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.

I'm in my 60s and retiring this summer. I feel the opposite. Agents have removed most of the satisfaction and fulfilment from designing, building, testing and completing a feature or component. And if frameworks are a problem, learning to create simply and efficiently without them has its own sense of satisfaction.

Maybe it's a question of expectations. I suspect weavers felt the same with the arrival of mechanised looms in the industrial revolution. And it may be that future coders learn to get their fulfilment otherwise using agents.

I can absolutely see the attraction to business of agents and they may well make projects viable that weren't previously. But for this Luddite, they have removed the joy.

A year ago, cursor was flummoxed by simple things Claude code navigates with ease. But there are still corner cases where it hallucinates on the strangest seemingly obvious things. I'm working on getting it to write code to make what's going on in front of its face more visible to it currently.

I guess it's a question of where you find joy in life. I find no joy in frameworks and APIs. I find it entirely in doing the impossible out of sample things for which these agents are not competitive yet.

I will even say IMO AI coding agents are the coolest thing I've seen since I saw the first cut of cuda 20 years ago. And I expect the same level of belligerence and resistance to it that I saw deployed against cuda. People hate change by and large.

Lots of people wanted (and Intel tried to sell, somewhat succesfully) something they could just plug-and-play and just run the parallel implementations they'd already written for supercomputes using x86. It seemed easier. Why invest all of this effort into CUDA when Intel are going to come and make your current code work just as fast as this strange CUDA stuff in a year or two.

Deep learning is quite different from the earlier uses of CUDA. Those use cases were often massive, often old, FORTRAN programs where to get things running well you had to write many separate kernels targeting each bit. And it all had to be on there to avoid expensive copies between GPU and CPU, and early CUDA was a lot less programmable than it is now, with huge performance penalties for relatively small "mistakes". Also many of your key contributers are scientists rather than profressional programmers who see programming as getting in the way of doing what they acutally want to do. They don't want to spend time completely rewriting their applications and optimizing CUDA kernels, they want to keep on with their incremental modifications to existing codebases.

Then deep learning came along and researchers were already using frameworks (Lua Torch, Caffe, Theano). The framework authors only had to support the few operations required to get Convnets working very fast on GPUs, and it was minimal effort for researchers to run. It grew a lot from there, but going from "nothing" to "most people can run their Convnet research" on GPUs was much eaiser for these frameworks than it was for any large traditional HPC scientific application.

It seems funny though: The advantages of GPGPU are so obvious and unambiguous compared to AI. But then again, with every new technology you probably also had management pushing to use technology_a for <enter something inappropriate for technology_a>.

Like in a few decades when the way we work with AI has matured and become completely normal it might be hard to imagine why people nowadays questioned its use. But they won't know about the million stupid uses of AI we're confronted with every day :)

I remember being a bit surprised when I started reading about GPUs being tasked with processes that weren't what we'd previously understood to be their role (way before I heard of CUDA). For some reason that I don't recall, I was thinking about that moment in tech just the other day.

It wasn't always obvious that the earth rotated around the sun. Or that using a mouse would be a standard for computing. Knowledge is built. We're pretty lucky to stand atop the giants who came before us.

I didn't know about CUDA until however many years ago. Definitely didn't know how early it began. Definitely didn't know there was pushback when it was introduced. Interesting stuff.

*But TBF Xeon Phi was not a complete disaster for if you coded it in assembler you could squeeze out Fermi class GPU performance. Good luck getting the "recompile and run" crowd to do that though as they segued from that to relying on compiler directives going forward and that's how NVDA got a decade+ headstart that should never have happened, but did. Today a lot of these sorts are insisting that because of autograd, everything should be written in Python and compiled with an autograd DSL like torch. I am so glad I am close to retirement on that front. I already trust coding agents more than I trust this mindset.

But good luck explaining that to technical leaders who hadn't written a line of code in over a decade and yet somehow were in charge of things. People really need to consider the backstory here if they want to do better going forward, but I don't think they will. I think history is going to rhyme again.

Won't name names anymore, it really doesn't matter. But I feel the same way about people still characterizing LLMs as stochastic parrots and glorified autocomplete as I feel about certain CPU luminaries (won't name names) continuing to state that GPUs are bad because they were designed for gaming. Neither sorts are keeping up with how fast things change.

If it's the former, you hate AI agents. If it's the latter, you love AI agents.

Bear in mind also that the inputs to train LLMs on future languages and frameworks necessarily have to come from the hacker types. Somebody has to get their hands dirty, the "micro" of the parent post, to write a high quality corpus of code in the new tech so that LLMs have a basis to work from to emit their results.

I don't think you're a hacker. I think you enjoy writing code (good for you). Some of us just enjoy making the computer execute our ideas - like a digital magician. I've also gotten very good at the code writing and debugging part. I've even enjoyed it for long periods of time but there's times where I can't execute my ideas because they're bigger than what I can reasonably do by myself. Then my job becomes pitching, hiring, and managing humans. Now I write code to write code and no project seems too big.

But I'm looking forward to collapsing the many layers of abstraction we've created to move bits and control devices. It was always about what we could do with the computers for me.

What I want to do is create bespoke components that I can use to create a larger solution to solve a problem I have.

What I don't want to do is spend 45 minutes wrangling JSON to a struct so that I can get the damn component working =)

A quick example: I wanted a component that could see if I have new replies on HN using the Algol API. ~10 minutes of wall clock time with Claude, maybe a minute of my time. Just reading through the API spec is 15 minutes. Not my idea of fun.

The reason your login is slow is not because someone didn’t use the right algorithm.

Most game developers are just using other company’s engines.

While yes you need to learn how the architecture, the code isn’t the gating factor.

One example is the Amazon Prime Video team using AWS Step functions when they shouldn’t have and it led to inefficiencies. This was a public article that I can’t find right now.

(And before someone from Amazon Retail chimes in and says much of Amazon Retail doesn’t run on AWS and uses the legacy CDO infrastructure - yes I know. I am a former AWS employee).

These are not toys. I want to make money. The customers want feature after feature, in a steady stream. It's bad business if the third or fourth feature takes ages. The longer stream, the better financially.

That the code "works" on any level is elementary, Watson, what must "work" is that stream of new features/pivots/redesigns/fixes flowing.

But the important thing is getting solutions to users. Claude makes that easier.

At the moment I am trying to fix a vibe coded application and while each individual function is ok, the overall application is a dog’s breakfast of spaghetti which is causing many problems.

If you derive all your pleasure from actually typing the code then you’re probably toast, but if you like building whole systems (that run on production infrastructure) there is still heaps of work to do.

The need for assembly programmers diminished over the decades. A similar thing will happen here.

I highly recommend not using these tools in their "agentic" modes. Stay in control. Tell them exactly what to write, direct the architecture explicitly.

You still get the tremendous benefit of being unlocked from learning tedious syntax and overcoming arcane infra bottlenecks that suck the joy out of the process for me, but you get freed from the tedious and soul crushing parts.

Obviously you should do whatever you want, however you want to do it, and not just do whatever some Internet rando tells you to do, but glorified autocomplete is so 1 year ago. Everyone knows the $20/month plans aren't going to last, time will tell if the $100/month ones do. The satisfaction is now in completing a component and getting to polish it in a way you never had time for before. And then totally crushing the next one in record time. To each their own, of course, but personally, what's been lost with agentic mode has been replaced by quantity and quality.

Their tag line: "Don't outsource your thinking to AI. Instead, use AI to become a better problem solver, clearer thinker, and more elegant coder."

I have followed the course myself and it reignited my passion. During the course I built a cool side project from scratch, small steps, no vibe coding using the course's principals. It was really satisfying, I felt in control again while learning new things.

Congrats! I'm in that age where I'm envying more the ones like you than the 20-something :)

It is still weird to me that I talk to a remote Python app but that's how we write code nowadays. Still, I felt almost mocked when Claude plauded my "accomplishments".

So I'd say that I am definitely more productive than I used to be but I enjoy the work less on one level. But on another level, I feel like I can build a lot more and tackle problems I wouldn't have tackled in the past. It's a mixed blessing. It's also a WIP, I expect that the way we write code will change even more, over the next few years.

I love it, I hate it, it's the Brave New World of software development.

Its almost like it reignites novelty at things that were to administratively heavy to figure out. Im not sure if its fleeting or lasting.

