Programming is Art

I'm grumbled in other comments here about definitions of art that are so vague that they're just "anything where you think a bit"

But I'd take them every time over a definition of art being "something for which there is a market". Screw that.

Most people who are good at their job understand the greatest value comes from deep understanding and internalization. Some people have been in roles where alienation from the code is institutionalized, and they've been playing that role for so long that the idea of having a strong understanding and near prescient mental map of a 2 million line codebase sounds absurd.

I think the question is less about the capabilities of LLMs and their outputs (because obviously that can change), but because they represent a workflow which codifies exactly that role I mention. You fulfill the role of a manager more than a programmer. You have a vague idea of the codebase, but lack any kind of intuitive mechanical understanding. There are many cases in which I can see an issue, identify the problem, and have it fixed faster than you could prompt an LLM to investigate. After a certain point of familiarity, all reasoning is local reasoning. That's the power of mechanical sympathy.

Many, many large shops don't give a shit. They've long since galloped into the dark woods of bureaucratic know-nothing hell. Let the LLM figure that shit out. Those shops are already optimizing for maximal output and churn, so the value proposition seems obvious. I've been smugly saying it for years, but LLMs have me saying it quite a lot more these days: "Sure looks like it sucks down there!"

Completely understand you, that's why people are starting projects like this one I found yesterday https://www.programmingmuseum.com/ , looks to me, that looking at code instead of prompts is becoming a pleasure rather than a necessity.

AI actually can take away desire to code. If what you do becomes cheap for the masses, you lose your motivation bit by bit. Also you were surrounded by art, now you are surrounded by crap. It does not kill your passion fully but drains you slowly.

I've actually been thinking a lot about this. I do absolutely 100% lament that we are possibly going to see the loss of the art of programming. Not entirely, sure; some people will do it by hand the way some people do anything by hand. But, I think we will lose a lot of it. I wouldn't compare it to painting, I would compare it to the lost art in manufacturing (something relatable in the U.S. where we've long outsourced much of it and the people who have relevant knowledge and skills are literally dying out now.)

The reason why is because for however much programming is an art, it is even moreso not. I know someone will argue about this but by and large code that looks pretty but doesn't execute or do anything meaningful rarely has much value. The most purely artistic endeavor I can think of that involves code would be something like IOCCC. Which is awesome. But, if the code didn't actually compile or run, it just simply wouldn't be interesting at all. I think there is a strong argument to be made that no matter how artistic or creative you can argue programming is, it will always be fundamentally about mechanical side of things first, then about creative expression. Insofar as code is expression (it is) it is so only secondary to being executable code.

People will continue to "fight back" against the spread of LLM text and diffusion art because we want authentic human expression and those things are simply going too far in literally replacing human expression with automated output. If creativity involving the use of tools is a spectrum, using LLMs to write text and diffusion models to generate "drawings" is almost as far to the no-creative-expression end as "simply stealing something someone else made" as you can get, because it's really similar to that.

But for code, I think ultimately nobody is coming to save programming. There will be plenty of people lamenting the loss of the art, but if it really is possible to make LLMs that can write code better than people can, then it's pretty close to over. At this point, LLMs definitely can not write code better than people can, for similar reasons to why they can't write English prose better than people can, but they are nonetheless superhuman in some ways (which, as a statement on its own, isn't that scary; I mean, strictly speaking, a calculator also is superhuman in some ways.)

Oddly enough though, despite major early success, I'm still not convinced that we have hit the tipping point where human code is guaranteed to be dead soon. Actually, I think people have been vastly over-eager to adopt AI code and many codebases are suffering because they weren't careful enough about how they did it. Not to mention, the mountains of slop that Microsoft blames all of the GitHub downtime for is mostly aspirational and a lot of it is probably going to unceremoniously fall out of maintenance, at which point there isn't really much of a reason to pick it up since if you really wanted slop you could probably spend a few million tokens just doing it yourself with a model that is better than the one the author used anyways. I do prefer GPT-5.6 Sol and GPT-6 Astra for producing code that I personally feel is more tasteful, with comments that are less stupid and redundant, but I still personally think the code sucks too much to ship as is. It's hard to square this with how absurdly powerful these tools really are. We are really having to confront the complexity of what it means to be "intelligent" or what makes code "good".

But to me, it doesn't change the likely eventuality, it only impacts the timeframe. And it does suck, because even ignoring the art of programming, it really was a pretty good career for a long time. With this career facing its potential death, I'm sitting here wondering what the hell we'll even do. Guess I hope the AI coding singularity is willing to wait for me to finish paying off my mortgage at least.