I’m a 46-year-old programmer from Mexico. I’ve been doing this work for a little more than twenty-six years now, and I don’t want to stop. Writing ...
For further actions, you may consider blocking this person and/or reporting abuse
This is such a honest take on the current state of tech. I can tell you the "AI noise" feels completely deafening. One day it’s a new agentic framework, the next it’s a model that supposedly makes everything we learned last week redundant. And now it is either you catch up or you get left behind. With the current state of how quickly technology is growing, I think AI might actually come to stay. It is sad and frustrating, and I cant imagine how it must be for you who has been in the industry for that long
Thank you, but to be fair, it's always been like this; we just move faster nowadays. If it's not a new JavaScript framework, it is a new tool, a new packer, or even a new language every other month. You get used to it. But AI is getting complicated to keep up with, let alone find our place in it.
Rightly said Ed
Man, your two-screen experiment is actually the setup I landed on too. And I think you accidentally discovered the sweet spot that a lot of people are still searching for.
The "I feel like I'm cheating" thing resonates. I've got about 8 years in and I still get that twinge when an agent spits out something that works on the first try. But here's what shifted it for me: I stopped thinking of it as "the AI wrote my code" and started thinking of it as "I reviewed and approved this code." The skill didn't go away — it just moved from writing to evaluating.
Your 26 years of pattern recognition is exactly what makes you good at prompting, even if it doesn't feel that way yet. You know what good code looks like. You know when something smells off. A junior dev using Claude Code can't spot the subtle issues you'd catch in 2 seconds.
Also, don't apologize for the post being sad. This is the most honest thing I've read on dev.to in months. More people feel this way than are willing to admit it.
Thank you, Kai. You gifted me with an awesome point of view. I'll keep experimenting and refining this new way of work. After all, this might be what I was looking for years ago: a way to build fast, better software.
I'll keep in mind what you said: I'm reviewing the generated code and finding a way to meet my clients' needs. But also, I don't have to stop writing code if that is what I love. There's something magical in thinking about code and writing those thoughts on a keyboard.
Wow what chutzpah to write a freaking AI comment under this post. And people are liking this shi. Or are they even people? Maybe the dead internet theory really is true.
I was reading about this a couple of days ago. I don't know if the dead Internet theory is true, but it is worrisome.
Same boat man. 20+ years doing web dev full stack [actual full stack, you know, dns, servers, email, along with web code and databases etc]
And yea the AI stuff is overwhelming. I'm using various tools and mostly using them alongside my editors, but not embedded as i just don't trust em yet to not screw up something massively as i have dozens of projects on my machine at a given time.
I have no idea where it's going, but I know we all have to be using them and learning to use them, just so we learn what to do when things go wrong.
And i agree... i love playing 5D chess too... and we still can... for fun, but most work stuff we probably need integrate AI since companies/clients/bosses are going to be expecting it even though they don't understand how any of it works.
Keep your manual backups and recovery plans handy :) and ask for a raise every time you have to use them :D
That's a beautiful way of seeing things. I'll keep in mind that raise trick :)
It is an whole other mindset using agents to code. You need to rethink your old way of working. That is more impactful than for example leaning a new language.
It is not going to be easy and it will have teething problems. But when you find the way that works for you it will be a benefit.
I'm sad to hear things aren't going well. I do suggest hiring someone to find out to work in an agent-people team. The speed of AI is going to cause burnouts faster because the hype is all about one person companies. Having people you can rely on is better than the fastest smartest model.
Thank you, David. It is true that having a team you can rely on is priceless.
Hello Ed, I really share the same feeling you described. Rather than pushing against the AI wave, I believe we should learn and grow with it.
Would you be okay if I translate your article and share it on cafeincode.com (a Vietnamese engineering community) with full credits and a link to the original post?
Thanks for the great write-up!
Thank you, and I think you're right. We should always be improving ourselves. That's the only way we can achieve marvelous things.
It's ok on the publication, but make sure dev.to is the canonical post so we don't affect this site.