I got completely fed up of continually having to learn new incantations to do the same shit I’ve been doing for decades without enough of a value add on top. I know what I want to build, and I know how to architect and structure it, but it’s simply not a good investment of my increasingly limited time to learn the umpteenth way to type code in simply to display text, data, and images on the web - especially when I know that knowledge will be useful for maybe, if I’m lucky, a handful of years before I have to relearn it again for some other opinionated framework.

It’s just not interesting and I’ve become increasingly resentful of and uninterested in wasting time on it.

Claude, on the other hand, is a massive force multiplier that enables me to focus on the parts of software development I do enjoy: solving the problems without the bother of having to type it all in (like, in days of old, I’d already solved the problem before my fingers touched the keyboard but the time-consuming bit was always typing it all in, testing and debugging - all of that is now faster but especially the typing part), focussing on use cases and user experience.

And I don’t ever have to deal directly with CSS or Tailwind: I simply describe the way I want things to look and that’s how the end up looking.

It’s - so far at any rate - the ultimate in declarative programming. It’s awesome, and it means I can really focus on the quality of the solution, which I’m a big fan of.

Computers do not feature at all in my ideal retirement. Maybe a phone or tablet so I can do the minimal email and bill paying.

"Resentful" is a perfect way of putting it - I may just be old and grumpy now, but I think it's sad what we as a community have done to the process of web development. It's such a circle jerk. Node in my view is the worst thing that ever happened to building web applications.

Enter Claude Cowork. I've spent the past few days building an app that would have taken me weeks of time in the past. It's using a framework I've never built with, and I don't have to learn the intricacies. Shipping this to Vercel and hosting the database on Supabase is incredibly easy and it's very exciting. The only drawback so far is the unsettling fear of the unknown regarding leaking secrets and whatnot, so I'm going to have to manually audit the finished project before deploying.

And here I thought my days of "side projects" were completely over.

- Root cause and fix failures.

- Run any code "what if scenario".

- Performance optimizations.

- Refactor.

There's no reason why you shouldn't (and you should) read all the code and understand it after Claude does any work for you but the experience vs. the "old" SO model of looking for some technical detail is very different.

Some of my colleagues didn’t make the jump. Those that were the most into AngularJS back then are still writing Angular apps today.

I'm only in my forties. I've been nostalgic for the days when I'd stay up all night exploring new frontiers (for me) in tech for a number of years. I could not disagree more with your take on this.

Someone said they value their time before death and you're pretty dismissive. Priorities change. Values change. Conditions change.

> Especially coupled with the fact that tech has never moved so fast as right now, being on top of the AI-game is a target changing a hundred times faster than frontend frameworks back in the days.

I mean, isn't that what people in this thread have been saying about frameworks? How many hours have been lost relearning how to solve a problem that has already been solved? It's like when I tried to fix a date-time issue on Windows as a Mac / Linux user. I knew NTP was the answer but I had to search the web to find out where to turn it on. Stuff like that is pretty frustrating and I didn't even have to do it every five to ten years.

Implementation details can very much matter though. I see this attitude from my managers that now submit huge PRs, and it is becoming a big problem.

I definitely agree that these tools allow one with an in-depth developer background to cover territory that was too much work previously. But plop me into a Haskell codebase, and I guarantee I’d cause all kinds of problems even with the best intentions and newest models. But the ramp up for learning these things has collapsed dramatically, and that’s very cool.

I still don’t want to have to learn all the pitfalls of those frameworks though. Hopefully we will converge on a smaller number, even if it’s on tooling that isn’t my favourite.

And a rewrite of a non-trivial application, even with the AI goodness, is still a big proposition and full of all kinds of risk. If you have a trivial application, you probably don’t have much protecting you from someone else vibing up a competing replacement either.

Coding has never _stopped_ being a passion for me, but my increasingly limited time becomes an issue.

And Claude code (and cursor) saves me So. Much. Time.

I only have 10-20 active years ahead of me, so this is really, really important. Young ppl don’t get it.

Now I do fun code on a laptop on the sofa with my family. I’m only typing in tiny breaks between socializing and I’m still getting lots of fun stuff done.

I can play with my 7yo, or help the 11yo with his homework, or go for a run!! - while LLM is cooking a well-spec'ed agentic task. This sounds embarrassing, but LLMs have made our lives healthier, unbelievable.

PS. Not to mention all the "boring" sh*t that I used to postpone/stretch indefinitely, like writng docs or polishing copy for my website... No more stress, no procrastination, just let the LLM do it.

Phone goes back in my pocket and I'll have something to read when I get back home. Literal game changer.

Before those ideas just flowed through my brain and either out the other ear or got filed into "don't have time anyway shrug" and never completed or even tried.

10-20 years of coding/technology or living?

Young ppl rarely get life experience/choices, youth is wasted on the young, 1000x for my own.

Agreed. Leetcode caused more harm than good.

Where do I even begin...yes, you should care about implementation details unless you're only going to write stuff you run locally for your own amusement.

They often do solve business problems around responsive design, security and ux.

Currently working maintenance with one foot in a real legacy system and the other foot in modern systems the difference is immense.

>> tired of the never ending rat race of keeping up with the latest bizarre web stacks, frameworks

Instead of just keeping up with the latest development frameworks, I also now have to keep up with the latest AI frameworks. I spent a week at my $ job just installing plugins, requesting permissions, debugging issues with the agents, before I went back to writing code myself (plumbing between the latest frameworks) because I'm expected to get stuff done in addition to managing agents that were supposedly going to do my work for me.

My personal agentic AI coding setup never fully materialized while I have been chasing the latest crazes and I am back to handwriting my personal code too (with AI chat help) until I manage to stick to a particular setup.

Anyway, I feel like the rat race just opened yet another front. And I bet I'll still be expected to leetcode in my next technical interview (still was in 2025) in addition to leetprompting or whatever the next segment on interviews will be.

But the real talk we need to have is... "Uber for cats"

Now a lot can be cast off to LLMs to focus on the problem space and the innovative computing use around them. It’s been exciting to not worry about arbitrary idiosyncrasies and machete through jungles of technical minutia to get to the clearing. I still have to deal with them but less of them. And I don’t have to commit nearly as much in the technical space to memory to address problems, I can often focus on higher level architectural decisions or new approaches to problems. It’s been quite enjoyable as well.

I kinda feel the same way when I visit Home Depot once a year

Have you tried Claude? No, Opus? No, not that version, it's two weeks old, positively ancient lol. Oh wait, now OpenClaw is the cool thing around the block.

My dude, the rat race just became a rat sprint. I hope you're keeping up, you're no spring chicken any more.

It makes it so easy to cut through the bullshit. And I've never considered myself scared of asking "stupid" questions. But after using these AI tools I've noticed that there are actually quite a few cases where I wouldn't ask (another human) a question.

Two examples: - What the hell does React mean when they say "rendering"? Doesn't it just output HTML/a DOM tree and the browser does the actual rendering? Why do they call it rendering? - Why are the three vectors in transformer models named query, key & value? It doesn't really make sense, why do they call it that?

In both cases it turns out, the question wasn't really that stupid. But they're not the kind of question I'd have turned to Stackoverflow for.

It really is a bit like having a non-human quasi-expert on most topics at your fingertips.

And yet, having customers and listening to them is the whole point.

Anything that re-ignites a person's zest for thinking and creating is a net gain.

That said, it is paradoxical that the catalyst in this case is a technology that replaces thinking.

My experience is that people who weren't very good at writing software are the ones now "most excited" to "create" with a LLM.

You have decades of expert knowledge, which you can use to drive the LLMs in an expert way. Thats where the value is. The industry or narrative might not have figured that out yet, but its inevitable.

Garbage in, garbage out still very much applies in this new world.

And just to add, the key metric to good software hasn't changed, and won't change. It's not even about writing the code, the language, the style, the clever tricks. What really matters is how well does the code performs 1 month after it goes live, 6 months, 5 years. This game is a long game. And not just how well does the computer run the code, but how well can the humans work with the code.

Use your experience to generate the value from the LLMs, cuase they aren't going to generate anything by themselves.

Massive job cuts, bad job market, AI tools everywhere, probable bubble, it seems naive to be optimistic at this juncture.

LLMs may be accelerating the process, but definitely not the cause.

If you want a career in technology, a durable one, you learn to adapt. Your primary skill is NOT to master a given technology, it is the ability to master a given technology. This is a university that has no graduation!

