This little line from the article scares me: "but a software engineer would iron out the remaining potential bugs that I could not find quickly"

Every sw dev knows this is a very dangerous, and unrealistic, assumption.

it's basically a tiny statement that kind of hand waves all the 'actual stuff'.

It's "I did the first/easy 90% now someone else do the second/hard 90%". Same as it ever was.

We're not supposed to be crude on HN but that's some real Dilbert level stuff right there. Like spit out my coffee laughing, cringing. It's too bad the Dilbert guy seemed to have lost his mind in meta level cynicism (and maybe his legacy as well) and also passed away, because we kind of need him now. Dilbert is almost made more for the AI age than the computing age.

Phew, was this a test? ;P

Tnx for checking, and reconfirming!

This is exactly the reason why I like to work with local models on a regular specced machine. The fact that the agent moves slower allows me to stay in the loop much better, compared to skimming through a huge amount of generated content and data and then going to the end really fast to make sense of it all, in the interest of time (and thus losing track and quality). The fact that I can run it locally makes it (much) cheaper too.

An LLM should not "generate specs", a human should. The LLM can work from the specs. It can never infer meaning from a vague prompt. If so, it will start guessing. Every human that ever did functional specification or information analysis at some point knows this. Or has learned the hard way, something with assumptions and asses ;)

The guessing of a LLM for a vague prompt is better than the one of your average developer.

A prompt like "write these two files on disk" will very likely make the LLM do some sort of an atomic write/swap operation, unlike the average developer which will just write the two files and maybe later encounter a race condition bug. You can argue the LLM output is overkill, but it will also be more robust on average.

What kind of race condition do you have in mind?

Gitea might be an option also.

Out of interest: what does "better AI integration" mean? Any specific functional or non-functional requirements?

I didn't challenge them on this but it's because of Claude integrations with github. I'm not sure what that gives them over just running it against the codebase, but I didn't want to lose the opportunity to finally move them from that EoL server

Out of interest: what VCS where you using with CircleCI? (as CircleCI is VCS agnostic)

Correction: Gitlab is not Dutch. It has a Dutch co-founder, and was a B.V. for about one year before moving to the US and become an Inc. The original founder is from Ukraine.

Let engineers do Vibe-accounting because, AI.

Debug with SSH(1) is still one of our (CircleCI) most loved and praised features. I really believe that these little QoL features can make a world of difference for sw developers and engineers, and this stays a strong focus for us.

1: https://circleci.com/docs/guides/execution-managed/ssh-acces...

(Disclaimer: i work at CircleCI)