about LoC as a metric of good code, it was more about the effort put in felt really tangible)

The world today, is fundamentally different. Today coding is really cheap, even disregarding the best SOTA models, OSS models are catching up in how consistent of an output they produce. Note the word "consistent", I won't say the code wirten by Opus or Sol is "good", but it's consistent enough to where with the right infrastructure around it (aka harnesses) you can get it to work really well.

2 years back, saying "LLMs don't produce good code" was objectively true. They were good for maybe a function, heck I remember AI straight up producing incorrect logic. Today with the right context, the "code quality" (interestignly, quality is a human metric, maintainability would be more apt, and the cost of building a change) is upto par or even better than what I would write on the first iteration.

Given that the cost of coding is now 0, my job becomes to solve problems, in the best way possible, WHILE directing the HLD/LLD of the system to understand the limitations of what I have built. The latter, I only do because it's much faster knwoing a limitation in a codebase than having an agent "recongize" it through a bunch of tests, and loops (it's cheaper too) -- THAT is where my experience really shines, knowing which patterns to apply to what problem and expect what results. Heck, we live in a time, where I can trivially have an LLM run experiments to quantify 3 other approaches at the same time.

Fundamentally what I do, is managing context, scoping problems & requirements, understanding limitations and directing the codebase (again, also drawing from experience, since an LLM may arbitiarily do something which leads to increasing entropy in the codebase)

Think of it like this; If software engineer, was like maintaing a garden, then 5 years back the process was -- think about how you want your garden to look, with this in mind begin planting the right seeds, pruning the bushes, weeding the overgrowth. Today, I have robot to do all that stuff for me, that leaves me a WHOLE lot of time to plan out my garden, thus, today, I can maintain an entire ecosystem instead of a garden. Not the best analogy, but it models my fear of the LLM not "pruning it right" when in the grand scheme of things, it may not matter at all.

Whew, that was long. I just wanted to get my thoughts out haha