We all have different thresholds for what is acceptable, and our roles as engineers typically reflect that preference. I can grind on a single piece of code for hours, iterating over and over until I like the way it works, the parameter names, etc.
Other people do not see the value in that whatsoever, and something that works is good enough. We both are valuable in different ways.
Also, theres the pace of advancement of the models. Many people formed their opinions last year, and the landscape has changed a lot. There’s also some effort requires in honing your skill using them. The “default” output is average quality, but with some coaxing higher quality output is easily attained.
I’m happy people are skeptical though, there are a lot of things that do require deep thought, connecting ideas in new ways, etc., and LLMs aren’t good at that in my experience.
I think there are multiple dimensions that people fall on regarding the issue and it's leading to a divide based on where everyone falls on those dimensions.
Quality and standards are probably in there but I think risk-tolerance/aversion could be behind some how you look at quality and standards. If you're high on risk-taking, you might be more likely to forego verifying all LLM-generated code, whereas if you're very risk-averse, you're going to want to go over every line of code to make sure it works just right for fear of anything blowing up.
Desire for control is probably related, too. If you desire more control in how something is achieved, you probably aren't going to like a machine doing a lot of the thinking for you.
I am a high quality/craftsmanship person. I like coding and puzzling. I am highly skilled in functional leaning object oriented deconstruction and systems design. I'm also pretty risk averse.
I also have always believed that you should always be "sharpening your axe". For things like Java delelopment or things where I couldn't use a concise syntax would make extensive use of dynamic templating in my IDE. Want a builder pattern, bam, auto-generated.
Now when LLMs came out they really took this to another level. I'm still working on the problems.. even when I'm not writing the lines of code. I'm decomposing the problems.. I'm looking at (or now debating with the AI) what is the best algorithm for something.
It is incredibly powerful.. and I still care about the structure.. I still care about the "flow" of the code.. how the seams line up. I still care about how extensible and flexible it is for extension (based on where I think the business or problem is going).
At the same time.. I definately can tell you, I don't like migrating projects from Tensorflow v.X to Tenserflow v.Y.
That line always makes me laugh. There’s only 2 points of an algorithm, domain correctness and technical performance. For the first, you need to step out of the code. And for the second you need proofs. Not sure what is there to debate about.
There may be other considerations as well -- licensing terms, resources, etc.
I, for example, would claim to be rather risk-tolerant, but I (typically) don't like AI-generated code.
The solution to the paradox this creates if one considers the model of your post is simple:
- I deeply love highly elegant code, which the AI models do not generate.
- I cannot stand people (and AIs) bullshitting me; this makes me furious. I thus have an insanely low tolerance for conmen (and conwomen and conAIs).
Sometimes you need something to be extremely robust and fool-proof, and iterating for hours/days/weeks and even months might make sense. Things that are related to security or money are good examples.
Other times, it's much more preferable to put something in front of users that works so that they start getting value from it quickly and provide feedback that can inform the iterative improvements.
And sometimes you don't need to iterate at all. Good enough is good enough. Ship it and forget about it.
I don't buy that AI users favor any particular approach. You can use AI to ship fast, or you can use it to test, critique, refactor and optimize your code to hell and back until it meets the required quality and standards.
People have been saying this every year for the last 3 years. It hasn't been true before, and it isn't true now. The models haven't actually gotten smarter, they still don't actually understand a thing, and they still routinely make basic syntax and logic errors. Yes, even (insert your model of choice here).
The truth is that there just isn't any juice to squeeze in this tech. There are a lot of people eagerly trying to get on board the hype train, but the tech doesn't work and there's no sign in sight that it ever will.
Now I can ask an agent to code a full feature and it has been handling it more often than not, often getting almost all of the way there with just a few paragraphs of description.
Having said that, I’ve changed the way I work too: more focused chunks of work with tight descriptions and sample data and it’s like having a 2nd brain.
I'm curious about this discrepancy too. I assume that you're being facetious and the discrepancy is with people's perceptions of AI’s capabilities or usefulness or whatever subjective metric. Some, myself included, seem to perceive it as basically useless, while others, yourself included, seem to imply that it's at a level where it genuinely replaces competent coders.
If the discrepancy were small, it could just be chalked up to the metric being subjective. But it seems to be like night and day. A difference of orders of magnitude. I wanna know what's going on there.
To this day I remember being delighted by Perl on a regular basis. I wasn't concerned with the aesthetics of it, though I was aware it was considered inscrutable and that I could read & write it filled me with pride. So yea, programming Perl was delightful.
Emacs vs. vi. Command-line editor vs. IDE. IntelliJ vs. VS Code. I could do like twenty more of these: dev teams have always split on technology choices.