Your post is incredibly honest. After 26 years, you aren't just "typing". you’re a craftsman. It makes total sense that outsourcing your "5D chess" to an agent feels empty.
You aren't "getting old", you’re a professional who cares about the soul of his work. Keep writing the code you dreamed of, even if you let the AI handle the "grunt work" on the side.
Thank you for your kind words. You made me realize that maybe we're starting to learn how to play chess in 6D. I can see the possibilities now.
It might be a new form of impostor syndrome magnified by AI?
It is not easy to tell the AI what you want. The more specific, the less room for the AI to do "whatever". At the end you are just coding in natural language.
Yes, AI is getting better and better "guessing" what you want, but my feeling is that a lot of times the developer is (and still will be) the one who can really find "what the customer wants".
Anyway, I started my career few years ago, and I feel the same as you.
I send you saludos cordiales!
Thank you, Gabriel. You might be on to something with that impostor syndrome hypothesis. In my case, I think it was more of a conflict about what I'm supposed to be doing or how I should feel about it. Reading through the comments on this post, like yours, made me realize that with these new transformers, there's also more to it than meets the eye.
Te mando un abrazo fuerte y saludos cordiales.
You're not alone ...
Thank you, Johnny.
First of all, thank you for writing this so honestly. As a fresher developer, I can’t fully relate to having decades of experience before AI arrived, but I deeply relate to that feeling of disconnection you described. When I code myself, I feel immersed in the problem, like I’m shaping the system directly. When an agent writes most of it, the satisfaction feels different, more like reviewing than creating.
At the same time, reading your two-screen experiment made me realize something important: maybe the role isn’t disappearing, it’s shifting. Your experience still defines the direction, the architecture, and the judgment. The AI just accelerates execution. It feels less like replacement and more like leverage, but the emotional adjustment is real.
It’s genuinely inspiring to see someone with your level of experience still adapting and exploring instead of resisting. That mindset alone says a lot about what kind of engineer you are!
Thank you, Aryan. Truth is, being able to adapt and evolve is a core part of this profession. I've seen a lot of friends get left behind, and I get it: change is difficult. You are right, these technologies should be used as leverage to build bigger and more important things. I see that now.
I hope you find your way through this technology shift and that it helps you become a better developer.
Hi Ed, You are not alone to have such feelings but you are way ahead of me. I fell in love with coding 45 years ago and have enjoyed a successful career of over 30 years as a result. I cannot bring myself to use any AI tools and accept that makes me a bit of a Luddite. I am a few more years away from retirement than I am comfortable with but fear my career in software development is drawing to a close.
I work for a large multi-national company that is pushing AI tools on developers hard. Fortunately in a way, I have not been assigned a coding task of more than a few days in the last years, otherwise I might have felt more pressed.
I have seen many changes in the industry and rode with them but software development will never be the same again.
I "doff my hat" to you and wish you the best for the rest of your journey.
Thank you, my friend. Forty-five years is a long road. You don't get that far without being a resilient, wonderful developer. I truly hope the future holds great things for you. Reading through some comments on this post, I came to the idea that AI is not going anywhere but forward. Maybe we should move in the same direction. We might find greener fields behind that mountain.
Man, I feel you. I’m 41, so I’m right there in that xennial zone too, and I’ve had those same moments of “how the hell do I keep up with this pace?” AI is generating code like a firehose, and it’s easy to feel like we’re getting pushed out of our own field.
But here’s the crazy part: in the last three months, I’ve generated more code than in my entire life as a developer. And that’s not because I suddenly became a 10x engineer — it’s because I stopped trying to fight the wave and started riding it. I’ve been building a whole portfolio of systems, tools, and experiments, and I’m still figuring out how to integrate all of it into something profitable.
And that word — integration — is where everything clicked for me.
I come from a background of implementation work: electronics, C programming, hands‑on problem solving, even upholstery crafting. I’ve lived through multiple tech transitions already. Every time, the value shifted away from “how much you can manually produce” and toward “how well you can connect things, design systems, and make them work together.”
