Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Effects of Gen AI on High Skilled Work: Experiments with Software Developers

papers.ssrn.com

Post reply on HN
Print view View original

471–480 of 532 posts

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#471

Post by Aeolun »

badpun wrote: post #346
Aeolun wrote: post #341

Earlier quoted context omitted.

> Startup architecture and a 500+ engineer org's architecture are fundamentally different Certainly. The startup architecture is often better. I don’t know what exactly leads to those overcomplicated enterprise things, but I suspect its a lack of limitations.

Large org architecture is, most importantly, just much larger. The bank I worked in had 6000 _systems_. A lot of that was redundancy due to a lot of mergers and acquisitions that didn't get cleaned up (because it was really hard to do). Compare that to your typical SaaS startup, which typically has at most a couple systems.

There’s that, sure. But that’s also because everyone is way too happy to add another new system to the pile. Large enterprises generally have only so many business functions. I can get to about 100 necessary systems before feeling like they’re redundant.

And all these people have been given fodder by the microservices revolution.

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#472

Post by danielmarkbruce »

dimal wrote: post #402

It matters what you measure. The studies only looked at Copilot usage. I’m an experienced engineer. Copilot is worse than useless for me. I spend most of my time understanding the problem space, understanding the constraints and affordances of the environment I’m in and thinking about the code I’m going to write app. When I start typing code, I know what I’m going to write, and so a “helpful” Copilot autocomplete is…

Just for clarity - are you saying that going back and forth with ChatGPT is more useful than Co-pilot? The reason I ask is I have both and 95% of the benefit is ChatGPT.

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#473

Post by lagrange77 »

acchow wrote: post #460
lagrange77 wrote: post #446

Earlier quoted context omitted.

Excuse my ignorance, i have avoided copilot until now.. Does it have (some of) the other files of the project in it's context, when you use it in a test file?

Copilot has your open tabs in context. Cursor has that plus whatever files you want to specifically add. Or it has a mode where you can feed it the entire project and it searches to decide which files to add

Thanks!

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#474

Post by wahnfrieden »

jacobsenscott wrote: post #408
wahnfrieden wrote: post #100

Earlier quoted context omitted.

Maybe not at your workplaces, but at mine, we measured bugs, change failure rate, uptime, "critical functionality" uptime, regressions, performance, CSAT, etc. in addition to qualitative research on quality in-team and with customers

That's 0.001% of the workplaces out there. But good for them. Do you think it actually improved software quality, or was it all busy work?

Yah

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#475

Post by knallfrosch »

An interesting topic that will, given this site's community, attract a lot of strong opinions.

I, for one, only decide whether CoPilot's productivity increase is worth the $10 it costs per month.

It doesn't really matter whether you're an employer getting a 3–30% increase in productivity or whether you pay for it personally and finish 2 hours faster every week and log off illegaly. It's easily worth its money. What more to consider?

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#476

Post by ThalesX »

knallfrosch wrote: post #475

An interesting topic that will, given this site's community, attract a lot of strong opinions. I, for one, only decide whether CoPilot's productivity increase is worth the $10 it costs per month. It doesn't really matter whether you're an employer getting a 3–30% increase in productivity or whether you pay for it personally and finish 2 hours faster every week and log off illegaly. It's easily worth its money. What m…

> What more to consider?

Potentially sharing company IP with a 3rd party?

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#477

Post by LouisSayers »

sum1ren wrote: post #38

When I'm using genai to write some code for me, I lose the internal mental state of what my code is doing. As such, when I do have to debug problems myself, or dream up ideas of improvements, I no longer can do this properly due to lack of internal mental state. Wonder how people who have used genai coding successfully get around this?

I use Claude and there's a couple things I'd suggest.

1) You need to be the boss with the AI being your assistant. You are now a project manager coming up with strict requirements of what you'd like done. Your developer (AI) needs context, constraints and needs to be told exactly what you'd like created without necessarily diving into the technical details.

2) Planning - you need to have a high level plan of roughly how you'd like to structure your code. Think of it like you're drawing the outline and AI is filling in the gaps.

3) Separation of concerns - use software principles to drive your code design. Break problems down into separate components, AI is good at filling in components that are well defined.

Once you change your thinking to a higher level, then you can maintain flow state. Of course the AI isn't perfect and will make mistakes - you do need to question it as you go. The more creative you become with a solution the harder time the AI will have and sometimes you'll have to break out and fix things up yourself.

For me it's a huge productivity boost.

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#478

Post by 29athrowaway »

jillesvangurp wrote: post #348
29athrowaway wrote: post #61

The study doesn't cover the long term effects of using generative AI on a project, which is the deskilling of your workforce. Because development will become an auction-like activity where the one that accepts more suggestions wins.

I look at this very differently. There are a lot of grumpy people jealously guarding their "skills" and artisanal coding prowess getting really annoyed at the juniors that come in and devalue what they do by just asking an AI to do the same thing and then moving on with their day. Young people are more mentally agile and most young people are now growing up with LLMs as a tool chain that was always there. I'm actuall…

At the end of the day, all the layers of abstraction will be removed... interpreters, transpilers, compilers, microcode, instruction sets... all the layers of convenience and APIs created for humans and the overhead they introduce will be gone.

It will be a machine game, just like assembly is mostly compiler generated today.

The AI will produce faster, smaller, more power efficient and more secure binaries than a human ever can.

The AI will learn all the compilation steps, fuse everything into a simplified pipeline and what we call compilers today will be erased from reality.

Re: Effects of Gen AI on High Skilled Work: Experiments with Software Developers

#480

Post by mewpmewp2 »

risyachka wrote: post #462
kbaker wrote: post #440

Earlier quoted context omitted.

Yeah, Copilot is meh. Aider-chat for things with GPT-4 earlier this year was a huge step up. But recently using Claude Sonnet + Haiku through OpenRouter also with aider, and it is like a new dimension of programming. Working on new projects in Rust and a separate SPA frontend, it just ... implements whatever you ask like magic. Gets it about 90-95% right at the first prompt. Since I am pretty new to Rust, there are a…

Thats the point. The less experienced you are the more gains you see, and vice versa. The issue in the first case is that you have no idea if it tells you good stuff or garbage. Also in simple projects it shines, when the project is more complex - it becomes mostly useless.

Isn't that the opposite? The more experienced you are the higher the gains since you are able to see what it outputs and immediately tell if it is what you expected. You also know how to put on the best input for it to auto complete rest of it while you are already planning next steps in your head. I feel like a superhuman with it, honestly.
Post reply on HN