[christkv]

I have found that the attention moves to thinking about the things I want done and planning, reading and iterating over the specs and other artefacts that will be part of running the agents. I still need to understand the code and iterate over it to get to a usable and maintainable point.

I find the problem is we are reaching the top of the slop curve. I will subside because it's impossible to actually do anything useful with all the output. There will just be a ton of half-finished and abandoned projects. Whatever gets into production will require more eyes on it.

I just think a lot of people are still stuck in the "holy f** I'm so productive" and working themselves into the ground being productive pumping out code. I think it's a phase that will pass.

[user43928]

I find that it depends a lot on the project.

I'm doing solo mobile app projects, and I have no need to iterate on specs. The bottleneck is QA testing whether it works on the phone.

I don't need to carefully review and understand the implementation. It's not important whether I understand the details of how exactly UICollectionView in Apple's UIKit works.

I see that my implementation works on different physical devices, my tests cover device rotation, and I checked the memory allocations in the Instruments tool.

It has been some months of part-time work on my side, and I will publish this iOS app soon.

Without AI I could not have done it, the scope of the features is too large. The project is around 100k LOC.

It is not true that projects become unmaintainable and abandoned because of agentic engineering, or even vibecoding if you want to call it that.

[christkv]

The problem starts with a project of 20 engineers pumping 1000+ lines of code a day into a shared code base.

Single dev it's amazing, tiny team still good, large team wohoooo

[ChrisLTD]

This 100%. Maybe I have a prompting skill issue, but without my guidance Opus (and now Fable) writes some gnarly stuff with tons of small bugs, and weird design decisions.

[nicce]

LLMs still do not have proper contextual understanding of their solutions. Just couple days ago I was using GPT 5.5 with xhigh to vide code some application, and yet it defaulted for sorting dates from new to old by using plain string comparison. Just one of the many bugs.

[properbrew]

This absolutely fascinates me. I had a friend who needed subtitle files generating for audio and using in CapCut yesterday yet none of the available stuff was suitable, so he asked if I could adapt some of my software to export subtitles.

2 hours later he's got a fully working piece of local software that does exactly what he wants, yet yours is not able to even sort dates correctly. Feel free to download it if you want to see for yourself, I didn't even do any UI tweaks as this was just a tool for him to use:

Linux - https://downloads.blazingbanana.com/whistle-subtitles/unstab...

Windows - https://downloads.blazingbanana.com/whistle-subtitles/unstab...

Mac - https://downloads.blazingbanana.com/whistle-subtitles/unstab...

How can there be such a massive gap in what can be produced?

[nicce]

> How can there be such a massive gap in what can be produced?

What I was doing looks really nice and mostly works on the surface, but it is all about the corner cases where these bugs appear. In another day I was able to generate Frida script with LLM help that bypasses Dart certificate pinning/validation and proxies all the traffic by injecting the runtime binaries. With the latest Flutter/Dart version on Android when doing security analysis.

[properbrew]

Ahhh ok I totally understand what you mean. Yea the edge cases are absolutely where you start to feel the pain and things look good on the surface until you dig in. I think even in the age of LLMs the adage of 90% of the time is spent of the last 10% will ring true.

Sure an app can be built and spun up in an afternoon, but are you willing to spend another 6 months ironing out all those little bugs, tuning it a bit, testing, tweaking, testing etc.

[christkv]

LLMs is a little like the guy in Memento right. Every single conversation is it looking at the compressed scribbles of the context.

[polotics]

it sometimes goes like this: vibecoding viberequirements, vibeleadership vibemanagement, vibecustomers

as a matter of taste, you can substitute "slop" as a prefix

[christkv]

I think we as a professional class have gotten a bit overwhelmed with the magic slot machine. 2 in the morning let me just do one more pull on the slop machine. I WILL win this time.

It's a phase. The problem is the managerial class sees it as a magic black box and don't understand it's limitations. Calling it AI does not help either. It's the "rockstar developer" illness but on crack.

[pydry]

I've never seen a greater disconnect between what I read on social media about vibe coding and what I've seen in real life.

In particular the whole "the best people are the ones who will use it the best". IME the best ones are the ones keeping it the most at arm's length while the people who embrace it the most churn out epic amounts of utter slop.