Tools like Claude Code are the ultimate cheat code for me and have breathed new life into my desire to create. I know more than enough about architecture and coding to understand the plumbing and effectively debug, yet I don't have to know or care about implementation details. It's almost an unfair unlock.
It'll also be good to see leetcode die.
My experience is that people who weren't very good at writing software are the ones now "most excited" to "create" with a LLM.
Fast forward 30 years later, I thought those days were gone forever. I'd accepted that I'd never experienced that kind of obsession again. Maybe because I got older. Maybe those feelings were something exclusively for the young. Maybe because my energy wasn't what it used to be. Yada yada, 1000s of reasons.
I was so shocked when I found out that I could experience that feeling again with Claude Code and Codex. I guess it was like experiencing your first love all over again? I slept late, I woke up early, I couldn't wait to go back to my Codex and Claude. It was to the point I created an orchestrator agent so I could continue chatting with my containerized agents via Telegram.
"What a time to be alive" <-- a trite, meaningless saying, that was infused by real meaning, by some basic maths that run really, really, really fast, on really, really expensive hardware. How about that!
With experience, you see these dead ends before they have a chance to take hold and you know when and how to adjust course. It's literally like one poster said: coding with some buddies without ego and without the need to constantly talk people out of using the latest and greatest shiny objects/tools/frameworks.
I've really enjoyed going back a revisiting old ideas and projects with the help of AI. As the OP stated -- it has restored my energy and drive.
I’m probably going to go back and redo everything with my own code.
“Peter Steinberger is a great example of how AI is catnip very specifically for middle-aged tech guys. they spend their 20s and 30s writing code, burn out or do management stuff for a decade, then come back in their late 40s/50s and want to try to throw that fastball again. Claude Code makes them feel like they still got it.”
Been working for about a month, and I’m halfway through. The server’s done (but I’m sure that I’ll still need to tweak and fix bugs), and I’m developing the communication layer and client model, now. It took seven months to write the first version of the server, and about six months to write a less-capable communication driver, the first time.
This is not a “vibe-coded” toy for personal use. It’s a high-Quality shipping app, with thousands of users. There’s still a ton of work, ahead, but it looks like an achievable goal. I do feel as if my experience, writing shipping software, is crucial to using the LLM to develop something that can be shipped.
I’ve had to learn how to work with an LLM, but I think I’ve found my stride. I certainly could not do this, without an LLM.
The thing that most upset me, since retirement, has been the lack of folks willing to work with me. I spent my entire career, working in teams, and being forced to work alone, reduced my scope. I feel as if LLMs have allowed me to dream big, again.
>"Wait a moment! Being forced to use AI gave me depression, and I'm really aware of the fact that it's only going to become better and better the more developers are using it, to the point where the 10 job openings of yesterday are 1 job opening tomorrow. Why are people so excited", remember this:
You are reading HN, the survivorship bias and groupthink is just as high as any other self-calibrating online community ("upvote if you agree" -> self-calibration of the popular opinion), and there's an extremely high survivorship bias because people who are into this LLM craze have a higher probability of browsing HN.
As for you, OP, I have no idea why age is a factor to consider to this. I'm 45, and while I programmed as a hobby since I was 16 I turned it into a career during COVID, and all the pressure cooking LLM watch-six-agents-writing-and-you-proofreading gave me so much existential crisis and depression that I seriously can't even get myself writing anything "over the weekend".
I hope to God the next generation of wonder kids that is the equivalent of the 12 year old discovering how to bent the computer to do what they want it to do enjoy arguing with multiple agents concurrently back and forth.
So excited to be getting to my backlog of apps that I've wanted but couldn't take the time to develop on my own. I'm 66 and have been in the software field in various capacities (but programming mostly as a hobby). Here's a partial list of apps I've completed in the last few months:
- Media Watch app to keep a list of movies and shows my wife and I want to watch- Grocery List with some tracking of frequent purchases
- Health Log for medical history, doc appointments and past visits
- Habits Tracker with trends I’m interested
- Daily Wisdom Reader instead of having multiple ebooks to keep track of where I'm at
- A task manager similar to the old LifeBalance app
- A Home Inventory app so that I can track what I have, warranty, and maintenance
- An ios watch app to see when I'm asleep so that it can turn off my music or audiobook
- An ios watch chess tactics trainer app
- some games
Many of these are similar to paid offerings, but those didn't check off all the features I really wanted, so I vibe-coded my own. They all do what I want, the way I want it to.
Staying up and re-learning what I used to love long ago has given me a new found passion as well. Even if I do vibe code some scripts, at least I have the background now to go through them and make sure they make sense. They're things I'm using in my own homelab and not something that I'm trying to spin up a Github repo for. I'm not shipping anything. I'm refreshing my old skills and trying to bring some of them up to date. An unfortunate reality is that my healthcare career is going to be limited due to multiple injuries along the way, and I need to try to be as current as I can in case something happens. My safety net is limited.
The teams get reduced, as now one can do effectively more with less, and in South Europe, in IT there is hardly a place to get a job above 50 years old, unless one goes consulting as the company owner, and even then the market cannot hold everyone.
As kid I have seen this happening, as factory automation replaced jobs in complete villages, the jobs that weren't offshored into Asia or Eastern Europe for clothing and shoes, got replaced with robots.
