AI coding • Hacker News Board

Re: AI coding

#191

Post by mihaic » Sat, Sep 13, 2025, 4:23 PM UTC

dehugger wrote: post #183

mihaic wrote: post #166

Earlier quoted context omitted.

I actually have seen plenty of people that could have done something like this, but did not because they simply never tried. Being daring by itself is a skill, but we're talking raw technical ability here. I've actually seen another developer that was probably in the same category write his own self-driving software. It kind of worked, but couldn't have ever been production ready, so it was just an exercise in flexin…

So, if I understand correctly, you've seen plenty of people that didn't do what he did? This was not a compelling argument.

> you've seen plenty of people that didn't do what he did?

Yes, because I've seen them build software that was actually used. And I've seen a few that did just like him, impressive sounding projects that had no usage.

I understand it's something subjective. I get the same feeling when looking at Damien Hirst's monstrouly expensive stuff that leaves me cold. Even after I get the concepts behind the works, my end feeling is of "so what?".

Re: AI coding

#192

Post by redhale » Sat, Sep 13, 2025, 5:19 PM UTC

> The only reason it works for many common programming workflows is because they are common. The minute you try to do new things, you need to be as verbose as the underlying language.

You can stop here. This is enough to change the game. SO MUCH time and money is spent on mostly-boilerplate line of business apps. If you don't understand this, you are in a bubble of some kind. The corporate world's IT departments and consulting agencies are filled with developers who write nothing but what boil down to custom-tailored CRUD apps. AI is absolutely KILLER for this kind of work.

If you're doing novel research, perhaps not. But that describes a relatively small portion of the worldwide developer community.

Re: AI coding

#193

Post by scuff3d » Sat, Sep 13, 2025, 5:24 PM UTC

dm3 wrote: post #134

krystofee wrote: post #104

Earlier quoted context omitted.

I’m experiencing something similar. We have a codebase of about 150k lines of backend code. On one hand, I feel significantly more productive - perhaps 400% more efficient when it comes to actually writing code. I can iterate on the same feature multiple times, refining it until it’s perfect. However, the challenge has shifted to code review. I now spend the vast majority of my time reading code rather than writing i…

I'm very surprised you find this workflow more efficient than just writing the code. I find constructing the mental model of the solution and how it fits into existing system and codebase to be 90% of effort, then actually writing the code is 10%. Admittedly, I don't have to write any boilerplate due to the problem domain and tech choices. Coding agents definitely help with the last 10% and also all the adjacent work…

I doubt it actually is. All the extra effort it takes to make the AI do something useful on non trivial tasks is going to end up being a wash in terms of productivity, if not a net negative. But it feels more productive because of how fast the AI can iterate.

And you get to pay some big corporation for the privilege.

Re: AI coding

#194

Post by scuff3d » Sat, Sep 13, 2025, 5:27 PM UTC

matt3D wrote: post #70

This is a more extreme example of the general hacker news group think about AI. Geohot is easily a 99.999 percentile developer, and yet he can’t seem to reconcile that the other 99.999 percent are doing something much more basic than he can ever comprehend. It’s some kind of expert paradox, if everyone was as smart and capable as the experts, then they wouldn’t be experts. I have come across many developers that beha…

I work in aerospace, trust me some of the aerospace engineers aren't any better.

But don't worry. The company puts them somewhere they can't do any damage. Most of them become managers.

Re: AI coding

#195

Post by Wowfunhappy » Sat, Sep 13, 2025, 5:27 PM UTC

layer8 wrote: post #186

jb3689 wrote: post #165

Earlier quoted context omitted.

100% agree. I am interested in seeing how this will change how I work. I'm finding that I'm now more concerned with how I can keep the AI busy and how I can keep the quality of outputs high. I believe it has a lot to do with how my projects are structured and documented. There are also some menial issues (e.g. structuring projects to avoid merge conflicts becoming bottlenecks) I expect that in a year my relationship…

An important skill for software developers is the ability to reason about what the effects of their coce will be, over all possible conditions and inputs, as opposed to trial and error limited to specific inputs, or (as is the case with non-deterministic LLMs) limited to single executions. This skill is independent of whether you are coding in assembly or are using higher-level languages and tooling. Using LLMs exact…

> Using LLMs exactly doesn’t train that skill

I actually think this is one skill LLMs _do_ train, albeit for an entirely different reason. Claude is fairly bad at considering edge cases in my experience, so I generally have to prompt for them specifically.