If you're a great programmer, can you can stop using Angular and master React? Yes. Can you stop telling the computer what to do, and master formal proof assistants? Maybe. Can you stop using the computer except as a tool and go master agricultural technology? Probably not. (Which is not to say you can't be a good programmer at an agritech company)

Let’s talk about the industries the computer killed: travel agents, musician, the entire film development industry, local newspapers built on classified ads, the encyclopedia industry, phone operators, projectionists, physical media industries, and a few dozen other random industries.

We aren’t special because we are coders. Creativity and engineering thoughtfulness will still exist even with LLMs, it will just take a different form.

This is the fundamental problem with how so many people think about LLMs. By the time you get to Principal, you've usually developed a range of skills where actual coding represents like 10% of what you need to do to get your job done.

People very often underestimate the sheer amount of "soft" skills required to perform well at Staff+ levels that would require true AGI to automate.

I remember a cinema theater projectionist telling me exactly that while I was wiring a software controlling numeric projector, replacing the 35mm ones.

[citation needed]

It has just merely moved from "almost, but not entirely useless" to "sometimes useful". The models themselves may perhaps be capable already, but they will need much better tooling than what's available today to get more useful that that, and since it's AI enthusiasts who will happily let LLMs code them that work on these tools it will still take a while to get there :)

[citation needed]

:P

This thing has changed the way I work. I barely touch my editor to actually edit anymore, because speaking into the chat field what changes I want it to make is more efficient

The tooling does need to get better, yes, but anecdotally, I do a fundamentally different job (more thinking, less typing, less sifting through docs, less wiring up) than 3 months ago

So much of my career was spent on especially rummaging in docs and googling and wiring things up. I believe that's the same for most of us

If we truly value human creativity, then things that decrease the rote mechanical aspects of the job are enablers, not impediments.

Society can interpret sociopathy as damage and route around it, if we do the work to make it happen. It will not happen by itself without effort.

After 40 years in this industry—I started at 10 and hit 50 this year—I’ve developed a low tolerance for architectural decay.

Last night, I used Claude to spin up a website editor. My baseline for this project was a minimal JavaScript UI I’ve been running that clocks in at a lean 2.7KB (https://ponder.joeldare.com). It’s fast, it’s stable, and I understand every line. But for this session, I opted for Node and neglected to include my usual "zero-framework" constraint in the prompt.

The result is a functional, working piece of software that is also a total disaster. It’s a 48KB bundle with 5 direct dependencies—which exploded into 89 total dependencies. In a world where we prioritize "velocity" over maintenance, this is the status quo. For me, it’s unacceptable.

If a simple editor requires 89 third-party packages to exist, it won't survive the 5-year test. I'm going back to basics.

I'll try again but we NEED to expertly drive these tools, at least right now.

> neglected to include my usual "zero-framework" constraint in the prompt

And then your complaint is that it included a bunch of dependencies?

AI's do what you tell them. I don't understand how you conclude:

> If a simple editor requires 89 third-party packages to exist

It obviously doesn't. Why even bother complaining about an AI's default choices when it's so trivial to change them just by asking?

You call it a trivial instruction, but it is not trivial. It was a core requirement for your own design that you neglected to specify. This is not different than leaving out any other core requirement for a engineering specification.

Most people would NOT want this requirement. Meaning most people wouldn't care if there are package dependencies are not, so the agent 100% did the right thing.

What's missing is another LLM dialog between you and Claude. One that figures out your priorities, your non-functional requirements, and instructs Claude appropriately.

We'll get there.

There are already spec frameworks that do precisely this. I've been using BMAD for planning and speccing out something fairly elaborate, and it's been a blast.

I have been consistently skeptical of LLM coding but the latest batch of models seems to have crossed some threshold. Just like everyone, I've been reading lots of news about LLMs. A week ago I decided to give Claude a serious try - use it as the main tool for my current work, with a thought out context file, planning etc. The results are impressive, it took about four hours to do a non-trivial refactor I had wanted but would have needed a few days to complete myself. A simpler feature where I'd need an hour of mostly mechanical work got completed in ten minutes by Claude.

But, I was keeping a close eye on Claude's plan and gradual changes. On several occasions I corrected the model because it was going to do something too complicated, or neglected a corner case that might occur, or other such issues that need actual technical skill to spot.

Sure, now a PM whose only skills are PowerPoint and office politics can create a product demo, change the output formatting in a real program and so on. But the PM has no technical understanding and can't even prompt well, let alone guide the LLM as it makes a wrong choice.

Technical experts should be in as much demand as ever, once the delirious "nobody will need to touch code ever again gives way to a realistic understanding that LLMs, like every other tool, work much better in expert hands. The bigger question to me is how new experts are going to appear. If nobody's hiring junior devs because LLMs can do junior work faster and cheaper, how is anyone going to become an expert?

It’s refreshing to hear I’m not the only one who feels this way. I went from using almost none of my copilot quota to burning through half of it in 3 days after switching to sonnet 4.6. I’m about to have to start lobbying for more tokens or buy my own subscription because it’s just that much more useful now.

I'm still not ready to sing praises about how awesome LLMs are, but after two years of incremental improvements since the first ChatGPT release, I feel these late-2025 models are the first substantial qualitative improvement.

I see unreliable software like openclaw explode in popularity while a Director of Alignment at Meta publicly shares how it shredded her inbox while continuing to use openclaw [1], because that's still good enough innit? I see much buggier releases from macOS & Windows. The biggest military in the world is insisting on getting rid of any existing safeguards and limitations on its AI use and is reportedly using Claude to pick bombing targets [2] in a bombing campaign that we know has made mistakes hitting hospitals [3] and a school [4]. AI-generated slop now floods social networks with high popularity and engagement.

It's a known effect that economies of scale lowers average quality but creates massive abundance. There never really was a fundamental quality bar to software or creative work, it just has to be barely better than not existing, and that bar is lower than you might imagine.

[1] https://x.com/summeryue0/status/2025774069124399363

[2] https://archive.ph/bDTxE

[3] https://www.reuters.com/world/middle-east/who-says-has-it-ha...

[4] https://www.nbcnews.com/world/iran/iran-school-strike-us-mil...

I must say I find this idea, and this wording, elitist in a negative way.

I don't see any fundamental problem with democratization of abilities and removal of gatekeeping.

Chances are, you were able to accumulate your expert knowledge only because:

- book writing and authorship was democratized away from the church and academia

- web content publication and production were democratized away from academia and corporations

- OSes/software/software libraries were all democratized away from corporations through open-source projects

- computer hardware was democratized away from corporations and universities

Each of the above must have cost some gatekeepers some revenue and opportunities. You were not really an idiot just because you benefited from any of them. Analogously, when someone else benefits at some cost to you, that doesn't make them an idiot either.

It would be like if you put in all this time to get fit and skilled on mountain bikes and there was a whole community of people, quiet nature, yada yada, and then suddenly they just changed the rules and anyone with a dirt bike could go on the same trails.

It's double damage for anyone who isn't close to retirement and built their career and invested time (i.e. opportunity cost) into something that might become a lot less valuable and then they are fearful for future economic issues.

I enjoy using LLMs and have stopped writing code, but I also don't pretend that change isn't painful.

However, our personal emotions need not turn into disparaging others' use of the same skills for their satisfaction / welfare / security.

Additionally, our personal emotions need not color the objective analysis of a social phenomenon.

Those two principles are the rationales behind my reply.

I suppose I see "any idiot" as a more general phrase, like "idiot proof", not directly meaning that anyone who uses a LLM is an idiot. However I can also see how it would be seen as disparaging.

Also, while there's a lot of examples of people entrenching into a certain behavior or status and causing problems, I also think society is a bit harsh on people who struggle with change. For people who are less predisposed to be ok with change feels like a lot of the time the response is "just deal with it and don't be selfish, this new XYZ is better for society overall".

Society is pretty much made up of personal emotions on some level. I don't think we should go around attacking people, but very few things can be considered truly objective in the world of societal analysis.

This parroted argument is getting really tired. It signals either astroturfing or someone who just accepts what they are sold without thinking.

LLMs aren’t “democratising” anything. There’s no democracy in being mostly beholden to a few companies which own the largest and most powerful models, who can cut you off at any time, jack up the prices to inaccessibility, or unilaterally change the terms of the deal.

You know what’s truly “democratic” and without “gatekeeping”? Exactly what we had before, an internet run by collaboration filled with free resources for anyone keen enough to learn.

There are loads of high performance open source LLMs on the market that compete with the big 3. I have not seen this level of community engagement and collaboration since the open-source boom 20 years ago.