The few lucky ones were the ones pressing the buttons and unloading trucks.
Likewise a few ones will be lucky AI magicians, some will press buttons, and the large majority better get newer skills beyond computing.
The note read something like as follows : I don’t exactly agree with the framing that we will all get left behind if we all don’t learn to adapt to AI. More accurately, I see it this way. While the company definitely stands to gain from all the hyper increase in productivity by the use of said AI tools, I stand to pay a personal price and that personal price is this - I’m may very slowly stop exercising my critical thinking muscles because I am accustomed passing that to AI for everything and this will render me less employable, it is this personal price that I feel reluctant to pay. There has always been a delicate balance between an employer and employee. We learn new technologies on the job and we’re more employable for transferring that to other companies. This equation is now unbalanced. The company trapped more value, but there is skill erosion on my side . For instance, our team actually has to perform a Cassandra DB migration this year . Usually, I’d have to take a small textbook and read about the internals of CassandraDB, and maybe learn a guide on how to write Cassandra queries. What do I put in my resume now? That I vibe coded Cassandra migration? How employable is that? And I’m not sure if others felt the same way. But I definitely felt like the odd one out here for asking that question because everyone else in the meeting was on board with AI adoption.
The leader did respond to me and he said that learning agentic AI actually will make me more employable. So there is a fundamental disagreement as to what constitutes skill. I think he just spoke past me. Oh well at least I tried.
Was able to build a large financial application just with the 20 USD subscription in the last 12 month - without Claude, I would have required 5 - 6 people and at least 1 year of funding.
This was by far my best investment in my whole life 12x20 USD vs. 750.000 salary :-)
It is especially inspiring since it brings you usually a few new ideas into your context; also just joking around with it can yield new inspirations.
I'm wondering how long it will stay at 20 USD for the smallest subscription, no chance that they can keep this price, I'd say? Its impressive that they are giving it away for nearly free.
Until I realized that no one here is going to be in the blast radius. So many people who agree with this admit to being in their 40s, 50s, 60s. All of them have already had the time to learn without LLMs, get industry experience, network, climb their career ladders as high as they could. These people are now sitting on piles of assets, and they know that if LLMs start pushing out people from the industry, it'll be us juniors and new grads. They will either remain relevant in the industry due to seniority/experience/pivoting to managerial duty, use their money and connections to easily learn new skills and pivot, or punch out and coast through retirement before it affects them.
While I have never developed software professionally, in the four decades I have been using computers I have often written scripts and done other simple programming for my own purposes. When I was in my thirties and forties especially, I would often get enjoyably immersed in my little projects.
These days, I am feeling a new rush of drive and energy using Claude Code. At first, though, the feeling would come and go. I would come up with fun projects (in-browser synthesizers, multi-LLM translation engines) and get a brief thrill from being able to create them so quickly, but the fever would fade after a while. I started paying for the Max plan last June, but there were weeks at a time when I barely used it. I was thinking of downgrading to Pro when Opus 4.5 came along, I saw that it could handle more sophisticated tasks, and I got an idea for a big project that I wanted to do.
I have now spent the last two months having Claude write and build something I really wanted forty years ago, when I was learning Japanese and starting out as a Japanese-to-English translator: a dictionary that explains the meanings, nuances, and usages of Japanese words in English in a way accessible to an intermediate or advanced learner. Here is where it stands now:
https://github.com/tkgally/je-dict-1
It will take a few more months before the dictionary is more or less finished, but it has already reached a stage where it should be useful for some learners. I am releasing all of the content into the public domain, so people can use and adapt it however they like.
What changed for me was the feedback loop. Before AI tooling, I'd have an idea, realize it would take weeks to prototype, and let it die. Now I go from concept to working MVP in a weekend. The constraint shifted from "can I build this" to "should I build this" - which is a much better problem to have.
The stack that works for me: Lovable for frontend, Replit for backend, Claude API for the AI layer, Neon for Postgres. Not fancy, but it ships.
The biggest lesson: AI doesn't replace the need for experience and taste. It amplifies it. Your decades of context about what makes good software - that's the real asset. Claude is just fast hands.
I have a sense that AI could have something to do with it.
AI is degrading the status of our profession; its perception in the public eye.
At the same time, it is stealing our work and letting cretins pretend to be software engineers.
It's a bad taste in the mouth.
Of course you love it, you don't have to worry about retirement anymore.
Give me your 401k, then tell you feel about Claude Code.
* Implementing a raw Git reader is daunting.
* Codifying syntax highlighting rules is laborious.
* Developing a nice UI/UX is not super enjoyable for me.
* Hardening with latest security measures would be tricky.
* Crafting a templating language is time-consuming.
Being able to orchestrate and design the high-level architecture while letting the LLM take care of the details is extremely rewarding. Moving all my repositories away from GitLab, GitHub, and BitBucket to a single repo under my own control is priceless.
"in (language I'm familiar with) I use (some pattern or whatever) what's the equivalent in (other language)?"
It's really great for doing bits and then get it to explain or you look and see what's wrong and modify it and learn.
In a weird way it’s making software development feel more like engineering again rather than constant framework churn.
