Comment by ChrisMarshallNY | Hacker News Reader

Comment by ChrisMarshallNY | Hacker News Reader

Parent Full thread

ChrisMarshallNY

I find myself sympathetic to the author's PoV, but I am incorporating LLMs into my workflow, with a resultant jaw-dropping (to me) increase in velocity.

But I am not just dispatching to agents. I work interactively with a chat interface, and sometimes, I will just bin a whole hour's worth of back-and-forth, because we're not getting anywhere (in fact, I did exactly that, about 30 minutes ago).

But that hour is peanuts, compared to the ten hours that I would have spent, trying to figure it out on my own. With an LLM (and git), I can "run something up the flagpole, and see who salutes." I can afford to experiment with very large code bases, and toss out a whole bunch of stuff, if need be.

That said, I know damn well, that quite a few folks here, would sneer at my methodology, as "awkward, stodgy, and slow." Nevertheless, I am pretty chuffed with the results. Yeah, it's slower than some folks would do it, but the Quality is really high, and I'm happy with the results.

My favorite thing to do, is (for example) toss all 5 of my SDK files into the LLM, paste in the JSON server interaction, describe the bug, and ask it to help me figure it out.

Nine times out of ten, it finds the bug quickly. The real bug. I am not always happy with the proposed solutions, but finding the root cause is always the time-consuming part.

Same approach as mine. :)

One more thing I try is give same prompt once in a while to ChatGPT/Gemini/Grok. Than take the 2 out of 3 ideas forward.

All leading AI seems to have some blind spot. Like some kind of intricsic character. 1 will completely overlook some particular bug while finding other excellent bugs and edge cases.

Getting the code through all 3 before commiting has shown excellent results to me.

Hope you'll be able to afford that workflow once the money dries up and costs skyrocket.

I think that is the exact friction with the original post title.

For an experienced coder, writing code is never the problem. Some tasks are more raw writing effort then others, sure. The problem is understanding the problem, the task, the constraints and how to express them correctly, in a way that suits your environment (i.e. company, ecosystem, platform).

My brain is the actual problem, how fast I can digest information and make sense of it. For me, this is the process that always consumes the most time.

Last week I've wasted a full day, just because I didn't had a decent overview of what I was working on. It was heavily LLM chatbot assisted, at no point in time the LLM nudged me into the right direction, mainly because I was confidently asking the wrong questions.

No AI will help you if you insist being dumb, but it will accelerate you in that regard. That at least works for me.

I do the same. Code is cheap - try it, see if it works, if it doesn’t completely throw it away. LLMs and agents get into doom loops and keep making the situation worse if it’s not going well from the outset. Usually if you’re engaged you can notice this happening and change tack.

A perfectly good methodology with the right guardrails in place imho.

> the Quality is really high, and I'm happy with the results.

>toss all 5 of my SDK files into the LLM, paste in the JSON server interaction, describe the bug, and ask it to help me figure it out.

I wonder if this "quality" code wouldn't have that many bugs to dive into if it was more carefully considered and produced up front?

This harkens back to a study in 2024 were senior devs were actually less productive with LLMS but they felt more productive, even after being told they were less productive.

I’m sure you’re right.

I’m actually a terrible programmer. This is just how I compensate.

Of course, you are more than welcome to visit my portfolio (easily found from my HN profile), and make fun of it.

I don't mean to belittle your expertise. You likely have triple my experience in tech.

Its just more that we should understand that there's no free lunch.even pre-Ai, codegen has the offset of creating nigh unmaintainable code, and making proper iteration (changing the codegen itself) take longer. But ideally, you gain long term speed at the cost of massive tech debt.

This ultimately operates as faster codegen, wrapped in a black box. If shipping is all that matters, then it is what it is. But we should understand the sacrifices for that.

That's one of the advantages of doing it iteratively, and in an arena that I already know well (Swift native application development). I'm able to evaluate each step, and make decisions (and adjustments, but I'm usually better off, asking for another try).

I am not yet comfortable with just asking an agent to "write me a Facebook." That seems to be what all the kids are into, these days, but I am not convinced the tech has reached these lofty expectations.

I just use it as I would, a coding partner, who knows a lot, is a bit obsessive, and does things quickly. I've worked with some aspies, in my day, that actually come very close. Working with them, has helped me to work with an LLM, better.

I have been around, long enough, to have seen several "watershed" developments in software technology. Each one, was accompanied by wailing, rending of clothes, and gnashing of teeth; yet, here we are. We have not just survived. We have thrived.

> were senior devs were actually less productive with LLMS but they felt more productive

I'm certain this is true for me. The only thing LLMs do get me is the ability to make forward progress on tasks while I'm in meetings, which is a net positive at least.

Reply on news.ycombinator.com