Hacker News Reader

Comment by jwittmayer | Hacker News Reader

Implying that nobody has put AI generated code into production yet?

Stuff that's going into production now (actual production, not startup MVP production) would have been being written just before Claude Code came out, so pretty much by definition no. There's some copilot-style assisted stuff in the wild, I guess? But not really more of it than pre-copilot so the productivity argument kind of falls through there.

Cursor came out 3 years ago. "Agentic" refactors have been a thing for 1.5 years. Vibecoding as a term has been created 1 year ago.

There are multiple companies that deploy to production daily. What are we even talking about?

Right but this agentic stuff was supposed to be the wave where we would finally actually see increased output, so we should probably be seeing it soon if it's real. Like, my dev team should definitely have the actual code they keep talking about their agents making, ready for me to put into production. As should my vendors. Any day now.

What is this nonsense?

You said that none of this was in production and then when people pointed out that it was obviously in production, you shifted the goal post to some other measure that you just imagined in your head.

Well, if it's in production, it's not at my company, any of my vendors, or for that matter any of the software I use in my private life; the pace of all of that is exactly what it was 2 years ago. When it shows up I'll form an opinion.

Let me amend that: one of my vendors has a new diffusion-based noise-reduction plugin that's pretty good though the resource usage is still too high. I imagine that will come down as they improve it. And that's pretty cool. But it didn't come out any faster it's just that it uses diffusion in the plugin itself. But docker was a much bigger impact on the software we use at work than AI has been so far.

I was even trying to come up with a list of software I use in my personal life to see if any of that has started coming out faster, and I came up with:

KDE

Supercollider

Puredata

Mixxx

Renoise

CUDA and ROCM

none of which have had any kind of release acceleration that I know of (though obviously the hardware to use the last two has gotten mind-blowingly expensive, alas). I use maybe three apps on my phone and they aren't updating any more frequently than they used to.

I get that for whatever reason this bugs people, but I'm in a very tech job and have a very tech personal life (just not webdev in either case) and literally have not seen anything I deal with change other than needing to learn to scroll past the AI summary at the top of search results.

What do you expect that it’s gonna announce itself in a modal dialogue when you run the software?

This isn’t like AI image generation where you’re going to convince yourself that you can tell the difference based on how you think it looks. Do you really think no one in the production chain of any of the software that you use picked up copilot in the last two years?

What signal are you hoping to receive that this is happening?

Well like I said in the sibling post to this one I'd expect really any of the software vendors in my professional or personal life to release either more rapidly or with a wider array of features than they were a few years ago, and that hasn't been my experience, at all.

The coding was never the slow part.

I'm certainly sympathetic to that argument, but if you scroll way back this thread started with the question of whether or not AI is transformative, and if it is neither faster nor better that would suggest "no".

Reply on news.ycombinator.com