I love coding with agents. Claude Code now almost exclusively. The 20x max subscription is endless until you start writing custom multi-agent processes, and even then. Still takes quite a bit of effort to burn through.
I get so much more done, and can be productive with languages/frameworks I'm not familiar with.
To everybody worried that AI will kill jobs. There have been many points in the evolution of software dev where some new efficiency was predicted to kill off jobs. The opposite happens. Dev becomes more economical, and all of the places where dev was previously too expensive open up. Maybe this time won't work out that way, but history isn't on the side of that prediction.
An experienced software dev can get multiples of efficiency out of AI coding tools compared to non-devs, and can use them in scaled projects, where non-devs are only going to compound a mess. Some of those non-devs will learn how to be more efficient and work with scaled projects. How? They'll learn to be devs.
I'd be building several side projects for myself if I wasn't super busy with the primary work I'm doing. The AI tools take over the tedious work, and remove a lot of work that would just add mental load. Love it.
Let’s get you to bed, gramps, you can talk to your French friend tomorrow.
I’m a programmer for life. My hobbies revolve around programming and hardware as well (demos for retro hardware: XTs up to Pentium, Sega Master System, custom built hardware). I stay up late working on this stuff and I still have the drive to do it with two young kids who take up a lot of my time.
I have zero interest in an AI doing any of it for me. I don’t think I’ll be replaced by an AI but I might be forced to use one by an employer at which point I think I’ll retire and just work on my hobby projects!
So I decided that I wanted web apps, something that is probably beyond me in any reasonable time, if at all, if I was to code myself by hand.
For my coding AI "stack" I am now running OpenClaw sitting on top of Claude Code, I find the OpenClaw can prompt Claude Code better and keep it running for me without it stopping for stupid questions. Plus I have connected OpenClaw to my Whatsapp so I can ask how it is going or give instructions to the OpenClaw while not at the keyboard.
One app was a little complex with 35,000 loc, plus libraries etc. I reckon I had spent maybe 2500 hours on it over some years, but a significant part of that was developing the algorithm/workflow that it implemented - I only knew roughly what I wanted when I started, writing several to throw away at the beginning.
AI converted it to a webapp overnight, with a two sentance prompt, without intervention of any kind.
It took me another 15 minutes and a couple of small changes, mostly dependancies issues, and I had a working version of the same app that was literally 95%+ of the original, in terms of funcitonality and use.
I have a bunch of ideas for things I want to make that I probably never would have been able to otherwise.
I am just totally unable to fathom people that just make a blanket proclamation that AI is good for nothing. I can accept that it is not good for everything, it may cause some social disruption and the energy use is questionable, but improving, but not useful? Wake up.
It's given me the guts to be a solo-founder (for now). I
My current passion is pushing small LLMs as far as I can using tools and agentic frameworks. The latest Qwen 3.5 models have me over the moon. I still like to design and code myself but I also find it pleasurable to sometimes use Claude Code and Antigravity.
The re-ignition thing resonates though. There's something about having a collaborator that removes the activation energy of starting. The blank file problem is real and brutal at 25, probably more so at 60 when you know exactly how much work lies ahead. AI doesn't eliminate the hard parts but it compresses the "ok where do I even begin" phase from hours to minutes.
What are you building?
Such a big part of coding becomes mundane after a while. Constantly solving variations of the same kinds of problems.
Now Claude does it at my direction and I get so much more done!
But maybe even more important: It gets me to go outside my comfort zone and try things I wouldn't normally try because of the time it would take me to figure it out.
Like: Wat if I used this other audio library? I don't have to figure it out, I just pass in the interface I need to implement and get 90% of a working solution.
AI augmented programming couldn't have come at a better time and I'm really happy with it!
I started out with an 8 bit micro so I really enjoy tinkering and coding. AI doesn't seem attractive at all.
It's not only about what you do, but also about how you do it.
Take anthropic for example, they have created MCP/claude code.
MCP has the good parts of how to expose an API surface and also the bad parts of keeping the implementation stuck and force workarounds instead of pushing required changes upstream or to safely fork an implementation.
Claude code is orders of magnitude inefficient than plainly asking an llm to go through an architecture implementation. The sedentary black-box loops in claude code are mind bending for anyone who wants to know how it did something.
And anthropic/openai seems to just rely of user momentum to not innovate on these fundamentals because it keeps the token usage high and as everyone knows by now a unpredictable product is more addictive than a deterministic one.
We are currently in the "Script Monkey" phase of AI dev tools. We are automating the typing, but we haven't yet automated the design. The danger is that we’re building a generation of "copy-paste" architects who can’t see the debt they’re accruing until the system collapses under its own weight.
The problems, as ever, are 1) what negative things are enabled by the technology, 2) do the positive things that are enabled by the technology outweigh those ("is the price worth paying?"), and 3) how much harm will "stupid" and/or "evil" cause as a result of the technology?
And so on.
The fact that a thing is exciting or interesting or stimulating is neat, for sure, but as always there is no relevant thought given to ramifications.
Humans lag well behind technological advancement, and this particular wave is moving faster than perhaps anything else (because prior technological advances enable it, etc).
It's cool that you enjoy it. Me, too. I might enjoy shooting heroin into my eyeballs, too, right up until I don't.