That’s exactly what’s happening now. AI isn’t replacing us — it’s changing what our job is. I let the models generate the repetitive stuff, and I focus on architecture, decisions, creativity, and the glue that holds everything together.
Watching videos about our generation helped me realize something important: xennials have already survived the analog‑to‑digital jump, the internet explosion, the mobile revolution, and the cloud era. We’ve always been the bridge generation. We know how to adapt because we’ve been adapting since we were kids.
So yeah, the pace is insane. But the real challenge isn’t keeping up with AI’s code output — it’s figuring out how to monetize our experience, our judgment, and our ability to integrate all these new tools into something meaningful. And honestly, that’s where we shine. We’ve done it before. We’ll do it again.
Thank you, Cristian. You gifted me some gems I need to wrap my mind around. I almost forgot that thing about the mythical 10X developer. Maybe we are all just becoming one using these new augmented abilities.
You are right about something important: there's a key factor in this software-building world: monetization. As employees, it's hard to see this, but as freelance or independent developers, our time is currency. I should be focusing on how these AI agents can help me build more and earn more. Build better software and reclaim some time to do the things we love.
I get that monetization is important, but personally I don’t really see the upside yet. Right now AI mostly eats my time: I have to read, test, and adapt to new tools, and I don’t see who is going to pay for the extra 10x content or code being produced. It feels more like overload than real value so far.
Hello, Karlis. As a freelancer, I was exploring the idea that if AI agents could let me handle more projects in the same time frame, I could earn more money by charging for more work done. But this is only in the assumption that AI agents help me produce 10X solutions for my customers, and that I have enough customer projects lined up.
I was reading another article on the site and realized that it might not be as easy as it sounds, as IA agents have a way of consuming tokens that is not always predictable. You know how many tokens they consume after the job is done, so transferring the cost of said tokens to the client can be a little tricky.
I really appreciate your message, man. What you said about augmented abilities hit me hard, because that’s exactly how this whole shift feels from my side too. I’m 41, and in the last three months I’ve written more code than in my entire life. Not because I suddenly became some mythical 10x dev, but because I finally accepted that my real value is in how I integrate things — how I connect systems, shape workflows, and give direction to these new tools.
Something personal I haven’t shared yet: I’ve never actually worked online. My whole career has been physical implementation work — electronics, C programming, hands‑on problem solving, even upholstery crafting. But lately I’ve been designing and building systems every night, trying to understand how all these pieces can fit together into something meaningful, useful, and eventually profitable.
That’s why your point about monetization resonated so much. When you’re independent, your time really is currency. And I’m starting to see that AI isn’t here to replace that — it’s here to multiply it. The real challenge now is figuring out where our experience and these new tools meet in a way that creates real value.
I’m also at a moment where I want to start connecting with people, building a network, and exploring where these ideas can land — whether that’s projects, collaborations, startups, or just conversations that help refine direction. Nothing pushy, just keeping the window open. I’m still figuring out the applicability of everything I’m building, and exchanges like this help more than you know.
Thanks again for the dialogue. It’s rare to find someone who’s thinking about this transition with the same mix of curiosity and honesty. It makes the whole process feel a little less solitary.
see i am a beginner , i just entered the developer world when the ai boom occured , i have a brother who is in industry for 3-4 years , so i compare his college life with mine. it's not same ,besides i think current state of the community is quite unstable. Hackathons feel useless to me , i am not saying that people who win are cheating , but i think hackathons don't feel like hackathons i though they what they would be like, the barrier between a person who doesn't know required skills and a person knows it is becoming minimal day by day. you either make your project using agents while getting frustrated and sitting for hours or you make your project using agents knowing what to tell them. There's nothing like if someone doesn't know one tech stack ,he needs to devote a significant amount. evenif you try , you get overwhelmed , in my case prompting agents has gotten so accessible , its like an involuntary action ,whenever i get stuck , copilot suddenly opens.
