The people vs the AI overlords - anarcat
Previously in this series: The Four Horsemen of the LLM<br>Apocalypse.
In a post to oss-security, my (Debian) co-developer Russ Allbery<br>stated that "open source software [OSS] is coming face to face with a<br>motivation crisis that has been building for a long time". His point<br>is essentially that large language models (LLMs1) are making the<br>existing OSS community crisis worse. For him, it's the flood of code<br>reviews, but he argues that varies according to people's desires, for<br>others it's security issues and so on.
I think Russ is right, but I would argue there's something much bigger<br>than our open communities going on here, and it's about the entire<br>field of computing. This pressure is on all of us, regardless of<br>whether we work on open source software or not.
How people use models
People using LLMs in their workflow have radically changed how<br>programming works, even for people who claim to avoid<br>vibe-coding. And I'm sorry to single out one poor maintainer here:<br>it's not you, Brian, you're just one example among many. But this is<br>typical use of those models nowadays:
Once it’s done, I’ll use /code-review and let Claude spawn<br>sub-agents to do a full review of the new code. This usually finds<br>some problems, even problems that the “main” Claude instance didn’t<br>find during its validation. I usually keep running /code-review<br>again and again after finding and fixing issues, until there aren’t<br>any left.
Think about what that means for a minute. This is automation built to<br>fire up dozens of agents crunching at a problem for minutes if not<br>hours of GPU compute time, in parallel. This is essentially a couple<br>of shelves in a datacenter rack, totally maxed out on power and<br>cooling, abstracted behind a cute little /code-review command.
The author, here, is rightly concerned that "Anthropic could pull the<br>rug out and require API pricing", which is perhaps a code word for<br>"charging something closer to actual costs". Brian also pays lip<br>service to environmental and societal costs but those are largely<br>abstracted away, so let's keep that conversation aside here as well,<br>as we have discussed it before anyways.
But clearly, this way of working has an (externalized) cost, to<br>say the least.
Paying for non-free tools
For decades my work has been focused on free and open source<br>software. I've long stopped using proprietary operating systems like<br>Windows or Mac, and even before that switch, I was mostly using free<br>software on those platforms, partly out of principle, but also because<br>I was too poor. So the tools of my trade are free, and I build free<br>tools with them.
It feels like we're going backwards: when I was in school, a millennia<br>ago, my classmates didn't have access to a compiler and were wondering<br>how they would scrape the money to buy a compiler like Borland's<br>or Microsoft's. I had a compiler built into my operating system<br>(FreeBSD at the time), so that wasn't a problem for me. For them,<br>it was a significant expense, but at least those expenses (or more<br>shady sourcing of programs) were a one-shot deal.
Fast forward 30 years, and software is rented: you pay monthly for<br>Adobe's Photoshop and Microsoft's office suite just like you pay for<br>Netflix, Disney+ or Spotify2. And now you need to add dozens (if not<br>hundreds of dollars) of monthly credits to access LLMs on top of that.
So, now we have to pay to get anything done? This is peak<br>enshitification of our job: first they steal our work to train their<br>models, and then they sell it back to us at a profit.
Attacking the engineers
AI is coming for our jobs, as engineers, if not everyone, according<br>to the narrative. For a while now, our job market has deteriorated:<br>less jobs, for less pay. Lots of skilled engineers looking for work<br>and finding crap jobs then still looking while working.
This is not by accident.3 We engineers have a lot of power, it is not<br>organized, but that's just a couple of unions away (easy!). Tech<br>overlords know this, so they are attacking our profession, directly,<br>by forcing us to train and use models that they can control.
Even in environments where programmers are not forced to use LLMs,<br>the mere pressure of other people's LLM-generated work is huge. One can<br>be forced to review LLM outputs, or just peer pressured you into<br>producing more.
We're now supposed to accelerate delivery, because models can<br>presumably do things so much better and faster. With supply chain<br>security becoming such a large vector that we now have worms crawling<br>around developers accounts on NPM, increasing the delivery cadence<br>seems like a really bad idea.4
The LLM hype is part of the larger wave of cyberwar against workers,<br>against water, against the Earth, against all the people. This is not<br>a matter of individually "adapting to the reality" or personal choice,<br>but a political, social, hard problem we need to address collectively.
Previously in this series: The Four Horsemen of the LLM<br>Apocalypse.
I again prefer the term...