I am 43. I used to code as a kid and I've dabbled in it here and there, but I quickly realised I didn't want to code as a career, but now with these new tools I am building again and it's great, because I'm building the things that work for me.
To manage my life there was a todo app I used, now I've built my own, don't need to pay for it and it works exactly as I want it and now I also have a few ideas for other things I Want to do.
It's great, it feels like we might be able to start taking control of our tech back again now, when we can build the tools ourselves that work the way we want, we don't have to worry about the nonsense companies are sticking into there products, we can make things work exactly as we want it.
First with LOGO on the Apple ][, making the turtle move around the screen and follow your commands. It was magic.
Then discovering BASIC, and the ability to turn the pixels on and off and make them any color you like.
Making my Amiga talk with the "SAY" command.
The first time I dialed a BBS in the dead of night with my Commodore 64 and my 300 baud modem, watching those colorful letters sloowly make their way across the TV screen...
Running my own BBS software and dialing in from my cousin's house at Thanksgiving...
Putting up my own web page and cgi-bin scripts....
It's all been magic, and it's all been just for me.
So when you remove everything else, all the cruft and crap,
I will still be programming just for me.
I landed on GitHub Copilot. I now manage a team, but just last night snuck away to code some features. I find my experience and knowing how to review the output helps me adopt and know how much to prompt the agent for. Is software development changing? Absolutely. But it always has been. These tools help me get back to that first freedom I felt when I dragged a control onto a VB6 designer, but keep the benefits of code in text files. I can focus on feature, pay attention to UX detail, and pivot without taking hours.
Google's "Ask AI" and ChatGPT's free models seem to be consistently bad to the point where I've mostly stopped using them.
I've lost track of how many times it was like "yes, you're right, I've looked at the code you've linked and I see it is using a newer version than what I had access to. I've thoroughly scanned it and here's the final solution that works".
And then the solution fails because it references a flag or option that doesn't even exist. Not even in the old or new version, a complete hallucination.
It also seems like the more context it has, the worse it becomes and it starts blending in previous solutions that you explained didn't work already that are organized slightly different in the code but does the wrong thing.
This happens to me almost every time I use it. I couldn't imagine paying for these results, it would be a huge waste of money and time.
I'm now dealing with a lot of stuff via codex, including technical debt that I identified years ago but never had the time to deal with. And I'm doing new projects. I've created a few CLIs, created a websites on cloudflare in a spare half hour, landed several big features on our five year old backend and created a couple of new projects on Github. Including a few that are in languages I don't normally use. Because it's the better technical choice and my lack of skills with those languages no longer matters.
I also undertook a migration of our system from GCP to Hetzner and used codex to do the ansible automation, diagnosing all sorts of weirdness that came up during that process, and finding workarounds for that stuff. That also includes diagnosing failed builds, fixing github action automation, sshing into remote vms to diagnose issues, etc. Kind of scary to watch that happen but it definitely works. I've done stuff like this for the last 25 years or so using various technologies. I know how to do this and do it well. But there's no point in me doing this slowly by hand anymore.
All this is since the new codex desktop app came out. Before Christmas I was using the cli and web version of codex on and off. It kind of worked for small things. But with recent codex versions things started working a lot better and more reliably. I've compressed what should be well over half a year of work in a few weeks.
It's early days but as the saying goes, this is the worst and slowest its ever going to be. I still consider myself a software maker. But the whole frontend/backend/devops specialization just went out of the window. But I actually enjoy being this empowered. I hate getting bogged down in grinding away at stupid issues when I'm trying to get to the end state of having built this grand thing I have in my head. There definitely is this endorphin rush you get when stuff works. And it's cool to go from idea to working code in a few minutes.
Hoping to start blogging about some of these projects in the future.
But I have been haranguing Claude/Gemini to help me on an analog computer project for some months now that has sent me on a deep dive into op-amps and other electronics esoterica that I had previously only dabbled a bit in.
Along the way I've learned about relaxation oscillators, using PWM to multiply two voltages, integrating, voltage-following…
I could lean on electronics.stackexchange (where my Google searches often lead) but 1) I first have to know what I am even searching for and 2) even the EEs disagree on how to solve a problem (as you might expect) so I am still with no clear answer. Might as well trust a sometimes hallucinating LLM?
I guess I like the first point above the best—when the LLM just out of the blue (seemingly) suggests a PWM multiplier when I was thinking log/anti-log was the only way to multiply voltages. So I get to learn a new topology.
Or I'm focused on user-adjustable pots for setting machine voltages and the LLM suggests a chip with its own internal 2.45V reference that you can use to get specific voltages without burdening the user to dial it in, own a multimeter. So I get to learn about a chip I was unfamiliar with.
It just goes on an on.
(And, Mr. Eater, I only let the magic smoke out once so far, ha ha.)
I'm really sorry (and accept down-vote storm) to disappoint you but you won't be young again and burning midnight oil may remind old days and bring excitement, but in the end it will harm your health.
Learning like crazy, late night hacking and other attributes of fresh engineers is sometimes a necessity to build a career, knowledge base, equity to comfortably start a family. Some people enjoy it and many hate, but most of us did it at some point.