But, all of those were rational separations. Emacs and vi, IntelliJ and VS Code ... they're all viable options, so they boil down to subjective preference. By definition, anything subjective will vary between different humans.
What makes AI different to me is the fear. Nobody decided not to use emacs because they were afraid it was going to take their job ... but a huge portion of the anti-AI crowd is motivated by irrational fear, related to that concern.
LLMs are not a tool like an editor or an IDE, they make up code in an unpredictable way; I can't see how anyone who enjoyed software development could like that.
That's kind of how this is starting to feel to me, like I'm turning more into a project manager that has to review code from a team of juniors, when it comes to A.I. Although those juniors are now starting to show that they're more capable than even I am, especially when it comes to speed.
Now given that LLMs are known to not produce 100% correct code you should review every single line. Now the production rate of LLMs is so high that it becomes very hard to really read and understand every line of the output. While at the same time you are gradually losing the ability to understand everything because you stopped actively coding. And at the same time there are others in your team who aren't that diligent adding more to the crufty code base.
What is this if not a recipe for disaster?
Yeah, this is a concern. I remember when I took a break from coding to work as a video game producer for a couple of years and I felt like my ability to code was atrophying and that drove me nuts.
Now I'm not so sure. There's just so much dumb garbage that's accumulated around coding nowadays, especially with web dev. So much time just gluing together or puzzling out how to get APIs or libraries to works and play nice together.
It's not like back in the days where it was relatively simple with PHP and HTML, when I first started. Much less you could do back then, sure, but expectations were a lot lower back then as well.
I might just content myself with doing Sudoku or playing/designing board games to help keep that brain plasticity going, and stop fighting so hard to understand all this junk. Or finally figure out how to get half-decent at shader math or something, at least that seems less trivial and ephemeral.
And then it changes every 6 months or goes in circles - and I suppose now we just gave up and are letting LLMs YOLO code out the door with whatever works.
Like I remember learning all about Redux Sagas, and then suddenly the whole concept is gone, but also I'm not actually particularly clear on what replaced it (might be time to go back to that well since I need to write a web interface soon again).
I'm afraid my inherent laziness will lead to something like that if I don't have to do it for my job.
I love the exciting ideation phase, I love putting together the puzzle that makes the product work, and I also take pride in the craft.
Yes, those things should have been automated long ago, but they weren't, and now with coding agents much of them are.
The two emotions I personally feel are fear and excitement. Fear that the machines will soon replace me. Excitement about the things I can build now and the opportunities I’m racing towards. I can’t say it’s the most enjoyable experience. The combo is hellish on sleep. But the excitement balances things out a bit.
Maybe I’d feel a sense of sadness if I didn’t feel such urgency to try and ride this tsunami instead of being totally swept away by it.
Sure, they could have come up with those optimizations without AI... but they didn't. What's your theory for why that is?
If you're the size of Shopify that represents a huge saving in server costs and improved customer-facing latency.
This percentage is meaningless on its own. It’s 4 ms shaved off a 7 ms process. You would need to time a whole flow (and I believe databases would add a lot to it, especially with network latency) and figure out how significant the performance improvement is actually. And that without considering if the code changes is not conflicting with some architectural change that is being planned.
In fact it being easier to get them out there I might care less that they should be marketable and have a chance to make serious money, as opposed to when I was sinking hundreds of hours into them and second guessing what direction I should take the games to make them better all the time.
The art wasn't the problem (the art wasn't great, but I could make functional art at least), it was finding the time and energy and focus to see them through to completion (focus has always been a problem for me, but it's been even worse now that I'm an adult with other responsibilities).
And that hasn't always been the issue, I did release about a dozen games back in the day (although I haven't in quite a few years at this point).
Of course someone may say 'well that's slop then', and yeah, maybe by your standards, sure. These games aren't and never were going to be the next Slay The Spire or Balatro. But people can and do enjoy playing them, and not every game needs to be the next big hit to be worth putting out into the world, just like not every book needs to be the next 1984 or Great Gatsby.
Money, opportunity, status. It is all status games. Think of it as a nuclear war on old order and new players trying to take the niche. Or maybe commies killing whites and taking over Russia?
For me specifically it means two products, one that is something I have been working on for a long time, well before the Claude Code era, and another that is more of a passion project in the music space. Both have been vastly accelerated by these tools. The reason I say “racing” is because I suspect there are competitors in both spaces who are also making great progress because of these tools, so I feel this intense pressure to get to launch day, especially for the first project.
And yes it is very frenetic, and it’s certainly taking a toll on me. I’m self-employed, with a family to support, and I’m deeply worried about where this is all going, which is also fuelling this intense drive.
A few years ago I felt secure in my expertise and confident of my economic future. Not any more. In all honesty, I would happily trade the fear and excitement I feel now for the confidence and contentment I felt then. I certainly slept better. But that’s not the world we live in. I don’t know if my attempts to create a more secure future will work, but at least I will be able to say I tried as hard as I was able.