The issue arises from it not being that person’s opinion but a talking point. People didn’t all individually arrive at this “democratisation” argument by themselves, they were sold what to say by the big players with vested interest in succeeding.

I’m very much for discussing thoughts one has come up with themselves, especially if they disagree with mine. But what is not productive is arguing with a proxy.

> I have not seen this level of community engagement and collaboration

Nor this level of spam and bad submissions.

Programming is a tricky skill and takes a long time to get good at. Lots of people aren't good at it. AI helps them program anyway, and allows them to sometimes produce useful programs. That's it.

It's not a talking point. It's just the reality of what the technology enables, and it's a simple enough observation that millions of people can independently arrive at that conclusion, and some of them might even refer to it as "democratization".

This is a good thing. It's a filter for the careless, lazy, and incompetent. LLMs are to programming what a microwave is to food. I'm not a chef because I can nuke a hot pocket. "Vibe coders" (not AI-assisted coding) are the programming equivalent of the people on Kitchen Nightmares. Go figure, it's a community rife with narcissism, too.

> Nor this level of spam and bad submissions.

Your comments seem pretty aggressive for what you’re replying to. Maybe take a beat to assess your biases? I thought the main comment was pretty fair and sensible, yet somehow you landed on calling them a spammer/bad submitter/astroturfer/non-thinker. Maybe they are? I could be wrong, but that's quite a strong reaction for what they asserted at face value. Not really trying to police anything here, I just thought the initial comment had merit and this devolved quite quickly.

It is what we are talking about, hence not "counterproductive".

They absolutely are. Anytime new knowledge or skills become widely available to everyone, that's a term used for it.

> There’s no democracy in being mostly beholden to a few companies which own the largest and most powerful models, who can cut you off at any time, jack up the prices to inaccessibility, or unilaterally change the terms of the deal.

None of that has anything to do with anything. There's competition between companies to keep prices low and accessibility high.

I think you are simply misunderstanding the word "democratic". It isn't just political. From MW:

> 3 : relating, appealing, or available to the broad masses of the people : designed for or liked by most people

Here, it's specifically about making things available to the broad masses of the people that wasn't before.

This isn't a matter of opinion. It's just the meaning of the word.

There was no tool that could easily whip up a 300 line script to do something for me when I didn't know how to code, and do it in just seconds.

The topic here is the democratization of a whole set of new abilities. Not just knowledge.

....so your only option was to just learn? learn from the freely & vastly available resources? Oh the horrifying tragedy. How elitist. How gatekeeping. To think you had to put in effort.

LLMs have a lot of good and bad, but saying they "democratised" anything is plain bullshit.

But that's the whole point...

When things that previously took hundreds of hours of effort to learn now become available with just a few minutes, they become available to all -- even those without all that extra time, which is most people who have a lot of other competing priorities in their lives.

That's democratization.

I don't understand why you're trying to argue against that. It's a dictionary definition. It's just a meaning of the word.

Whatever you seem to be upset about is something else, I don't know what.

The knowledge was freely available for the past few decades, way before LLMs - ergo democratised. The skill was freely available to learn - ergo democratised.

By your logic,

- cooking is a skill that is not democratised (despite the fact that you can do it everyday).

- A person who doesn't take time to learn cooking from cookbooks or youtube tutorials is somehow "lacking access" and in an undemocratic position

- An experienced chef is monopolising his mastery of cooking skills bc they dedicated time to learning them

- a robot that manufactures food is "democratising cooking skills"

See how silly that sounds?

Not to mention, using an AI to do a job is NOT equivalent to learning the skills of the job. Me having a printer doesn't mean "painting skills were democratised".

That would not happen, simply because those companies' interest will never be aligned entirely. There are at least three SOA models at the moment plus many open weight models. Anthropic vs. Pentagon is exactly what would play out.

And what is a precedence? Don't say Google, because search is well and alive.

> You know what’s truly “democratic” and without “gatekeeping”? Exactly what we had before, an internet run by collaboration filled with free resources for anyone keen enough to learn.

We have way more free resources at the moment. Name anything you'd like to learn, someone will be able to point you to a relevant resource. There are also better ways of surfacing that resource.

> This parroted argument

Most of arguments here on HN have been discussed ad nauseam, for or against AI. It's only parroted (or biased) if it's against your own beliefs.

Everyone already had the option to write any code, fork any open source project, publish any of their code, run any of their code but suddenly AI appears and THAT is what makes it democratic? What was undemocratic about it? Is this democracy where idiots are running ai agents that publish smear campaigns, or harass maintainers for not accepting their slop is the democratic future you wish for?

How many (job) positions do you see today that want a backend developer? Frontend developer? Not much because now everyone is expected to be at least full stack, if not also devops as well. The exact same thing is playing out right now with AI, people are expected to produce 5x the amount of code before, if you don't, someone else will take your job that is willing to do it.

Already bloated programs will bloat further, they will require even more resources to run, you will have to pay even more for hardware, they will be slower, less responsive, you will have to pay yet another monthly fee to big tech for their AIs, and people will happily do it and pat themselves that we democratized programming, while running towards the future where nobody will be able to own hardware capable of general computing.

Why blame big tech when they're just providing a service at a fair cost (3rd party inference is incredibly cheap)? I'm not sure how that makes sense.

LOL. Maybe you are referring to OpenAI and Anthropic? Yes they have codex and opus. But about 1-2 months behind them is Grok, Gemini, and then 2-3 months behind them are all the other models available in cursor, from chinese open source models to composer etc.

How you can possibly use this "big company takes everything away" narrative is ridiculous, when you can probably use models for free that are abour 2 months behind the best models. This is probably the most uncentralised tech boom ever.

(I mean openAI is in such a bad state, I wouldn't be surprised if they lose almost their entire lead and user base within 6-12 months and are basically at the level of small chinese llm developers).

> when you can probably use models for free that are about 2 months behind the best models.

You can use them for free, but training of near-SOTA foundation models is currently not an open-source process, and it's funded almost exclusively by large and/or wealthy corporations. That's a weak point at the moment from the perspective of openness.

It was very democratized before, almost anyone could pick up a book or learn these skills on the internet.

Opportunity was democratized for a very long time, all that was needed was the desire to put in the work.

OP sounds frustrated but at the same time the societal promise that was working for longest time (spend personal time specializing and be rewarded) has been broken so I can understand that frustration..

/s, obviously I would hope except I've actually seen this sentiment expressed seriously.

It does not. Legumes, whole grains, vegetables, and yogurt have always been cheaper than processed food.

People prefer eating carbohydrates and saturated fats.

It doesn’t.

They are not obese because they cannot afford the necessary amounts of protein and calories from healthy sources in the grocery store.

If you eschew those highs and settle for some sprouted moong bean salad with a little bit of salt/lime/black pepper, or hummus and veggies, or eggs with some smashed avocado on toast, tofu and some broccoli, etc, then it does not cost much time.

There is no baking involved, just cutting, mixing, blending, and maybe soaking. Sautéing or pan frying in a little bit of olive oil or canola oil is also quick.

There is a range of food processing going from raw vegetables to hyper-processed fast food.

Most people fall in the middle and expect their food to be boiled, baked, fried, etc. They will most likely not want to eat just salad 24/7.

And once you start boiling/baking/frying, you're adding a lot more processing time just for the food itself, let alone cleaning everything afterwards.

Plus people want varied diets so they need a wide range of different tasting foods, most like 10-20.

Cooking well is really hard and very time consuming (and potentially costly!) to learn.

Healthy living enthusiasts as well as cooking enthusiasts (sometimes the same group, many times different groups) REALLY love to minimize this aspect.

>Most people fall in the middle and expect their food to be boiled, baked, fried, etc. They will most likely not want to eat just salad 24/7.

> Plus people want varied diets so they need a wide range of different tasting foods, most like 10-20.

Right, so the problem is people want to eat unhealthy foods. The problem is not lack of money or time.

Did you actually read the 2 paragraphs you quoted?

It doesn't. Carbs like rice, potatoes, etc. are incredibly cheap. Protein like ground beef and basic cuts of chicken are not expensive. And broccoli, carrots, green peppers, apples -- these are not exactly breaking the bank. Product is seasonal, so you vary what you buy according to what is cheapest this week.

Meanwhile, stuff like breakfast cereal and potato chips and Oreo cookies actually are surprisingly expensive.

Eating too many carbs is not a healthy diet dude

is there price to be paid for getting any desired result imaginable without effort on a press of a button?