I wouldn't oppose it if it wasn't harmful for the industry. What all those engineers who are excited again would think of a startup that stole all free land, building material and doubled housing? I bet all youngsters would be excited to have their own place for $20 monthly mortgage payment, telling everyone who has paid most of their salary over last 30 years how energizing is feeling you don't need to work for your house your whole life and ignoring equity crash for those folks.
Try to tell Claude Code to refactor some code and see if it doesn't just delete the entire file and rewrite it. Sure that's cute, but it's absolutely not okay in a real software environment.
I do find this stuff great for hobbyist projects. I don't know if I'd be willing to put money on the line yet
With Claude Code specifically, I've noticed that the longer it runs autonomously, the more cost anxiety creeps in. You stop thinking about the problem and start watching the token counter.
What finally let me stop worrying and just build again was building a hard budget limit outside the app — not just alerts, but an actual kill switch.
Glad you found the spark. It's worth protecting.
Been programming off and on since I was a kid, though I went into a career of systems architect instead, because I found the actual process of churning out code kinda tedious.
But I still had all these ideas in my head that I wanted to make reality, and now I finally can.
A project that would normally take weeks, and significantly affect the rest of my life, now only takes hours.
But remember that all those projects need to be maintained too, you can't just release a bunch of new code into the open source ecosystem without maintaining it.
I was getting Claude to implement a popular TS drag and drop library, and asked it to do something that, it turns out, wasn't supported by the library.
Claude read the minified code in node_modules and npm patched the library with the feature. It worked, too.
Obviously not ideal for future proofing but completely mind blowing that it can do that.
I think you really hit the jackpot because you got a full career out of it, saw an amazing evolution etc. So you can hopefully enjoy the ride now being more as a spectator without the fear of being personally affected by job displacement. Enjoy the retirement!
Which immediately surfaces the next problem: how do those agents communicate back to you while running?
Most setups default to tailing a log file, or a Slack/Telegram bot bolted on as an afterthought. Works for one agent. Falls apart when you have five running overnight and one hits an edge case at 2am that needs a human call.
The agent-to-human communication layer is still surprisingly ad-hoc. You can generate more ideas and actually implement them now — but the infrastructure for keeping humans in the loop as agents execute is still duct tape. Feels like the next interesting problem after the coding unlock.
Walked into work Monday morning, bleary eyed and told everybody, “This is the solution. This is how you build rapidly and bypass all of the long term maintenance issues that we always have to fix in every other codebase. It makes the hard things easy, it makes perfect sense and it’s FUN.”
I highly recommend this blog post about vibe coding, gambling, and flow. Glad you're having a great time! Just something to consider.
I feel selfish in that I am towards the end of my career rather than right at the start.
I think it's also somewhat addictive. I wonder if that's part of what's at play here.
A coworker that never argues with you, is happy to do endless toil... sometimes messes up but sometimes blows your mind...
“Oh shit, Hey Babe did you close my laptop?”
My not-very-technical friend as we returned home from a Sunday afternoon trip to the park with the kids to find his Claude Code session had been thwarted.
There are definitely a lot of limitations with Claude Code, but it's fun to work through the issues, figure out Claude's behavior, and create guardrails and workarounds. I do think that a lot of the poor behavior that agents exhibit can be fixed with more guardrails and scaffolding... so I'm looking forward to the future.
Sure, AI is exciting, and it reignites a passion. But everything you learn today will be obsolete a year from now. And that might tire you out again.
I can ask an LLM for specific help with my codebase and it can explain things in context and provide actual concrete relevant examples that make sense to me.
Then I can ask again for explanations about idiomatic code patterns that aren't familiar for me.
Working on my own, I don't get that feedback and code review loop.
Working with new languages and techniques, or diving into someone else's legacy code base is no longer as daunting with an LLM to ask for help!
Following this idea, what do people think "backend" work will involve? Building and tweaking models, and the infra around them? Obviously everyone will shift more into architecture and strategy, but in terms of hands-on technical work I'm interested in where people see this going.
Claude is for old people!
Anthropic can adapt the "Tai Chi" YouTube ads, where fat retired people become muscular in just three weeks!AI haters trend towards affection for the jargon, languages, and falling down that rabbit hole. They love Ruby, web apps, SaaS... the ecosystem of syntaxes. They love their job.
Those that dig AI see code as a historically necessary tool to get a machine to do a thing. I fall in this category.
I find the syntax and made up semantics boring, and doing interesting things with the machine interesting.
Ymmv but both online and in the real world I have only encountered these two schools of thought, as they say, when AI comes up.
I am only 43, but on the last year of my career, suddenly my level of care in big corporate politics nose dived to almost zero. To the point that I happily retired myself.
After messing around with some hard subjects, with the help of Claude Code, the little boy who used to love programming so much is waking up again.
Juniors prompt "build me X" and get frustrated when it goes sideways. Seniors architect the constraints first - acceptance criteria, test harness, API boundaries - then let the AI fill in mechanical work.
The real shift: AI makes the cost of prototyping near-zero, which paradoxically makes taste and judgment MORE valuable. When you can spin up 5 approaches in a weekend, knowing which one to actually ship becomes the bottleneck.
The folks who defined their value as "typing code" will struggle. The folks who defined their value as "knowing what to build and how to verify it works" are thriving.