also , according to me the , Ai is being marketed as an alternative to developers rather than it being marketed as a tool . there;s like so much politics involved and so much chaos.
everyone wants to just dominate this industry.
every company developing its agent framework.
still no one knows which framework to learn.
every day new techstacks comming , but no one using it . for example google has developed a2a protocol for agents , but i dont find anyone talking about it or using it.
this post may feel vague and written by a total beginner , but that's true. it what i feel about the current state
I was asking myself the same questions. That's why I wrote that entry on the site. I was feeling overwhelmed and even confused by the way things are turning around AI. But I've been reading the comments and learned two things: 1) we are all in the same boat, and 2) using AI can have a very positive impact on our work. It all depends on how we see and apply things.
Right now, AI companies are burning cash at an alarming rate. They're consuming a lot of energy in pursuing what they say is the future of software development, and companies around the world are buying it. There's a new invention around the corner, new tools, new rules, new everything. That's why there's so much chaos around the topic, and it's getting complicated to keep track of all that's happening.
Thanks for sharing your honest take here, and congratulations on your first article!
I personally think AI is lowering the barrier to entry for people to start building tools which I think is fundamentally a beautiful development. This in turn generates more software and increasingly sloppy software that would require experienced engineers to fix and maintain.
The nature and role of a software engineer is likely to shift again, as it has in the past. We need to adapt and learn the new technologies. I see that you are experimenting already on your end, that is the right mindset to have in my opinion.
Thank you, Julien. In the past, learning to code was really hard. We didn't have the connection, the toolset, and the shared code base we have today. Plus, AI agents make it too easy for everybody to start building software without any previous knowledge or experience. The problem is that AI also makes it difficult to get beginner developers hired for their first job. I think you are right, our role as software developers is shifting again, and we are in the middle of that transition.
I can feel this so much! You are definitely not alone in this. Most of all I feel like alienated from the code I "write" or let write. I came up with a different approach. After more than year of using claude code, chat gpt or co pilot exhaustingly I have just deleted everything and canceled subscriptions.
I will test, if I get back the joy of coding. And I know it is a radically solution. But for me personally it was inevitable. I am going to write an article about it next month after a month of not using AI for code (or anything else).
My two main reasons are: I lost the feeling of loving to write code and I am not learning. I have the feeling that I get more stupid with every prompt.
I'm looking forward to that article. You know, one thing I've seen plenty of times is that things tend to follow a circular pattern: popular today, forgotten next week, rediscovered next month, popular again, and so on. Maybe we're bound to love-hate our relationship with AI until we find out how it fits in with ourselves.
"I want it to work, but the thing is, it doesn’t make me feel a thing.
When I’m coding, I feel like I’m playing 5D chess. When I use these agents, I feel dumb. I feel like I’m cheating."
I'm a no-coder but can definitely relate to this. I love working with AI to brainstorm or challenge ideas, and draft or review non-creative content but not to build things.
It's not exactly that I feel dumb or feel like I'm cheating, it's more that I feel hollow and lose the joy of starting from a blank page and creating something out of nothing.
But I think that's normal. It's a bit like a musician using Suno to create a song: sure, the result is impressive and often way better than anything I could have produced myself! But the point of creating music goes beyond the end result. If you’re only prompting an AI to “make it groovy” or “write funny lyrics,” the output might be cool, but it doesn’t really feel like you.
I guess a big part of excelling at something is struggling to get there and feeling pride in what you achieved despite the challenges. If you remove the struggle and creativity, it becomes "just a job".
Thank you, Joyce. I think you are right. Instead of fighting the AI agents or feeling ashamed for using them, I should find a way to make the process my own. I could find a better me.
I can totally relate to feeling a bit lost with all the AI tools out there — it’s definitely a lot to navigate. Personally, I’ve found that using paid versions of tools like GitHub Copilot and Claude for coding has made a big difference in quality and reliability.