So it's not just software that's coming to an end, everything else is as well. But; billionaires wives will still need haircuts (women billionaires will also need haircuts), so hairdresser will be the last profession.

That said, if we zoom out and review such paradigm shifts over history, we find that they usually result in some new social contracts and value systems.

Both good expert writers and poor novice writers have been able to publish non-fiction books from a few centuries now. But society still doesn't perceive them as the same at all. A value system is still prevalent and estimated primarily from the writing itself. This is regardless of any other qualifications/disqualifications of authors based on education / experience / nationality / profession etc.

At the individual level too, just because book publishing is easy doesn't mean most people want to spend their time doing that. After some initial excitement, people will go do whatever are their main interests. Some may integrate these democratized skills into their main interests.

In my opinion, this historical pattern will turn out to be true with the superdrug as well as vibe coding.

Some new value will be seen in the swimming or running itself - maybe technique or additional training over and above the drug's benefits.

Some new value will be discovered in the code itself - maybe conceptual clarity, algorithmic novelty, structural cleanliness, readability, succinctness, etc. Those values will become the new foundations for future gatekeeping.

It's a nice idea, but I feel like that's only going to be the case for very small companies or open source projects. Or places that pride themselves on not using AI. Artisan code I call it.

At my company the prevailing thought is that code will only be written by AI in the future. Even if today that's not the case, they feel it's inevitable. I'm skeptical of this given the performance of AI currently. But their main point is, if the code solves the business requirements, passes tests and performs at an adequate level, it's as good as any hand written code. So the value of readable, succinct, novel code is completely lost on them. And I fear this will be the case all over the tech sector.

I'm hopeful for a bit of an anti-AI movement where people do value human created things more than AI created things. I'll never buy AI art, music, TV or film.

But I do agree, if everyone can build software then the allure of it along with the value will be lost. Vibe coding is only a superpower as long as you're one of the select few doing it. Although I imagine it will continue to become a niche thing, anyone who thinks everyone and their grandma will be vibing bespoke software is out to lunch.

Personally I think there is a certain je ne sais quoi about creating software that cannot be distilled to some mechanical construct, in the same way it exists for art, music, etc. So beyond assembly line programming, there will always be a human involved in the loop and that will be a differentiating factor.

Given how I've seen a lot of AI "artists" describe themselves and "their" works, yeah, probably a lot of them would.

I gatekeep my bike, I keep it behind a gate. If you break the gate open and democratize my bike, you're an idiot.

Maybe it's of value that any idiot can do this, but we're still idiots.

You gatekeep your bike, you keep it behind a gate, you don't let anyone else ride it.

Your neighbor got a nicer bike for Christmas, rode it by your house and now you are sad because you aren't the special kid with the bike any more, you are just regular kid like your neighbor.

We can dish it out, but we can’t take it.

As for the comparisons - some are partly comparable to the current situation, but there's some differences as well. Sure books and online content enabled others to join, thereby reducing the "moat" for those who built careers on esoteric knowledge. But it didn't make things _that_ easy - it still required years of invested time to become a good developer. Also, it happened very gradually and while the developer pie was growing, and the range of tech growing, so developers who kept on top of technology (like OP did) could still be valuable. Of course, no one knows fully how it will play out this time around; maybe the pie will get even bigger, maybe there's still room for lots of developers and the only difference is that the tedious work is done. Sure, then it is comparable. But let's be honest, this has a very real chance of being different (humans inventing AI surely is something special!) and could result in skill-sets collapsing in value at record time. And perhaps worse, without opening new doors. Sure, new types of jobs may appear but they may be so different that they are essentially completely different careers. It is not like in the past you just needed to learn a new programming language.

Gates were put in place for lawyers, doctors, and engineers (real ones, not software "engineers") because the cost of their negligence and malpractice was ruined lives and death. Gatekeeping has value.

Software quality, reliability, and security was already lousy before the advent of LLMs, making it increasingly clear that the gate needed to be kept. Gripes about "gatekeeping" are a dogwhistle for "I would personally benefit from the bar being lowered even further".

This discussion is specifically about lowering the barriers of programming and creating using software.

I haven't said anything at all about other professions nor do I think my arguments for democratizing software creation apply to law, medicine, or "real" engineering.

There's also a false equivalence in the software part of your comment. It equates lowering of barriers for recreational/hobby coding with software engineering for serious purposes.

Since you dismiss me as a dogwhistle, I hope my terming your argument as elitist, strawmanish, and full of false equivalences is only seen as fair.

Skill based one of course.

Open research papers, that everyone can access is democratizing knowledge. Accessibile worldwide courses, maybe (like open universities).

But LLMs are not quite the sane. This is taking knowledge from everyone and, in the best case, paywalling it.

I agree in spirit that the original comment was classist, but in this context your statements are also out of place, in my opinion.

- What if these centralized providers had restricted their LLMs to a small set of corporations / nations / qualified individuals?

- What if Google that invented the core transformer architecture had kept the research paper to themselves instead of openly publishing it?

- What if the universities / corporations, who had worked on concepts like the attention mechanism so essential for Google's paper, had instead gatekept it to themselves?

- What if the base models, recipes, datasets, and frameworks for training our own LLMs had never been open-sourced and published by Meta/Alibaba/DeepSeek/Mistral/many more?

I'm pretty sure that someone else would have come around the corner with a similar idea some time later, because the fundamentals of these stuff were already discussed decases before "Attention is all you need" paper, the novel thing they did was combining existing knowhow into a new idea and making it public. A couple of ingredients of the base research for this is decades old (interestingly back then some European universities were leading the field)

I am not trying to be dismissive, but this could apply to all research ever

Cell phones made communication easier for exactly zero people even though billions have been sold. Why? Because they come from just a few different companies.

Similar story to cell phones.

LLMs are in this state right out the gate.

I'm sure that's why they're investing 1 trillion in AI, for the poor illiterates.

—- from a ‘principal engineer’

It's funny you say that, because I've seen plenty of the reverse elitism from "AI bros" on HN, saying things like:

> Now that I no longer write code, I can focus on the engineering

or

> In my experience, it's the mediocre developers that are more attached to the physical act of writing code, instead of focusing on the engineering

As if getting further and further away from the instructions that the CPU or GPU actually execute is more, not less, a form of engineering, instead of something else, maybe respectable in its own way, but still different, like architecture.

It's akin to someone claiming that they're not only still a legitimate novelist for using ChatGPT or a legitimate illustrator for using stable diffusion, but that delegating the actual details of the arrangement of words into sentences or layers and shapes of pigment in an image, actually makes them more of a novelist or artist, than those who don't.

I've been a tech lead for years and have written business critical code many times. I don't ever want to go back to writing code. I am feeling supremely empowered to go 100x faster. My contribution is still judgement, taste, architecture, etc. And the models will keep getting better. And as a result, I'll want to (and be able to) do even more.

I also absolutely LOVE that non-programmers have access to this stuff now too. I am always in favor of tools that democratize abilities.

Any "idiot" can build their own software tailored to how their brains think, without having to assemble gobs of money to hire expensive software people. Most of them were never going to hire a programmer anyway. Those ideas would've died in their heads.

Programming was already “democratized” in the sense that anyone could learn to program for free, using only open-source software. Making everyone reliant on a few evil megacorporations is the opposite of democratization.

It's the same sort of argument artists use when it comes to AI generated media, there obviously is a qualitative difference in the people now able to generate whatever they want versus needing to draw something by hand, so saying "they could've just learned to draw themselves" is not very convincing. People don't want to do that yet still get an output, and I see nothing wrong with that, and if you do, it's just another sort of gatekeeping, that the "proper" way is to learn it by hand.

Lastly, many, many open weight models exist.

This is like democratizing house building.

They are not building anything; they are paying Anthropic to do it.

> Lastly, many, many open weight models exist.

Good luck running them when the megacorporations have bought out all hardware.

Open weight models run on consumer hardware, if you think corporations will make every single piece of hardware unaffordable then I'm not sure what to tell you.

One thing is for sure LLMs will bring down down the cost of software per some unit and increase the volume.

But..cost = revenue. What is a cost to one party is a revenue to another party. The revenue is what pays salaries.

So when software costs go down the revenues will go down too. When revenues go down lay offs will happen, salary cuts will happen.

This is not fictional. Markets already reacted to this and many software service companies took a hit.

I'm in the boat of SaaS myself, but feel a bit dishonesty from Senior devs complaining about technology stealing jobs. When it was them doing the stealing, it was fine. Now that the tables have turned, it's not technology is bad

You may not end up with a seat at the table.