Occasionally I remote in to help fix something, but the coding agent really takes a load off my back, and he can start learning without knowing where the endpoints are.
in 1 year I built three Laravel Apps from the ground up and sold one for $18,900.
That's my story and I'm sticking to it! I love Claude!
"Without tubes of paint, there would have been no Impressionism." - Renoir
100% agree even with half your experience.
Sorry, this "Tell HN" is 100% a stealth advertisement and the usual bots in the comments confirm the ad.
I review absolutely everything it does and do a lot of manual iteration on its work to clean up after it, but I think it's worth it, especially because it makes it so much easier to get ideas out that otherwise don't feel worth learning to make fully manually.
luckily i'm trusting my gut that staying away from cheap dompamine and following what's cool might just land somewherere
I wish I have the same energy once I am your age !
I wrote my first computer program in 1967. Since, it's been one fascinating thing after another but, for me, the modern age had become dull. The thought of figuring out another API or framework makes me need a nap.
Now I can have an idea, negotiate with Milo (Claude Code integrated with a neo4j graph database because now I can!) and it's off to the races.
Did I learn CYPHER, the neo4j query language? Nope. Am I the master of Agent SDK? Nope. Milo is my cognitive partner. I am inspired.
Ideas I had years ago are off the back burner. More new ideas flood my brain. I am set free. It feels like love. I lay awake at night thinking of things to do.
I am so grateful that I lived to see this day and still have the intellectual flexibility to enjoy it.
Fucking wild.
Claude has made my coding sessions WAY more productive and helps me find bugs and plan features like never before.
I'm also dealing with some career bullshit, so having a tool like this has helped me re-discover what I love about computing that capitalism has beaten out of me.
I loved coding before and love it still now.
I'm with you on the liberation not just with building, but I've also learned so much and so fast with LLM's the past few years.
Kinda scary like a motor bike, too.
God speed, you! And meh the haters and pontificators.
Here's a word I learned yesterday, my gift should you chose to accept - occhiolism.
What does your dev stack look like?
I've been leveraging a lay audience (one of my teams) to deep dive requirements, wants etc.
Anyway, I'm so torn. I like these people, I hate to see them lose their jobs. I'll retire soon, I want to find a better, "feel good role" than my current, yet very lucrative situation.
I want to leverage my years of good software design for good. Where, for who?
--old lost IT guy in FL
And I hear "why am I helping you code me out of a job". I scare them with "if you help, you'll stay", assuming they get that what I really mean is "if you duck away, bury you're head in the sand, you'll be out"
I want a game that generates its own mechanics on the fly using AI. Generates itself live.
Infinite game with infinite content. Not like no mans sky where everything is painfully predictable and schematic to a fault. No. Something that generates a whole method of generating. Some kind of ultra flexible communication protocol between engine and AI generator that is trained to program that protocol.
Develop it into a framework.
Use that framework to create one game. A dwarf fortress adventure mode 2.0
I have no other desires, I have no other goals, I don’t care. I or better yet - someone else, must do it.
This is likely fake and an ad. In case it isn't, consider treatment for AI psychosis.
I took a break from software, and over the last few years, it just felt repetitive, like I was solving or attempting to solve the same kinds of problems in different ways every 6 months. The feeling of "not a for loop again", "not a tree search again", "not a singleton again". There's an exciting new framework or a language that solves a problem - you learn it - and then there are new problems with the language - and there is a new language to solve that language's problem. And it is necessary, and the engineer in me does understand the why of it, but over time, it just starts to feel insane and like an endless loop. Then you come to an agreement: "Just build something with what I know," but you know so much that you sometimes get stuck in analysis paralysis, and then a shiny new thing catches your engineer or programmer brain. And before you get maintainable traction, I would have spent a lot of time, sometimes quitting even before starting, because it was logistically too much.
Claude Code does make it feel like I am in my early twenties. (I am middle-aged, not in 60s)
I see a lot of comments wondering what is being built -
Think about it like this, and you can try it in a day.
Take an idea of yours, and better if it is yours - not somebody else's - and definitely not AI's. And scope it and ground it first. It should not be like "If I sway my wand, an apple should appear". If you have been in software for long, you would have heard those things. Don't be that vague. You have to have some clarity - "wand sway detection with computer vision", "auto order with X if you want a real apple", etc.. AI is a catalyst and an amplifier, not a cheat code. You can't tell it, "build me code where I have tariffs replacing taxes, and it generates prosperity". You can brainstorm, maybe find solutions, but you can't break math with AI without a rigorous theory. And if you force AI without your own reasoning, it will start throwing BS at you.
There is this idea in your mind, discuss it with ChatGPT, Gemini, or Claude. See the flaws in the idea - discover better ideas. Discuss suggestions for frameworks, accept or argue with AI. In a few minutes, you ask it to provide a Markdown spec. Give it to Claude Code. Start building - not perfect, just start. Focus on the output. Does it look good enough for now? Does it look usable? Does it make sense? Is the output (not code) something you wanted? That is the MVP to yourself. There's a saying - customers don't care about your code, but that doesn't mean you shouldn't. In this case, make yourself the customer first - care about the code later (which in an AI era is like maybe a 30min to an hour later)
And at this point, bring in your engineer brain. Typically, at this point, the initial friction is gone, you have code and something that is working for you in real - not just on a paper or whiteboard. Take a pause. Review, ask it to refactor - make it better or make it align with your way, ask why it made the decisions it made. I always ask AI to write unit tests extensively - most of which I do not even review. The unit tests are there just to keep it predictable when I get involved, or if I ask AI to fix something. Even if you want to remove a file from the project, don't do it yourself - acclimatize to prompting and being vague sometimes. And use git so that you can revert when AI breaks things. From idea to a working thing, within an hour, and maybe 3-4 more hours once you start reviews, refactors, and engineering stuff.
