AI Coding Made Me Faster, but I Can't Code to Music Anymore

But I certainly wouldn’t assume that other people’s jobs are simple or boring just because they don’t look like yours.

The ones who know what they want to do, how it should be done, but can't really be arsed to read the man pages or API docs of all the tools required.

These people can craft a prompt (prompt engineering :P) for the LLM that gets good results pretty much directly.

LLMs are garbage in garbage out. Sometimes the statistical average is enough, sometimes you need to give it more details to use the available tools correctly.

Like the fact that `fd` has the `-exec` and `--exec-batch` parameters, there's no need to use xargs or pipes with it.

90% of what the average (or median) coder does isn't in any way novel or innovative. It's just API Glue in one form or another.

The AI knows the patterns and can replicate the same endpoints and simple queries easily.

Now you have more time to focus on the 10% that isn't just rehashing the same CRUD pattern.

Of course you need to check their work, but also the better your initial project plan and specifications are, the better the result.

For stuff with deterministic outputs it's easy to verify without reading every single line of code.

Without this it defaults to being ignorant about the trade-offs that you care about, or the relevant assumptions you're making which you think are obvious but really aren't.

The "simple stream" aspect is that each task I give to the LLM is narrowly scoped, and I don't want to put all aspects of the relevant theory that pertains just to that one narrow task into a more formal centralized doc. It's better off as an ephemeral part of the prompt that I can delete after the task is done. But I also do have more formal docs that describe the shared parts of the theory that every prompt will need access to, which is fed in as part of the normal context.

The hypothetical that we're 8x as productive but the work isn't as fun isn't "society becoming shittier".

You're right in general, but I don't think that'll save you/us from OP's statement. This is simple economic incentives at play. If AI-coding is even marginally more economically efficient (i.e. more for less) the "old way" will be swept aside at breathtaking pace (as we're seeing). The "from my cold dead hands" hand-coding crowd may be right, but they're a transitional historical footnote. Coding was always blue-collar white-collar work. No-one outside of coders will weep for what was lost.

This feels like kicking someone when they’re down! Given the current state of corporate and political America, it doesn’t look likely there will be any pressure for anything but enshittification to me. Telling people at the coal face to stay cheerful seems unlikely to help. What mechanism do you see for not giving up to actually change the experience of people in 10 ish years time?

Realizing that attitude in myself at times has given me so much more peace of mind. Just in general, not everything needs to be as fast and efficient as possible.

Not to mention the times where in the end I spend a lot of time and energy in trying to be faster only to end up with this xkcd: https://xkcd.com/1319/

As far as LLM use goes, I don't need moar velocity! so I don't try to min max my agentic workflow just to squeeze out X amount more lines code.

In fact, I don't really work with agentic workflows to begin with. I more or less still use LLMs as tools external to the process. Using them as interactive rubber duckies. Things like deciphering spaghetti code, do a sanity check on code I wrote (and being very critical of the suggestions they come up with), getting a quick jump start on stuff I haven't used again (how do I get started with X of Y again?), that sort of stuff.

Using LLMs in the IDE and other agentic use is something I have worked with. But to me it falls under what I call "lazy use" where you are further and further removed from the creation of code, the reasoning behind it, etc. I know it is a highly popular approach with many people on HN. But in my experience, it is an approach that makes skills of experienced developers atrophy and makes sure junior developers are less likely to pick them up. Making both overly reliant on tools that have been shown to be less than reliable when the output isn't properly reviewed.

I get the firm feeling that velocity crowd works in environments where they are judged by the amount of tickets closed. Basically "feature complete, test green, merged, move on!". In that context, it isn't really "important" that the tests that are green are also refactored by the thing itself, just that they are green. It is a symptom of a corporate environment where the focus is on these "productivity" metrics. From that perspective I can fully see the appeal for LLM heavy workflows as they most certainly will have an impact on metrics like "tickets closed" or "lines of code written".

If you're a salaried or hourly employee, you aren't paid for your output, you are paid for your time, with minimum expectations of productivity.

If you complete all your work in an hour... you still owe seven hours based on the expectations of your employment agreement, in order to earn your salary and benefits.

If you'd rather work in an output based capacity, you'll want to move to running your own contacting business in a fixed-bid type capacity.

If the structures and systems that are in-place only facilitate life getting more difficult in some way, then it probably will, unless it doesn't.

Housing getting nearly unownable is a symptom of that. Climate change is another.

Correct. But it becoming shittier is the strong default, with forces that you constantly have to fight against.

And the reason is very simple: Someone profits from it being shittier and they have a lot of patience and resources.

My company has been preparing for this for a while now, I guess, as my backlog clearly has years' worth of work in it and positions of people who have left the org remain unfilled. My colleagues at other companies are in a similar situation. Considering round after round of layoffs, if I got ahead a little bit and found that I had nothing else to do, I'd be worried for my job.

Society becoming shittier isn’t inevitable

Yes, I agree, but the deck is usually stacked against the worker, especially in America. I doubt this will be the issue that promotes any sort of collectivism.

I wish that the hype crowd would do that. It would make a for a much more enjoyable and sane experience on platforms like this. It's extremely difficult to have actual conversations subjects when there are crowds of fans involved who don't want to hear anything negative.

Yes, I also do realize there are people completely on the other side as well. But to honest, I can see why they are annoyed by the fan crowd.

If anything we'll see disposable systems (or parts) and the job of an SE will become even more like a plumber, connecting prebuilt business logic to prebuilt systems libraries. When one of those fails, have AI whip up a brand new one instead of troubleshooting the existing one(s). After all, for business leader it's the output that matters, not the code.

For 20+ years business leaders have been eager to shed the high overhead of developers via any means necessary while ignoring their most expensive employees' input. Anyone remember Dilbert? It was funny as a kid, and is now tragic in its timeless accuracy a generation later.