Then with AWS our infra was moving to proprietary platforms. Now our dev tools are moving to expensive proprietary platforms.
Combined with widespread enshittification, we've handed nearly everything to the tech bros now.
What opportunities? Anything you spend effort over, like PMF and discovery, etc... I can now clone with a few bucks of Claude Code and charge less than you for the same product, at the same quality level :-/
Where is the opportunity here? Technology and knowledge used to be the moat a startup or bootstrapped individual could use to produce a sustainable business.
Why exactly are you excited about producing something that can be cloned for less cost than it took you? Especially as the quality will be almost exactly the same?
If I say "doing $FOO is a losing proposition", and I believe what I say, why on earth would I then move on to actually doing $FOO?
I am pointing out that there is nothing to be gained by joining this recursive race - anything you produce using LLMs I can clone using LLMs, but anything I produce using LLMs can be cloned by someone else, using LLMs.
Why would you assume that I want to insert myself into this recursively descending race to the bottom?
Because they don't yet know that it is a losing proposition?
> and if nobody does it, isn't your original assertion false?
False dichotomy, it isn't an all-or-nothing scenario like you present, it's the fact that there are enough cloners to make every race a race to the bottom in a matter of days.
> Any argument relying on everyone cloning things with LLMs doesn't work if everyone doesn't clone things with LLMs.
Only if you believe your false dichotomy.
If I reflect for a moment about why I personally got into tech, I can find at least a few different reasons:
- because I like solving problems. It's sad that the specific types of problems I used to do are gone, but it's exciting that there are new ones.
- because I like using my skills to help other people. It's sad that one specific way I could do that is now less effective, but it's exciting that I can use my knowledge in new ways.
- because I like doing something where I can personally make a difference. Again, it cuts both ways.
I'm sure most people would cite similar examples.
The "make-it-go" people couldn't make anything go back then either. They build ridiculous unmaintainable code with or without AI. Since they are cowboys that don't know what they're doing, they play the blame game and kiss a ton of ass.
The "craft-lovers" got in the way just as much with their endless yak shaving. They now embrace AI because they were just using "craft" as an excuse for why they didn't know what they were doing. They might be slightly more productive now only because they can argue with themselves instead of the rest of the team.
The more experienced and pragmatic people have always been forced to pick up the slack. If they have a say, they will keep scope narrow for the other two groups so they don't cause much damage. Their use of AI is largely limited to google searches like it always was.
(I'll admit, though, that this also smells to me a bit too much like introvert/extrovert, or INTP/INTJ/etc so maybe I'm being reflexively rejective)
I, for one, remember when I could crash Netscape Navigator by using CSS too hard (i.e. at all) or trying to make a thing move 10px with DHTML. But I kept trying to make browser to thing.
Definitely not. Based on my observations from a career as an open source and agency developer it was obvious at a glance which of these camps any given developer lived in by their code quality. With few exceptions make-it-go types tended to produce brittle, hacky, short-sighted work that had a tendency to produce as many or more problems than it solved, and on more than one occasion I've seen developers of this stripe lose commit access to FOSS projects as a result of the low quality of their contributions.
"nobody got into it because programming in Perl was somehow aesthetically delightful."
Compared to trying to get stuff accomplished in C Perl was an absolute dream to work with and many devs I knew gravitated to web development specifically for their love of the added flexibility and expressiveness that Perl gave them compared to other commonly used languages at the time. Fortunately for us all language design and tooling progressed.
As an old school Perl coder, not true. Lots of people had a taste for Perl. TIMTOWTDI was sold as an actual advantage.
Perl caters to things almost nobody else does, like the way you have a negative "if" in "unless" and placing conditions after the code. So you can do things like:
frobnicate() unless ($skip_frobnicating);
if (!$skip_frobnicating) frobnicate();
The bigger problem with Perl IMO is that it started as a great idea and didn't evolve far enough -- a bunch of things had to be tacked on, and everyone tacked on them slightly differently for no real benefit, resulting in codebases that can be terribly fragile for no good reason and no benefit.
I've used Claude Code to build a few big things at work; I ask it questions ("where does this happen", "we have problem X, give me 3 potential causes", etc); I have it review things before I post PRs; our code review bot finds real heisenbugs. I have mixed success with all of this, but even so I find it overall useful. I'd be irritated if some place I worked, current or present, told me I couldn't use Claude Code or the like.