I also use it for iterative trading research. It is just an experiment for now, but it's quite interesting what it can do. I give it a custom backtesting engine to use, and then give it constraints and libraries like technical indicators and custom data indicators it can use (or you could call it skills) - I ask it to program a strategy (not just parameter optimize) - run, test, log, define the next iteration itself, repeat. And I also give it an exact time for when it should stop researching, so it does not eat up all my tokens. It just frees up so much time, where you can just watch the traffic from the window or think about a direction where you want AI to go.
I wanted to incorporate astrological features into some machine learning models. An old idea that I had, but I always got crapped out because of the mythological parts and sometimes mystical parts that didn't make sense. With AI, I could ask it to strip out those unwanted parts, explain them in a physics-first or logic-first way, and get deeper into the "why did they do this calculation", "why they reached this constant", and then AI obviously helps with the code and helps explain how it matches and how it works - helps me pin point the code and the theories. Just a few weeks ago, I implemented/ported an astronomy library in Go (github.com/anupshinde/goeph) to speed up my research - and what do I really know about astronomy! But the outputs are well verified and tested.
But, in my own examples, will I ever let AI unilaterally change the custom backtesting engine code? Never. A single mistake, a single oversight, can cost a lot of real money and wasted time in weeks or months. So the engine code is protected like a fortress. You should be very careful with AI modifying critical parts of your production systems - the bug double-counting in the ledger is not the same as a "notification not shown". I think managers who are blanket-forcing AI on their employees are soon going to realize the importance of the engineering aspect in software
Just like you don't trust just any car manufacturer or just any investment fund, you should not blindly trust the AI-generated code - otherwise, you are setting yourself up to get scammed.
Simply put, we delegate a freedom of use and cognitive power to complex tools and organizations that control and shapes them. One can argue that it's kind of the same if I decide to code any kind of programs the 'old' way, especially using native language, albeit their exist toolchains and OSes that are open source and thus technically free of monolithic take over.
Furthermore those LLMs tools seem to me like the transhumanists cybernetic enhancements of cyberpunk dystopia, splitting Humanity between those of us that would be able to afford them and the others that are left off the competitive arena. Again, an issue that were still there to some degree in a capitalist economy but where the real entry to programming was just a computer and an internet connection to some extent, a way more democratic and affordable goal than having a subscription to a Big Bad Corporation owning everything about you and your creation, where 'free' non local models are not a real answer here either.
Any new technology have some good potential, sure, it's obvious even. I don't think the path they naturally lead to are always the best we could take though, and I hope we wake up to the fact our society are nothing short of democratic* when the economical entities that govern us is nothing but.
* Well, I don't even think we could call our political systems democratic without any kind of random selection anyways. A pastiche of one at best.
If the software produced is for internal use, the point is probably moot. But if it isn't, this seems like a question that needs to be answered ASAP.
My first finished product: ZIB, a RSS Reader inspired by Innoreader, just free ;)
When it was just asking ChatGPT questions it was fine, I was having fun, I was able to unblock myself when I got non-trivial errors much quicker, and I still felt like I was learning stuff.
With Codex or Claude Code, it feels like I'm stuck LARPing as a middle manager instead of actually solving problems. Sometimes I literally just copy stuff from my assigned ticket into Claude and tell it to do that, I awkwardly wait for a bit, test it out to see if it's good enough, and make my pull request. It's honestly kind of demoralizing.
I suppose this is just the cost of progress; I'm sure there were people that loved raising and breeding horses but that's not an excuse to stop building cars.
I loved being able to figure out interesting solutions to software problems and hacking on them until something worked, and my willingness to do the math beforehand would occasionally give me an edge. Instead, now all I do is sit and wait while I'm cuckolded out of my work, and questioning why I bothered finishing my masters degree if the expectation now is to ship slop code lazily written by AI in a few minutes.
It was a good ride while it lasted; I got almost fifteen years of being paid to do my favorite thing. I should count my blessings that it lasted that long, though I'm a little jealous of people born fifteen years earlier who would be retiring now with their Silicon Valley shares. Instead, I get to sit here contemplating whether or not I can even salvage my career for the next five years (or if I need to make a radical pivot).
Wake me when we have ethically trained, open source models that run locally. Preferably high-quality ones.
At home, this has changed. Claude helped me setup a satellite dish, tune it, recompiled goesrec, for me and built a website to serve it - and my family dynamic was only “slightly interrupted” (daddy are you working still?). But it worked! And now I log in and tend to my projects with terminus instead of blindly go through the news or social media. Amazing! I’m still throwing myself at a new tech but way less invasve to my personal/family time.
