We fought for quality code mostly for selfish reasons. Because it sucked to work in a messy, fragile codebase.

But AI doesn't care. It can, and it will, happily churn through a messy codebase.

sunaurus 2 hours ago | parent | next [-]

> It can, and it will, happily churn through a messy codebase.

Do you have examples of where this actually worked for you? How large was the codebase, how long did you have agents generate the mess in it?

I ask because my experience is pretty different, even with Opus 5.5 and Astra - unleashing agents on a messy, vibe coded repo only seems to increase the mess. The agents just slow down more and more as they keep generating further mess, they make more and more incorrect assumptions based on bad code previous agents left behind, and I grow more disappointed in their capabilities every day.

Actually reviewing all agent output thoroughly is the only effective way I've personally found of working with agents. In order to ensure that every line of code they add is actually there for a good reason, and not organized in a way that future agents will miss it, or misinterpret it. It's the only way I can prevent them from producing worse results as the repo grows.

▲

greenowl 23 minutes ago | parent | next [-]

Actual effectiveness aside, the comment is more about willingness to even work with messy codebases. No matter how messy the code gets, agents can and will continue to push forward. While humans would just get fed up. In the past I've known multiple devs who literally quit and switched jobs from working with messy codebases (you can't always tell what you're walking into during the interview stage).

▲

discreteevent an hour ago | parent | prev [-]

The account you are replying to is a bot/shill who has been given the same talking points as the rest of them. There are trillions riding on this.

	▲	randysalami 40 minutes ago \| parent \| next [-]
		Dang can you confirm if this is true?
	▲	greenowl 33 minutes ago \| parent \| prev [-]
		What a bizarre conclusion. Read my comments you'll actually find they trend toward the opposite.

▲

dnautics an hour ago | parent | prev | next [-]

Yeah but if you don't keep tabs on it, you might not know what to prompt it to fix. For example, early on I established in my fun proof assistant project, that I did NOT want total functions (1/0 should not be permitted, 3-4 not permitted in N), and it seems to have forgotten that along the way and it took a 2 day struggle session to remove total functions from the standard library.

Managed to get it in time to have an accurate claim that "stdlib has no total functions" prior to presenting it to someone semifamous, who doesn't like that lean has total functions

▲

sandeepkd 2 hours ago | parent | prev | next [-]

I think you are underestimating the spiral effect. Working with a documented and structured code makes it easier to understand and reason with even for LLMs. At the end of the day its working with patterns, grammar, trying to deduce and implement intent. Its possible where we enter the era of blackbox applications where no one knows the internal, and yet for the required inputs it provides the expected outputs. Every failure would re-trigger the evaluation of the whole application to accomodate that one failure in requirement.

▲

mostertoaster 3 hours ago | parent | prev [-]

I agree. And while AI does better when there is less tech debt, at the same time paying down tech debt has gotten immensely cheaper.

I find myself desiring simplifying a codebase just because I still need to make so many decisions and the cognitive load of that has gone up with how much I can do in parallel and still keep track of.