That said, I've not gotten it to be useful in:
- building entire, complex features in brownfield projects
- solving systemic bugs
- system design/evolution
- feature/product design and planning
- replacing senior engineer code review
It will confidently tell you it's done these things, but when you actually force yourself through the mental slog of reviewing its output, you'll realize it's failed (you also have to be an expert to perform this analysis). Now, maybe it fails in an acceptable way; maybe only slight revision is required; maybe it one-shots the change and verifying success isn't a big mental slog. Those are the good cases. More annoying are the times it fails totally and obviously, but the real nightmares are when it fails totally, yet imperceptibly. It also sometimes can do (some of) these things! But it's inconsistent, such that its successes largely serve to lower your guard against its failures.
And the mental slog is real. The artifacts you have to produce/review/ensure the model adheres to are ponderous. The code generated is ponderous. Code review is even more tedious because there's no human mind behind the code, so you can't build a mental model of the author. Getting a codegen model to revise its work or take a different approach is very hit or miss. Revising the code yourself requires reading thousands and thousands of lines of generated code--again with no human behind it--and building a mental model of what's happening before you can effectively work, and that process is time-consuming and exhausting.
I'm also concerned about the second-order effects. Because switching into the often-required deep mental focus is very difficult (borderline painful), I've seen many, many people reach for LLMs in those moments instead, first a little, then entirely. I've watched people copy/paste API docs into Gemini prompts to explain them. I've watched people unable to find syntax errors in code and paste it into ChatGPT to fix it. I'm confident I'm not the only person who's observed this, and it's a little maddening it's not getting more play.
---
I'm not saying SWEs don't fail in similar ways. I've approved--and authored--human PRs that had insidious flaws with real consequences. I've been asked to "review" PRs pre-ChatGPT that were 10x the size they needed to be. I've seen people plagiarize code, or just copy/paste Stack Overflow constantly. The difference is we build process around these risks, everything from coding patterns, PR size limits, type systems, firing people, borderline ludicrous amounts of unit tests, CI/CD, design docs, staging environments, red/green deploys, QA lists, etc.
I hate all of it! It's a constant reminder of my flaws and it slows down mean time to dopamine squirt of released code. I'd be the first person to give all this shit the axe. I would love to point Claude at the crushingly long list of PRs I have to review. But I can't, because it still has huge, huge flaws. Code review bots miss obvious problems, and they don't have enough context/knowledge about the system/bug/feature to perform a sufficiently comprehensive review. It would be a net time waste because we'd then have to fix a bug in prod or revise an already-deployed feature/fix--things I like even less than code review, if you can believe it.
These models cannot adequately replace humans in other parts of the SDLC. But, because pesky things like design and code review cap codegen models' velocity, our industry is "rethinking" it all, with no consideration of the models' flaws; "rethinking" here meaning "we're considering having an LLM handle all our code review, or not doing it at all". The only way to describe that is reckless disregard. It's unprofessional and unethical.
So, I think my grief isn't about "the craft". I don't think that's gone and I don't think I'd care if it were. My grief is about the humiliation of our profession, the annihilation of our standards and the betrayal of any representation we made to our users--indeed to ourselves. We deserve software systems that do what they say they do, and up until recently I really thought we were working hard to get there. I don't think that anymore; like many other things in our era (community, truth, curiosity, generosity, trust, learning, rationality, practice, compassion) it has retreated in the face of some flavor of self-interested, shallow grift. I really don't know how or why this happened, but regardless of the cause we truly are in a dark time.
“I'm also concerned about the second-order effects. Because switching into the often-required deep mental focus is very difficult (borderline painful), I've seen many, many people reach for LLMs in those moments instead, first a little, then entirely. I've watched people copy/paste API docs into Gemini prompts to explain them. I've watched people unable to find syntax errors in code and paste it into ChatGPT to fix it. I'm confident I'm not the only person who's observed this, and it's a little maddening it's not getting more play”
That’s exactly why I stopped using LLM’s. Then people turn around and say “but.. you’ll get left behind.”
Yeah, nah. I value my ability to hold concepts and reason deeply and sit in those painful moments - I’m not letting go of this conditioning that pays dividends over the long term.
It was hard to believe the same person wrote the thoughtful rest of the comment and this insulting assumption.
> Before AI, both camps were doing the same thing every day. Writing code by hand. Using the same editors, the same languages, the same pull request workflows. The craft-lovers and the make-it-go people sat next to each other, shipped the same products, looked indistinguishable. The motivation behind the work was invisible because the process was identical.
Helps explain why some people are delighted to have AI write code for them while others are unhappy that the part they enjoyed so much has been greatly reduced.
Similar note from Kellan (a clear member of the make-it-go group) in https://laughingmeme.org/2026/02/09/code-has-always-been-the... :
> That feeling of loss though can be hard to understand emotionally for people my age who entered tech because we were addicted to feeling of agency it gave us. The web was objectively awful as a technology, and genuinely amazing, and nobody got into it because programming in Perl was somehow aesthetically delightful.