But my take on this is that accountability will still be a purely human factor. It still is. I recently let go of a contractor who was hired to run our projects as a Scrum/PM, and his tickets were so bad (there were tickets with 3 words in them, one ticket was in the current sprint, that was blocked by a ticket deep in the backlog, basic stuff). When I confronted him about them, he said the AI generated them.

So I told him that:

1. That's not an excuse, his job is to verify what it generated and ensure it's still good.

2. That actually makes it look WORSE, that not only did he do nearly 0 work, that he didn't even check the most basic outputs. And I'm not anti-AI, I expressly said that we should absolutely use AI tools to accelerate our work. But that's not what happened here.

So you won't get to say (at least I think for another few years) "my AI was at fault" – you are ultimately responsible, not your tools. So people will still want to delegate those things down the chain. But ultimately they'll have to delegate to fewer people.

I'm assuming that the software factory of the future is going to need Millwrights https://en.wikipedia.org/wiki/Millwright

But, builders are builders. These tools turn ideas into things, a builders dream.

I think much like you that AI is and will just continue to destroy the economy! At least I got to sell a house and make a profit--stash it away for when the big AI market crash happens (hopefully not a 2030 great depression tho). As then it's a down market and buying stocks, bitcoin and houses is always cheaper.

Because they can hire some "prompt engineer" to "steer the AI" for $30-50k instead of $150-$250k.

But..cost = revenue."

That is Karl Marx's Labor theory of value that has been completely disproven.

You don't charge what it costs to build something, you charge the maximum the customer is willing to pay.

- First, the LTV was not Marx's idea. Adam Smith held the same view, as did many many others during this era. Marx refined this idea, but there's nothing about your point that is unique to his version of it.

- Second, while LTV is not widely used today, this is not because it was "completely disproven" (can you cite anything to back this claim up?). It is because economics shifted to a different paradigm based on marginal utility. These two frameworks operate at different levels of abstraction and address different aspects of the price of goods. There is actually empirical evidence of a correlation between the cost of a good and the cost of the labour, at an aggregate level.

- Third, Marx explicitly differentiated between _value_ and _price_. LTV deals with value exclusively (in other words, what happens when externalities impacting price are accounted for). He would have had no issue accepting that externalities impacting supply and demand would impact price.

The final irony of your comment is that the commenter's claim that you are incorrectly analysing is actually also fully defensible under your (presumably) neoclassical view of economics. In competitive markets, reduced production costs lead to reduced equilibrium prices as competitors undercut each other. The proposition that in the long run, under competition, price tends toward cost is a standard result in microeconomics. The idea that "you charge the maximum the customer is willing to pay" only holds without qualification in monopoly or monopolistic competition with strong differentiation, which are precisely the conditions that increased software supply would erode.

Here's the other edge of that sword. A couple back-end devs in my department vibe-coded up a standard AI-tailwind front-end of their vision of revamping our entire platform at once, which is completely at odds with the modular approach that most of the team wants to take, and would involve building out a whole system based around one concrete app and 4 vaporware future maybe apps.

And of course the higher-ups are like “But this is halfway done! With AI we can build things in 2 weeks that used to six months! Let’s just build everything now!” Nevermind that we don’t even have the requirements now, and nailing those down is the hardest part of the whole project. But the higher-ups never live through that grind.

It was missing years of backend and had maybe 1/20th feature parity with what we already had and it would have, in hindsight, been literally impossible to implement some of the things we would need in the future if we had went down that path. But they were amazed by this flashy new thing that devs made in a weekend that looked great but was actually a disaster.

I fail to see how this is any different than what people are complaining about with vibe coded LLM stuff a decade and a half later now? This was always being done and will continue to be done; it's not a new problem.

If it isn't a product that needs to solve problems reliably over time then it was kind of silly to use a DBA that cost twice the Backend engineer and only handled the data niche. We progressed from there or regressed from there depending on why we are developing software.

AI will have to take a different direction.

My worry is that any idiot can prompt themselves to _bad_ software, and the differentiator is in having the right experience to prompt to _good_ software (which I believe is also possible!). As a very seasoned engineer, I don't feel personally rugpulled by LLM generated code in any way; I feel that it's a huge force multiplier for me.

Where my concern about LLM generated software comes in is much more existential: how do we train people who know the difference between bad software and good software in the future? What I've seen is a pattern where experienced engineers are excellent at steering AI to make themselves multiples more effective, and junior engineers are replacing their previous sloppy output with ten times their previous sloppy output.

For short-sighted management, this is all desirable since the sloppy output looks nice in the short term, and overall, many organizations strategically think they are pointed in the right direction doing this and are happy to downsize blaming "AI." And, for places where this never really mattered (like "make my small business landing page,") this is an complete upheaval, without a doubt.

My concern is basically: what will we do long term to get people from one end to another without the organic learning process that comes from having sloppy output curated and improved with a human touch by more senior engineers, and without an economic structure which allows "junior" engineers to subsidize themselves with low-end work while they learn? I worry greatly that in 5-10 years many organizations will end up with 10x larger balls of "legacy" garbage and 10x fewer knowledgeable people to fix it. For an experienced engineer I actually think this is a great career outlook and I can't understand the rug pull take at all; I think that today's strong and experienced engineer will be command a high amount of money and prestige in five years as the bottom drops out of software. From a "global outcomes" perspective this seems terrible, though, and I'm not quite sure what the solution is.

It was a sobering moment for me when I sat down to look at the places I have worked for over my career of 20-odd years. The correlation between high quality code and economic performance was not just non-existing, it was almost negative. As in: whenever I have worked at a place where engineering felt like a true priority, tech debt was well managed, principles followed, that place was not making any money.

I am not saying that this is a general rule, of course there are many places that perform well and have solid engineering. But what I am saying is that this short-sighted management might not be acting as irrationally as we prefer to think.

But, I have definitely seen failure due to persistent technical mistakes, as well, especially when combined with human factors. There’s a particularly deep spiral that comes from “our technical leadership made poor choices or left, we don’t know what to invest in strategically so we keep spending money on attempted refactors, reorgs, or rewrites that don’t add more value, and now nobody can fix or maintain the core product and customers are noticing;” I think that at least two companies I’ve worked at have had this spiral materially affect their stock price.

I think that generative coding can both help and hurt along this axis, but by and large I have not seen LLMs be promising at this kind of executive function (ie - “our aging codebase is getting hard to maintain, what do we need to do to ensure that it doesn’t erode our ability to compete”).

If you adequately specify what you want, then LLM's today are perfectly capable to produce code of a quality exceeding most humans.

But what has been going on is that many of the details of architecture and code have been implied as "good practice" or "experience" because it is time consuming to write a good specification, partly because you need to first work out exactly what you want.

1. We'll train the LLMs not to make sloppy code.

2. We'll come up with better techinques to make guardrails to help

Making up examples:

* right now, lots of people code with no tests. LLMs do better with tests. So, training LLMs to make new and better tests.

* right now, many things are left untested because it's work to build the infrastructure to test them. Now we have LLMs to help us build that infrustructure so we can use it make better tests for LLMs.

* ...?

It may look the same, but it isn't the same.

In fact if you took the time to truly learn how to do pure agentic coding (not vibe coding) you would realize as a principal engineer you have an advantage over engineers with less experience.

The more war stories, the more generalist experience, the more you can help shape the llm to make really good code and while retaining control of every line.

This is an unprecedented opportunity for experienced devs to use their hard won experience to level themselves up to the equivalence of a full team of google devs.

What I want when I'm coding, especially on open source side projects, is to retain copyright licensing over every line (cleanly, without lying about anything).

Whoops!

[1] https://cloud.google.com/blog/products/ai-machine-learning/p...

[2] https://www.anthropic.com/news/expanded-legal-protections-ap...

"We are going to use our deep billionaire pockets to squash the fucker artist who dares identify something of his that we stole, that made its way into your output ..."

For starters, because of the western values of giving credit.

We have diseases named after people, never mind inventions and ideas.

Plagiarism is kick-out-of-school grade academic misconduct, whereby you are pretending that someone's work (and the ability it implies) is your own.

> The only sort of property that actually exists is real and tangible.

Remember, I'm talking about works that are free to redistribute, use and even modify. Or in other cases, that the users to whom a compiled work is distributed have access to the buildable source code.

The authors put their names on it, and terms which says that their notices are to be preserved when copies are made.

This isn't good enough for the Altmans and Amodeis of the world.