Even for entirely “vibe-coded” apps I could theoretically have created without knowing any programming syntax, I was successful only because I knew about possible edge cases.

Re: AI coding

#196

Post by ac29 » Sat, Sep 13, 2025, 5:40 PM UTC

hereme888 wrote: post #95

I'm a 100% vibe-coder. AI/CS is not my field. I've made plenty of neat apps that are useful to me. Don't ask me how they work; they just do. Sure the engineering may be abysmal, but it's good enough to work. It only takes basic english to produce these results, plus complaining to the AI agent that "The GUI is ugly and overcrowded. Make it look better, and dark mode." Want specs? "include a specs.md" This isn't a 20%…

> I'm a 100% vibe-coder. AI/CS is not my field. I've made plenty of neat apps that are useful to me.

This describes me pretty well too, though I do have a tiny bit of programming experience. I wrote maybe 5000 lines of code unassisted between 1995-2024. I didn't enjoy it for the most part, nor did I ever feel I was particularly good at it. On the more complex stuff I made, it might take several weeks of effort to produce a couple hundred lines of working code.

Flash forward to 2025 and I co-wrote (with LLMs) a genuinely useful piece of back office code to automate a logistics problem I was previously solving via a manual process in a spreadsheet. It would hardly be difficult for most people here to write this program, its just making some API calls, doing basic arithmetic, and displaying the results in a TUI. But I took a crack at it several times on my own and unfortunately between the API documentation being crap and my own lack of experience, I never got to the point where I could even make a single API call. LLMs got me over that hump and greatly assisted with writing the rest of the codebase, though I did write some of it by hand and worked through some debugging to solve issues in edge cases. Unlike OP, I do think I reasonably well understand what >90% of the code is doing.

> This isn't a 20% more productive feeling. It's productivity beyond what I will ever do on my own, given this is not my field.

So yeah, to the people here saying the above sentiment is BS - its not. For people who have never worked in programming or even in tech, these tools can be immensely useful.

Re: AI coding

#197

Post by xeckr » Sat, Sep 13, 2025, 5:46 PM UTC

By the same token, a software developer could perhaps also be likened to a compiler, taking language and/or visual information as input and turning it into code.

Re: AI coding

#198

Post by sltr » Sat, Sep 13, 2025, 6:02 PM UTC

What an aggressive tone.

By the author's implied definition of compiler, a human is also a compiler. (Coffee in, code out, so the saying goes.)

But code is distinct from design, and unlike compilers, humans are synthesizers of design. LLMs let you spend more time as system architect instead of code monkey.

Re: AI coding

#199

Post by bdcravens » Sat, Sep 13, 2025, 6:09 PM UTC

saulpw wrote: post #187

bdcravens wrote: post #169

Earlier quoted context omitted.

I don't disagree, but I find a better use of my time is writing. Not code, but essentially a work journal. It's not big thoughts, it's bullet points. It's not documentation, but more of a open mind map: what's been done, what needs to be done, questions that inevitably pop up, etc. I use Obsidian for this, but if I write much more than what would go on a few post-it notes, it's too much.

Writing and coding use different parts of the brain. Writing is for creating a model to communicate about human things, and coding is for creating a model for communicating about computer things.

Sure, but we've all had one of those "ah-ha!" moments when we took ourselves out of the code. I feel like putting yourself in a position when you're thinking big picture is far more useful than hand-coding a routes file or adding validation to a model file.

Re: AI coding

#200

Post by bdcravens » Sat, Sep 13, 2025, 6:13 PM UTC

somewhereoutth wrote: post #102

bdcravens wrote: post #25

I'm almost 50, and have been writing code professionally since the late 90s. I can pretty much see projects in my head, and know exactly what to build. I also get paid pretty well for what I do. You'd think I'd be the prototype for anti-AI. I'm not. I can build anything, but often struggle with getting bogged down with all the basic work. I love AI for speed running through all the boring stuff and getting to the goo…

> developer somewhere between junior and mid-level Why the insistence on anthropomorphizing what is just a tool? It has no agency, does not 'think' in any meaningful manner, it is just pattern matching on a vast corpus of training data. That's not to say it can't be very useful - as you seem to have found - but it is still just a tool.

It's less about what the tool is, and more about the kind of work we often assign to less experienced developers. Pattern matching in meatspace is still pattern matching.