Also, Perplexity Pro has been a standout for me when it comes to research and getting deeper, clearer insights on topics I’m exploring. It really helps connect the dots when other tools feel too shallow or generic. Definitely worth checking out if you’re trying to make sense of complex AI ideas.
Thank you, Karlis. I haven't heard about Perplexity until now. I'll take a look into it.
Paying for proper tools is important and maybe unavoidable. Someone told me once: you can't make money without spending money, and so far it's been true.
I recall when I left my job 14 years ago and became a freelancer: one of my first investments was in a good laptop. It was expensive, but it was worth it. The second thing I invested in was Telerik Reports. It was/is an expensive tool to build reports with .NET, but I did the same work in a couple of hours that took me eight hours with the previous library. So you might be right, investing in a good AI tool is worth it.
You can try to spend 20 USD one month and You will see the difference.
Imagine not being a senior and trying to learn how to create solutions and now not only people don't want to hire you because AI can do it faster, but you are trying to learn and the code is already done in front of you. I try turning off the tools most of the time so I can do stuff myself, but I can't always do that for every client who wants something done within the hour. I'm evolving and building, but everytime claude starts spitting code I get a massive angst inside where I feel like there is no point. The thing is at least I can learn how to be a good software engineer by being an ARCHITECT of solutions and for that I need to know how to do the code logic, which I will (nothing will stop me. It doesnt matter). Still, I have high hopes that I will get to the point where I'm thinking and creating beyond the AI generated code, where I can get the basics and build amazing things. That seems to be the point of the AI right? Until then, I can still cheat at the basics, but as a designer and problem solver I want to be a good professional, so I will work on knowing better than the AI, correcting it and teaching it.
I recall feeling the same way when I started writing apps twenty-plus years ago. Please don't laugh, but back then, the Internet was not what it is today. In fact, it was a tool that few people had because it was expensive. The point is, I wanted to write apps, but I had limited resources. I felt I would never get very far at the rate I was learning to code. But the thing is, you can do whatever you put your mind to. Just keep moving forward. You are going to make it.
Thank you for your post. Like Kai said:
I hope that there will be more focus on the human aspect. Not just how can we have the AI generate as much code as possible, but also how can we make sure the humans that guide / manage them feel fulfilled.
That is a wonderful way of seeing things. Maybe we're at a point where AI is burning so much cash that these companies are inventing stories about how AI can replace programmers to lower costs, because they want to sell more subscriptions. But when the dust settles, we'll see improvements in how these AI agents enhance our lives as programmers.
I feel like I am connected to the mind of the universe when I am cooperating with AI in code.😀
I never thought about it like this before. In a way, you might be right. If we use these AI agents to improve ourselves rather than just accepting and using whatever they spit out, we are connected to something bigger.
I love your analogy about AI agents feeling like junior programmers. It is a very healthy way to look at it. If you had a real junior developer helping you, you would still feel like a 'real' coder because you are the one guiding them and reviewing their work. Maybe the trick is to treat AI the same way. Use it for the tedious parts to save your business, but keep those '5D chess' moments for yourself on the parts of the project you truly love. It does not have to be all or nothing. Finding that balance might be the key to not feeling lost.
You are right, it doesn't have to be one way or the other. When I wrote that article, I truly felt lost about how I was supposed to handle these AI agents, or whether I was going the opposite way I was supposed to. I was truly lost on the matter, but reading through comments like yours made me realize that I can explore my options and find a way that works for me.
As you said, I can have a new AI junior pal working with me, handling the boring, repetitive tasks on the projects, and I could reserve these 5D chess moments for myself.
Thank you, Ava.
You're not alone.
I've doing this from 20+ years too, I'm front-end developer (more towards the css side of things, UI engineer, css architecture, etc)
For me the hard thing is that when I complete a task I feel detached from the result, because I didn't type it myself, yes I 've planned, prompted, guided... but I didn't "build it myself"... maybe is just a matter of time and get used to it, but it feels odd...
When a bug arise from a thing I built with AI it cost me much more figure it out why it fails, because I have no "muscle memory" of the code...
