Maybe what the parent commenter was referring to is that recently CloudFlare have published multiple vibe-code demo blog posts whilst trying to pass it off as production grade, for example https://news.ycombinator.com/item?id=46781516 where they implemented an open standard communication protocol on CloudFlare Workers.

The blog post was chock full of factual errors, claimed to be based off project X but wasn't at all and even had the cheek to include that it was arguably the most secure way to deploy such a server, with their implementation apparently already being used by their team to serve real traffic. Meanwhile the repo was full of TODOs for all the security aspects of the protocol.

Of course after the backlash a lot of this was covered up so look at the archives if you are curious.

They have really done a disservice to themselves because their blog posts used to be excellent, but now I have to question whether it's another blogpost full of fakery like that one (and there was another since iirc). Given this blog post talks about reimplementing a popular project, it starts to give off the signs of being another one of these. Unfortunate if that's not the case

IMO, what matters is whether it's coded with AI or it's coded by AI.

Is AI merely being used as a tool to aid the engineer? Because that's what I do. I use it as essentially a super-autocomplete. It typically only writes a couple lines at a time for me. On rare occasions, I can write a function signature and let it fill out the body. That's coding with AI.

Anything more than that though? You're stepping into coding by AI, which utterly fails at anything beyond an MVP. Once you go over 2,000 lines of code or so, it falls apart. It can't reason about anything with even a small amount of complexity, and every "bug fix" either fails to fix the bug or it introduces two more.

CloudFlare announcements have caused them to lose credibility in this area.

whether people want to admit it or not, agent encoding is kind of the norm right now and I think the fear is the stories coming out of places like Block, Inc where they announced they fired 4,000 engineers a couple days ago because of what's the obvious truth today versus 6 months ago.... one expert software engineer can do the work of 20-40 people, so why do we need so many people? it's a hard pill to swallow, it's easier to claim that agentic coding doesn't work or that the code is sloppy and it doesn't work when in reality most companies are currently using it everyday, especially the large ones.