> it's an issue due to the effect that students won't learn well if they just copy everything

... and fraudulently obtain professional licensing, and use that to cause harm: medical malpractice, unsafe engineering.

It is fraud.

I am also not sure why you keep bringing up Altman et al, I really don't give a shit what they are talking about, that is not what I am discussing. You for some reason keep trying to inject your views on these people when they are not relevant to the points I made which are about the theoretical concepts of machine learning and training, and its intersection with intellectual property. I am not interested in your opinions on these people, and they are not the only ones who stand to benefit from democratization of AI models and publishing of weights for the public.

Anyway, I think we both fundamentally have different views on the freedom of information and the fallacious nature of IP that cannot be changed online so I will bid you a good day and won't continue this conversation further, as I don't think it's productive for either of us.

In the case of copyleft licenses like GPL, IP is applied in a way to ensure that users have the code.

These things are taken away when the code is laundered through AI.

Photography started displacing painting as a form of portraiture, but displacing a technique is not the same thing as appropriating the work itself.

Of course the AI robber barons would that it be so, but it must not be and should not be.

Training gobbles up works in their entirety, verbatim.

Fair use of the verbatim words of a written work requires the excerpt to be small.

Fair use also usually requires attribution, which is missing.

Transformative works like parodies are also fair use, but the LLM isn't transformative int his sense; it's strawman transformative like a meat grinder.

Parodies use the structure of something existing, as a vehicle for original thought which is why they are protected from copyright claims by the authors of whatever is pariodied.

I'm not even talking about big corporations with proprietary models, in fact I oppose their not being open source or weight, I want more open models not fewer as that at least democratizes the value of LLMs. The worst case is having copyright hawks allowing regulatory capture by big AI corps by pushing regulations about licensing content, which, of course, no open model company will be able to afford in the future. I find that infinitely worse than having more lax copyright laws, where only a few corporations can tell you want to think via usage of their LLMs.

Lastly, no one can tell me from first principles why LLM training is bad, on the copyright side, other than, it just is, because copyright law dictates it so. Perhaps copyright law is what needs to be abolished, not LLMs.

> Perhaps copyright law is what needs to be abolished, not LLMs.

Sure, now that it's inconvenient for some billionaires --- who themselves have nothing to protect, because everything they offer is a service the user can only access through the network, while they have a subscription.

No, not just now, since forever. I suppose Stallman being right all along is about this concept. And just to be clear, I'm not a supporter of current closed source AI companies, like I said I want to see open models succeed.

As I asked above, it really does look like no one can explain why LLM training is bad, besides saying it's bad. Therefore I will continue to reject IP as a concept.

So discussions revolving about doing so less directly through training a model just add distracting details that don't matter.

If everyone did that (due to there not being any rules against that), then fewer people would write programs under free licenses. Many such developers are volunteers, whose only payment is that the work product is theirs to license how they want.

Having that taken away from us is discouraging.

We haven't done anything to deserve such a "fuck you".

No, it can't. I use claude code and AMP a lot, and yet, unless I pay attention, it easily generate bad code, introduces regressions while trying to fix bugs, get stuck in suboptimal ideas. Modularity is usually terrible, 50 year ideas like cohesion and coupling are, by the very nature of it, mostly ignored except in the most formal rigid ways of mimicry introduced by post-training.

Coding agents are wonderful tools, but people who think they can create and mantain complex systems by themselves are not using them in an optmal way. They are being lazy, or they lack software engineering knowledge and can't see the issues, and in that case they should be using the time saved by coding agents to read hard stuff and elevate their technique.

I grew up without a mentor and my understanding of software stalled at certain points. When I couldn’t get a particular os API to work, in Google and stack overflow didn’t exist, and I had no one around me to ask. I wrote programs for years by just working around it.

After decades writing software I have done my best to be a mentor to those new to the field. My specialty is the ability to help people understand the technology they’re using, I’ve helped juniors understand and fix linker errors, engineers understand ARP poisoning, high school kids debug their robots. I’ve really enjoyed giving back.

But today, pretty much anyone except for a middle schooler could type their problems into a ChatGPT and get a more direct answer that I would be able to give. No one particularly needs mentorship as long as they know how to use an LLM correctly.

That said, I still feel strongly about mentorship though. It's just that you can spend your quality time with the busy person on higher-level things, like relationship building, rather than more basic questions.

Can't just offload all the hard things to the AI and let your brain waste away. There's a reason brain is equated to a muscle - you have to actively use it to grow it (not physically in size, obviously).

But I can tell you that, just like with most things in life, this is yet another area where we are increasingly getting to do just the things we WANT to do (like think about code or features and have it appear, pixel pushing, smoothing out the actual UX, porting to faster languages) and not have to do things most people don't want to do, like drudgery (writing tests, formatting code, refactoring manually, updating documentation, manually moving tickets around like a caveman). Or to use a non tech example, having to spend hours fixing word document formatting.

So we're getting more spoiled. For example, kids have never waited for a table at a restaurant for more than 20 mins (which most people used to do all the time before abundant food delivery or reservation systems). Not that we ever enjoyed it, but learning to be bored, learning to not just get instant gratification is something that's happening all over in life.

Now it's happening even with work. So I honestly don't know how it'll affect society.

The "as long as they know how..." is doing a lot of work there.

I expect developers with mentors who help give them the grounding they need to ask questions will get there a whole lot faster than developers without.

No they can't. They think they can, but they will still need to put in the elbow grease to get it done right.

But, in my case (also decades of experience), I have had to reconcile with the fact that I'll need to put down the quill pen, and learn to use a typewriter. The creativity, ideas, and obsession with Quality are still all mine, but the execution is something that I can delegate.

LLMs don't always produce correct code - sometimes it's subtly wrong and it takes an expert to notice the mistake(s).

It was never about writing the code—anyone can do that, students in college, junior engineers…

Experience is being able to recognize crap code when you see it, recognizing blind alleys long before days or weeks are invested heading down them. Creating an elegant API, a well structured (and well-organized) framework… Keeping it as simple as possible that just gets the job done. Designing the code-base in a way that anticipates expansion…

I've never felt the least bit threatened by LLMs.

Now if management sees it differently and experienced engineers are losing their jobs to LLMs, that's a tragedy. (Myself, I just retired a few years ago so I confess to no longer having a dog I this race.)

Retired, I have continued to code, and have used Claude to vibe code a number of projects—initially I dod so out of curiosity as to how good LLM are, and then to handle things like SwiftUI that I am hesitant to have to learn.

It's true then that I am not in a position of employment where I have to consider a performance review, pleasing my boss or impressing my coworkers. I don't doubt that would color my perception.

But speaking as someone who has used LLMs to code, while they impress me, again, I don't feel the threat. As others have pointed out in past threads here on HN, on blogs, LLMs feel like junior engineers. To be sure they have a lot of "facts" but they seem to lack… (thinking of a good word) insight? Foresight?

And this too is how I have felt as I was aging-out of my career and watched clever, junior engineers come on board. The newness, like Swift, was easy for them. (They no doubt have rushed headlong into Swift UI and have mastered it.) Never though did I feel threatened by them though.

The career itself, I have found, does in fact care little for "grey beards". I felt by age 50 I was being kind of… disregarded by the younger engineers. (It was too bad, I thought, because I had hoped that on my way out of the profession I might act more as mentor than coder. C'est la vie!)

But for all the new engineer's energy and eagerness, I was comfortable instead with my own sense of confidence and clarity that came from just having been around the block a few times.

Feel free to disregard my thoughts on LLMs and the degree to which they are threatening the industry. They may well be an existential threat. But, with junior engineers as also a kind of foil, I can only say that I still feel there is value in my experience and I don't disparage it.

When you have had to tackle dozens of frameworks/libraries/API over the years, you get to where you find you like this one, dislike that one.

Get/Set, Get/Set… The symmetry is good…

Calling convention is to pass a dictionary: all the params are keys. Extensible, sure, but not very self-documenting, kind of baroque?

An API that is almost entirely call-backs. Hard to wrap your head around, but seems to be pretty flexible… How better to write a parser API anyway?

(You get the idea.)

And as you design apps/frameworks yourself, then have to go through several cycles of adding features, refactoring, you start to think differently about structuring apps/frameworks that make the inevitable future work easier. Perhaps you break the features of a monolithic app into libraries/services…

None of this is novel, it's just that doing enough of it, putting in the sweat and hours, screwing up a number of times) is where "taste" (insight?) comes from.

It's no different from anything else.