At work though, i have been made into an absolute powerhouse. I invested the time years ago fussing with those oss projects and arch Linux or setting up lan parties and fixing my buddies rigs - toiling through terrible codebases at companies, deploying bad infrastructure, owning it and learning the hard way how to succeed - and it all is paying off and now 10x. AI can’t replace my judgement in the context of my org - maybe in time as the org shifts, but not for a few years.
The existential threat is not to me, at least for 5y - it’s when I’m asked - how do we get more features out the door?
* More headcount? Not unless they’re rockstars - more tokens.
* offshore talent? No, context switching and TZ - just more tokens.
* fly by night software startup xyz? No I’ll just write my own fault injection framework for $5 tailored to this project.
* consultants? Nope - pretty easy to try and fail fast and rewrite - again building to suite - software is disposable.
* oh no it was written in language xyz or deployed to cloud provider abc - no sweat, we’ll make it work on our cloud provider for $8.
Junior devs and offshore talent are the real losers here - I worry about them. Unless you’re die hard, I’d just assume do the work myself. But how do you accumulate this level of skill without getting paid to do it? I look back - I never got beyond baby projects or hobbies at home. I had to have someone roll the dice on me at a real job cause - rent and shit like that.
For those of you just starting out - I don’t have a great answer for you on how to start out, but - I can say you can install arch Linux, any oss project you want and all the things I did to get started in an afternoon - this is the new normal and embrace it.
For the rest of us it is our cloud moment - use the free tier - get your feet wet - we’re about to go for a hell of a ride. If you stick to the “took ur derbs” and want to keep treating your craft like artisian soap - go ahead, we’ll need those but don’t expect to survive on that
I am saying this in all seriousness, what difference is this to addiction?
This is something already talked about [1]. You are getting the sugar (results) and none of the nutrients (learning).
[1]https://quasa.io/media/the-hidden-dangers-of-ai-coding-agent...
https://hils.substack.com/p/help-my-husband-is-addicted-to-c...
Claude Code sure is great. Claud Code and my Codex reignited my passion for programming. Codex and Claude.
Ugh.
When you have no fucking idea what you're talking about, you cannot fix those issues. Simply telling opus "its broken, fix it" wont help. Sure, eventually it comes with a solution, but you have no idea if it's good.
Its like renting a bunch of construction tools and building a house. Unless you know what's important, you have no idea if your house will fall down tomorrow. At the end of the day, companies will always need an expert to sit there and confirm the code is good.
I only ever wanted to code.
I've spent decades developing mentorship, project management, and planning skills. I spent decades learning networking, databases, systems administration, testing, scrum, agile, waterfall, you name it. Every skill was necessary to build good software.
But I only ever wanted to code.
And I've spent decades burning out. I'm burt out on terrible documentation, tedious boiler plate and systems that don't interoperate well. I despise closed ecosystems, dependency management gone mad, terrible programming languages, over abstraction and I have fundamental and philosophical objections to modern software development practices.
I only ever wanted to code and I just couldn't do it anymore. And then AI happened.
This has been liberating for me.
The mountainous pile of terrible documentation written for somebody that has 36 years less experience? Ask the AI to find that one nugget I need.
That horrific mind numbingly tedious boilerplate? Doesn't matter if it's code, xml, yaml, or anything else. Have the AI do the busy work while I think about the bigger picture.
This nodejs npm dependency hell? Let the AI figure it out. Let the AI fix yet another breaking change and I'll review.
That hard to find bug? Let the AI comb through the logs and find the evidence. Present it to me with recommendations for a fix. I'll decide the path forward.
That legacy system nobody remembers? Let the AI reverse engineer it and generate docs and architectural diagrams. Use that to build the replacement strategy.
I've found a passion for active development that I've been missing for a very long time. The AI tools put power back in my hands that this bloated and sloppy industry took from me. Best of all it leverages the skills I've spent decades honing.
I can use the tools to engineer high quality solutions in an environment that has not been conducive to doing so on an individual level for a very long time. That is powerful and very motivating for somebody like me.
But I still fear the future. I fear a future where careless individuals vibe code a giant pile of garbage 10,000x the size of the pile of muck we have today. And those of us who actually try and follow good engineering practices will be right back to where we started: not able to get anything done because we're drowning in a sea of bullshit.
At least until that happens I'm going to be hyper productive and try to build the well engineered future I want to see. I've found my spark again. I hope others can do the same.
They started with co-opting DEI in open source so they could retain their positions without working. Part of the DEI people now probably pivoted to Trump.
Now they sell you out by promoting their intellectual wheelchairs, because they no longer care about future employment.
The three star bloggers that promote AI are all Gen-X.
Claude Code and it's parallels have extinguished multiple ones.
I was able to steer clear of the Bitcoin/NFT/Passport bros but it turns out they infiltrated the profession and their starry puppy delusional eyes are trying to tell me that iteration X of product Y released yesterday evening is "going to change everything".
They have started redefining what "I have build this" actually means, and they have outjerked the executives by slinging outrageous value creation narratives.
> I’m chasing the midnight hour and not getting any sleep.
You are 60; go spend some time with your grand-kids, smell a flower, touch grass forget chasing anything at this age cause a Tuesday like the others things are gonna wrap up.
Absolutely sincerely.