So you maybe need ask AI why it fails, like a worm hole you get kind of dependent ... So I try to write part of the code at least to feel better, too feel that the output code is somehow "mine"
Weird times...
That is exactly how I feel when I submit a change on a customer's project. It's like you passed the job off to someone else and just tested and committed the change. Weird times indeed. But at the end, you need to understand what happened and how the solution works, so you can fix any problems that may arise. Otherwise, you'll depend 100% on agents to write, debug, and fix the code. Your wallet might have a hard time ahead.
Thank you, Rodrigo.
Nice read. The future is pregnant!
Thank you. I hope the future brings you amazing opportunities.
This is exactly how I feel, also.
We shouldn't be afraid to admit that these tools suck for those who love programming.
And that the colleagues who love these "tools" are shit.
Many are afraid to admit it, but I think it's a perfectly normal attitude for those who love this job and realize that these "tools" are meant to replace us.
On the other hand, programmers are notorious for being miserable and will shoot themselves in the foot so as not to appear "inefficient" and to kiss the ass of their bosses who treat them like expendable animals.
I hate the programmers who work for these people who tell you exactly how many months are left until you're useless.
Personally, I only use copilot on my code; it seems like the best compromise.
I feel you and in sort of the same boat. Going on 44 this year and also been developing for the last 20+ years. Using Claude Code professionally, mostly as a reviewer of my written code and sometimes for brainstorming. But I mis the excitement of a piece of code that works by my own hand.
Owell, time goes on, I guess..
With the many videos and articles on the internet about AI coding agents, you begin to believe that you can get an autonomous AI agent to do all your coding projects with minimal supervision but in reality, it is not exactly like that.
Ed, you echo my sentiments exactly. I began my web development journey in 1998. This was a magical time for the internet for me. I switched careers in 2009 simply because I had a great opportunity come up. In 2021 I wanted to return to web development, since then I have been playing catch up and now with A.I - I feel the same as you. Completely lost. The fun factor for me has left the building. Things are moving at such fantastic speeds that I am completely overwhelmed and cannot help but question my purpose here. Lately I have returned to my programming roots, C++. I have the luxury of being able to fall back on a previous career so being productive at the moment is not imperative. My take now is to sit back and see what happens before I decide to jump back in the boat. At my age 52, returning to a coding career is unlikely but I am not ruling it out. Anyways, thank you for sharing - your not alone here buddy.
This is a very honest post, and I can absolutely relate to this - on the other hand, I would say let's not go crazy, at this point it's good to be a little bit skeptical about the whole AI craze ...
One thing I'm not hearing mentioned often, but which is real, is that this stuff simply costs money (of course it's a business model for AI companies) - you can easily burn through a lot of tokens, unless you're already a bit of an AI "expert" - and if the results aren't up to the same quality standards compared to when we'd have coded it ourselves, then what have we gained?
Bit of an eye opener: in reality, on many projects the actual coding might just take 15 or 20 percent of the time - the rest is spent on planning, design, testing, debugging and so on ...
So now the focus seems to be on those 15-20% - I'd argue for shifting our focus to the 80-85% :-)
Having AI help with the thinking part (as an advisor) may have several advantages - one if that you don't need to review/fix the code that AI generates, if it doesn't generate that much code ...
It can still be used to generate some of the code, but the "boring" part, the pure boilerplate code - if used like that, it might free up the developer to focus more on the "interesting" work ...
Well that's the more optimistic take, but basically my point is:
Let's not go crazy and jump into having AI generate complete codebases full of crap and security holes, and I'd say the same to companies who think they can go for a quick buck and fire all their juniors and replace them with AI - look before you leap, the benefits might not outweigh the costs when you make the complete calculation.
P.S. went off on a bit of a tangent here, but I hugely appreciate the honesty of your post - I think this is relatable for many people! Let's have a healthy dose of "well reasoned skepticism" as an antidote ...
i should change my name to hellsurfer
That's a hell of a name!
+1