Slop Forks
I maintain a very large open-source static site built with Jekyll, locally this takes nearly ~2m to build. Not really an issue in a CI deployment but when developing locally, even with incremental builds, it was painful enough to send me in search of alternatives.
Which is where I came across Rustkyll - a port of Jekyll, written entirely by AI agents. I think the disparaging term for this is "slop fork". However, the stats don't lie - my build has been cut from nearly ~2m to <15s. Not sloppy at all.
Now admittedly, I don't really care about the internals of Jekyll, it's a means to an end for me. Which I think highlights something interesting about where development sits these days - sometimes we get so caught up in the old idea that building software means writing and understanding every line of code - but at it's core, it is about solving problems.
This might read as a contradiction to my last piece, where I argued that understanding slips away when we stop engaging with the systems we build. The key word there is build. Jekyll was always a black box to me - I consume it, I didn't write it. The system I actually built is the site, and that part I understand. Swapping one build tool for another doesn't cost me knowledge I ever had to begin with.
Now I'll concede Rustkyll only succeeds here because the stakes are low and it has a source of truth to rely on. Yet precisely for those reasons, it succeeds today. In the past, no one — certainly not me — would have ported the whole of Jekyll to Rust just to build a side-project faster.
Yet that's the shift. The problem was never impossible — porting Jekyll to Rust has always been doable. What changed is the cost. A tool that only ever needed to serve an audience of one or two people suddenly clears the bar.
We tend to imagine AI coming for the software we already decided was worth building, the more interesting frontier though might be the inverse: all the software that was never worth building. The niche, the bespoke, the faster-build-for-my-side-project - the long tail of tools that solve real problems for tiny audiences, and that no rational person was ever going to fund with their own time.
Rustkyll is "slop" by someone's definition, sure, but it builds my site in 15 seconds and it exists purely because the price of "worth it" just fell through the floor. I'll take it.