Perhaps the best way to accelerate the above though is to give a junior dev ownership of an app (or if that is too big of a bite, then a piece of a thing).

"We need an image cache," you say to them. And then it's theirs.

They whiteboard it, they prototype it, they write it, they fix the bugs, they maintain it, they extend it. If they have to rewrite it a few times over the course of its lifetime (until it moves into maintenance mode), that's fine. It's exactly how they'll learn.

But it takes time.

> Nobody tells this to people who are beginners, and I really wish somebody had told this to me.

> All of us who do creative work, we get into it because we have good taste. But it's like there is this gap. For the first couple years that you're making stuff, what you're making isn't so good. It’s not that great. It’s trying to be good, it has ambition to be good, but it’s not that good.

> But your taste, the thing that got you into the game, is still killer. And your taste is good enough that you can tell that what you're making is kind of a disappointment to you. A lot of people never get past that phase. They quit.

> Everybody I know who does interesting, creative work they went through years where they had really good taste and they could tell that what they were making wasn't as good as they wanted it to be. They knew it fell short. Everybody goes through that.

> And if you are just starting out or if you are still in this phase, you gotta know its normal and the most important thing you can do is do a lot of work. Do a huge volume of work. Put yourself on a deadline so that every week or every month you know you're going to finish one story. It is only by going through a volume of work that you're going to catch up and close that gap. And the work you're making will be as good as your ambitions.

> I took longer to figure out how to do this than anyone I’ve ever met. It takes awhile. It’s gonna take you a while. It’s normal to take a while. You just have to fight your way through that.

> —Ira Glass

I'm excited to work with AI. Why? Because it magnifies the thing I do well: Make technical decisions. Coding is ONE place I do that, but architecture, debugging etc. All use that same skill. Making good technical decisions.

And if you can make good choices, AI is a MEGA force multiplier. You just have to be willing to let go of the reins a hair.

Any suggestions to overcome this deficit in design experience? My best guess is to read some texts on code design or alternatively get a job at a place to learn design in practice. Mainly learning javascript and web app development at the moment.

*Who has had a career in a previous field, and doesn't necessarily think that learning programming with lead to another career (and is okay with that).

I can tell you: Your problems are a layer higher than you think.

Coding, Architecture, etc. Those get the face time. Process, and Discipline, and where the money is made and lost in AI.

To give a minor example: My first attempt at a major project with AI failed HORRIBLY. But I stepped back and figured out why. What short-comings did my approach have, what short-comings did the AI have. Root Cause Analysis.

Next day I sat down with the AI and developed a PLAN of what to do. Yes, a day spent on a plan.

Then we executed the plan. (or it did and I kept it on track, and fixed problems in the plan as things happened.) On the third day I'd completed a VERY complex task. I mean STUPIDLY complex, something I knew WHAT I wanted to do, and roughly how, but not the exact details, and not at the level to implement it. I'm sure 1-2 weeks of research could have taught me. Or I could let the AI do it.

... And that formed my style of working with AI.

If you need a mentor pop in the Svalboard discord, and join #sval-dev. You should be able to figure out who I am.

I guess it sounds like a bit of a trap in some ways. Those without experience (noobs like me) are gifted a tool that can suggest to them solutions or strategies that you'd typically come across only after many years working/learning peeling back the onion. You don't really have the taste or canniness to know how best to use those tools without the experience, which you typically gain by working and gaining the experience!

I cycle regularly and it actually reminds me a little of the many people who've taken to ebikes recently. I'm generally in favour of ebikes for adding another mode of non-car transport out there but I definitely see a lot of newish ebikers attempting manouveres that only experienced cyclists would do, because you wouldn't do it if you didn't have the strength to get up to speed and have the handling. But ebikers are able to bypass the strength and skills training and get themselves in somewhat dangerous situations that you'd only find yourself in if you were a pretty strong cyclist and had the bike handling skills to match your speed.

A bit of a tangent there!

I do like spending time planning. Problem is as a relative beginner there's only so much you can add to your plan when you're still learning how to do so much. I do use AI to help advise on my strategies, toolkit approaches rather than do the code though. It does send me really inappropriate solutions though quite regularly that a human would (hopefully) avoid because the human understands the wider context or at least will ask questions.

Story: I'm dev for about 20 years. First time I had totally the same felling when desktop ui fading away in favor of html. I missed beauty of c# winforms controls with all their alignment and properties. My experience felt irrelevant anymore. Asp.net (framework which were sold as "web for backed developers") looked like evil joke.

Next time it have happened with the raise of clouds. So were all my lovely crafted bash scripts and notes about unix command irrelevant? This time however that was not that personal for me.

Next time - fall of scala as a primary language in big data and its replacement with python. This time it was pretty routine.

Oh and data bases... how many times I heard that rdbms is obsolete and everybody should use mongo/redis/clickhouse?

So learn new things and carry on. Understanding how "obsolete" things works helps a lot to avoid silly mistake especially in situation when world literally reinvent bicycle

If, as a principal engineer, you were performing basic work that can easily be replicated by an LLM, then you were wasted and mistasked.

Firstly, high-end engineers should be working on the hard work underlying advances in operating systems, compilers, databases, etc. Claude currently couldn't write competitive versions of Linux, GCC (as recently demonstrated), BigQuery, or Postgres.

Secondly, and probably more importantly, LLMs are good at doing work in fields already discovered and demonstrated by humans, but there's little evidence of them being able to make intuitive or innovative leaps forwards. (You can't just prompt Claude to "create a super-intelligent general AI"). To see the need for advances (in almost any field) and to make the leaps of innovation or understanding needed to achieve those advances still takes smart (+/- experienced) humans in 2026. And it's humans, not LLMs, that will make LLMs (or whatever comes after) better.

Thought experiment: imagine training a version of Claude, only all information (history, myriad research, tutorials, YouTube takes and videos, code for v1, v2, etc.) related to LLMs is removed from the training data. Then take that version and prompt it to create an LLM. What would happen?

Even regarding "chase something complex and difficult", there are currently only so many needs for that, so I think any given person is justified fearing they won't be picked. It may be several years between AI eating all the CRUD work from principal down, and when it expands the next generation of complex work on robotics or whatever.

Also, to speak on something I'm even less qualified – the economy feels weak, so I don't have a lot of hope for either businesses or entrepreneurs to say "Let's just start new lines of business now that one person can do what used to take a whole team." The businesses are going to pocket the safe extra profits, and too many entrepreneurs are not going to find a foothold regardless how fast they can code.

I'm having a lot of fun with AI. Any idiot can't prompt their way to the same software I can write. Not yet anyways.

My experience is the opposite. Those with a passion for the field and the ability to dig deeply into systems are really excited right now (literally all that power just waiting to be guided to do good...and oh does it need guidance!). Those who were just going through the motions and punching a clock are pretty unmotivated and getting ready to exit.

Sometimes I dream about being laid off from my FAANG job so I have some time to use this power in more interesting than I'm doing at work (although I already get to use it in fairly interesting ways in my job).

In my experience they are mostly the subset of engineers who enjoyed coding in and of itself and ——in some cases—— without concern for the end product.

What's missing from this is that iconic phrase that all the AI fans love to use: "I'm just having fun!"

This AI craze reminds me of a friend. He was always artistic but because of the way life goes he never really had opportunity to actively pursue art and drawing skills. When AI first came out, and specifically MidJourney he was super excited about it, used it a lot to make tons and tons of pictures for everything that his mind could think of. However, after awhile this excitement waned and he realized that he didn't actually learn anything at all. At that point he decided to find some time and spend more time practicing drawing to be able to make things by himself with his own skills, not by some chip on the other side of the world and he greatly improved in the past couple of years.

So, AI can certainly help create all the "fun!!!" projects for people who just want to see the end result, but in the end would they actually learn anything?

Playing with Claude, if you tell it to do something, it'll produce something. Sometimes it's output is ok, sometimes it's not.

I find I need to iterate with Claude, tell it no, tell it how to improve it's solution or do something in a different way. It's kind of like speed running iterating over my ideas without spending a few hours doing it manually, writing lots of code then deleting it to end with my final solution.

If I had no prior coding knowledge i'd go with what ever the LLM gave me and end up with poor quality applications.

Knowing how to code gives you the advantage still using an LLM. Saying that, i'm pessimistic what my future holds as an older software engineer starting to find age/experince is an issue when an employer can pay someone less with less experience to churn out code with prompts when a lot of time the industry lives by "it's good enough".