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

Vibe coding and agentic engineering are getting closer than I'd like

simonwillison.net

Post reply on HN
Print view View original

361–370 of 958 posts

Re: Vibe coding and agentic engineering are getting closer than I'd like

#361

Post by wilsonnb3 »

chromacity wrote: post #352
hintymad wrote: post #302

Earlier quoted context omitted.

> I don't think that AIs have become more trustworthy, the errors are just more subtle. Honest question: what about the counter-argument that humans make subtle mistakes all the time, so why do we treat AI any differently? A difference to me is that when we manually write code, we reason about the code carefully with a purpose. Yes we do make mistakes, but the mistakes are grounded in a certain range. In contrast, AI…

One answer, as another person pointed out, is that LLM mistakes are just different . They are less explicable, less predictable, and therefore harder to spot. I can easily anticipate how an inexperienced engineer is going to mess up their first pull request for my project. I have no idea what an LLM might do. Worse, I know it might ace the first fifty pull requests and then make an absolutely mind-boggling mistake in…

> I can easily anticipate how an inexperienced engineer is going to mess up their first pull request for my project.

Isn’t this just because you have seen a lot of PRs from inexperienced engineers? People learn LLM behavior over time, too.

Re: Vibe coding and agentic engineering are getting closer than I'd like

#362

Post by seanw444 »

empthought wrote: post #275
callc wrote: post #247

Earlier quoted context omitted.

Ah yes, we must force these obstinate engineers to the right path! Only after getting everyone to see the light will they understand and thank us for boundless productivity!! /s Perhaps these “obstinate” engineers have good reason in their decision. And it should be their decision! To be so confident in what is “the right way (TM)” and try to force it onto others is... revealing.

Engineers that didn't move past src.v35.final.zip version control don't really have jobs today, either.

I seriously don't see how version control and LLMs are comparable. A deterministic way to track code changes over time, versus an essentially non-deterministic statistical code generator that might get you what you want, and might do it in a reasonable time frame, and that might not land you in a minefield of short-term-good/long-term-bad design points.

Re: Vibe coding and agentic engineering are getting closer than I'd like

#364

Post by wilsonnb3 »

BoorishBears wrote: post #322
hintymad wrote: post #302

Earlier quoted context omitted.

> I don't think that AIs have become more trustworthy, the errors are just more subtle. Honest question: what about the counter-argument that humans make subtle mistakes all the time, so why do we treat AI any differently? A difference to me is that when we manually write code, we reason about the code carefully with a purpose. Yes we do make mistakes, but the mistakes are grounded in a certain range. In contrast, AI…

This is like having a coworker who's as skilled as you if not more skilled, but also an alien. Their mental model doesn't map cleanly enough to yours, and so where for a human you'd have some way to follow their thought patterns and identify mistakes, here the alien makes mistakes that don't add up. Like the alien has encyclopedic knowledge of op codes in some esoteric soviet MCU but sometimes forgets how to look for…

Dealing with the alien coworkers has always been the job, that is what software is to most people.

Software developers get paid big money because they can speak alien, the only thing that is changing is the dialect.

Re: Vibe coding and agentic engineering are getting closer than I'd like

#365

Post by wg0 »

Here's for the AI supremacists:

Let's assume AI is 10x perfect than humnas in accuracy and produces 10x less bugs and increases the speed by 1000x compared to a very capable software engineer.

Now imagine this: A car travels at a road that has 10x more bumps but it is traveling 1000x slower pace so even though there are 10x bumps, your ride will feel less bumpy because you're encountering them at far lower pace.

Now imagine a road that has 10x less bumps on the road but you're traveling at 1000x the speed. Your ride would be lot more bumpy.

That's the agentic coding for you. Your ride would be a lot more painful. There's lots of denial around that but as time progresses it'll be very hard to deny.

Lastly - vibe coding is honest but agentic coding is snake oil [0] and these arguments about having harnesses that have dozens of memory, agent and skill files with rules sprinkled in them pages and pages of them is absolutely wrong as well. Such paradigm assumes that LLMs are perfect reliable super accurate rule followers and only problem as industry that we have is not being able to specify enough rules clearly enough.

Such a belief could only be held by someone who hasn't worked with LLMs long enough or is a totally non technical person not knowledgeable enough to know how LLMs work but holding on to such wrong belief system by highly technical community is highly regrettable.

[0]. https://news.ycombinator.com/item?id=48018018

Re: Vibe coding and agentic engineering are getting closer than I'd like

#366

Post by andy_ppp »

Honestly, I think the need for devs is total copium, the progress made in two years is astounding and in two years time they will be better at programming than 99% of programmers. It’s incredible what they can do now. No it’s not perfect but imagine where we’ll be in 5 or 10 years.

Re: Vibe coding and agentic engineering are getting closer than I'd like

#367

Post by crimebrasil »

alishayk wrote: post #320
wek wrote: post #230

What an excellent article by a smart, humble, still-learning person! Favorite quote:" There are a whole bunch of reasons I’m not scared that my career as a software engineer is over now that computers can write their own code, partly because these things are amplifiers of existing experience. If you know what you’re doing, you can run so much faster with them. [...] I’m constantly reminded as I work with these tools…

What do you do if you don't have that existing experience? How do you build it up?

[flagged]

Re: Vibe coding and agentic engineering are getting closer than I'd like

#368

Post by seanw444 »

CharlieDigital wrote: post #347
asdfman123 wrote: post #293

Earlier quoted context omitted.

You can direct LLMs to do test-driven development, though. Write several tests, then make sure the code matches it. And also make sure the agent organizes the code correctly.

The LLM obliges and writes a lot of useless tests. I have asked devs to delete several tests in the last day alone.

"I don't trust this giant statistical model to generate correct code, so to fix it, I'm going to have this giant statistical model generate more code to confirm that the other code it generated is correct."

I swear I'm living through mass hysteria.

Re: Vibe coding and agentic engineering are getting closer than I'd like

#369

Post by gck1 »

mschuster91 wrote: post #305
retinaros wrote: post #288

Earlier quoted context omitted.

Its trash vibecoded markdown files around pi. This exemplifies well what op’is saying. We are at the ICO stage of llms. Hopefully there wont be an nft one

As much as I love to hate on AI: even the bad apples still produce something that one can reasonably work with. Cryptocurrencies? Barely any other use than money laundering, buying drugs and betting on the outcome of battles in war. And NFTs? No use at all other than money laundering and setting money ablaze.

Privacy and security from government overreach is not enough?

Re: Vibe coding and agentic engineering are getting closer than I'd like

#370

Post by autoconfig »

devin wrote: post #39

> If you can go from producing 200 lines of code a day to 2,000 lines of code a day, what else breaks? The entire software development lifecycle was, it turns out, designed around the idea that it takes a day to produce a few hundred lines of code. And now it doesn’t. It is so embarrassing that LOC is being used as a metric for engineering output.

The charitable interpretation here is obviously that the LoCs are equivalent in quality, in which case it is a very useful metric in the context that was presented. The inability to infer that should be embarrassing.
Post reply on HN