Of being thrown out the Overton window

I did not choose to somehow become radical on so-called AI. This is a major reason I am not comfortable embracing the "hater" label. All I did was stand still as the world shifted around me, and stick to my principles as near-everyone decided to toss them aside.

Anyone who's done physics or graphics knows how this works. Motion is always relative, never absolute: stay put on Earth and you're still moving around the Sun.

Falling silent

Picture a scene. You're a programmer, working on your own area of the code. It's quirky, full of tech debt - what isn't ? But you're making it better bit by bit. You're learning new things all the time. Progress is slow, because there is a lot to do and a lot to learn. Every time you are given room to think deeply and come up with a good solution, though, you notice that the next problem becomes a bit simpler.

Next desk over, some clown asks Claude Code to improve a tool that's annoying. It tunnels through three classes that should definitely not get a wider public API. Two of those classes are yours. You have a plan for them, six, twelve months from now. He never asked you before getting started.

If you're lucky, if you have enough pull to say no, you bounce the changes back and tell Chucky McChucklefuck to keep his shit local to proceed with the rest of his work. You will not help if it breaks.

If you're less lucky, if you look like anything other than One Of The Guys, do you get to say no? Do you say no anyway, and watch your coworkers get just a bit colder with you, just a bit more distant, just a bit less inclined to write down the praise you need for your next salary review?

This is poison to team culture. And when marginalized people - women, queers, people of color and so on - are refusing generative AI at greater rates than cishet white men, this is one more way the industry will homogenize.

I get to say no because I am white and look like One Of The Guys despite being openly genderqueer. I get to say no because if I walk then everyone knows very well that the pile of cursed flaming trash I so enjoy rolling in becomes their problem again. Every time I say no I still feel my heart rate ramp up and my stomach twist. How long will I get away with it?

Who doesn't get to say no? Is anyone paying attention to those who are not talking anymore?

Carve your heart open and show me that it bleeds red

I believe I first heard the phrase "There is no neutral on a conveyor belt" from Anita Sarkeesian, in a recorded talk about misogynistic online harassment. A more flippant, maybe more well-known version is "If you're not with us you're against us". Let's stick to the conveyor belt metaphor for now, as a convenient simple image.

Picture a conveyor belt. An escalator. A moving walkway. Whichever, so long as it's something that goes by itself. This is the simple point the expression makes: if you are standing still on a conveyor belt, you're going somewhere. Make sure you are aware of that motion and of your destination, lest you end up somewhere you did not wish to go. Make sure, should you wish to avoid ending up somewhere unpleasant, that you actively oppose that motion.

LLMs for coding are a conveyor belt that many, many people have decided to jump onto. I did not. I simply stayed put: I had no interest in the world it was heading towards, and I had pointed questions about how it was built. This is not, in itself, difficult - because you control the buttons you press.

What has made it difficult is everyone around me pretending the conveyor belt is actually the static ground, that I am the one making an ill-informed choice by refusing to get on, and that all the problems with the conveyor belt are overblown. Much like a lack of AI disclosures inverts the burden of due diligence, the normalization of these mediocre tools in programming inverts the burden of justification. The correct question is "Why would you use this shit when it has so many problems?", not "Why won't you annoying puritans get onboard?".

To a degree my Claude-brained peers are right - I too am making an active choice by staying put. But that choice is not radical and it is not strange and it is not difficult. I have always had a strong moral compass and generative AI being built on non-consent, ecological disasters in service of a deeply fascist project made it trivial to say no. Ethical boundaries should not be elastic, especially not when the end product is a piece of entertainment software. This truth lives in my gut.

There is therefore a particular type of violence in forcing me to argue in favor of staying put. It's a violence I'm intimately familiar with. It's the violence of fellow anarchists questioning the importance of accessibility. It's the violence of fellow programmers questioning the importance of feminism. It's the violence of privileged assholes treating something that ruins lives as a fun hypothetical to toy around with for debate club.

Technology was never neutral

There is a category of "anti-AI" critique that focuses near-exclusively on workers' rights and the evils of capitalism. Three years ago, I wrote something that can be considered part of that lineage: I believe it still holds up but has also shown itself woefully insufficient. As we spiral deeper into this new and hellish era, it is becoming worryingly common to argue that "sovereign" or "worker-owned" or "leftist" so-called AI is possible and desirable.

Counterpoint: no it fucking isn't. If all the workers consent to doing a harmful thing, it doesn't magically stop being harmful.

Following a day of intense Gamedev Discourse, videogame researcher Brendan Keogh published a piece titled AI is history which more or less argues that LLM technology could be good if only it was not tainted by capitalism. I am simplifying. Go read it.

I do not wish to pick on Brendan in particular, the sins here are common ones. However, the piece admits a lack of technical expertise and is entirely devoid of citations. This is probably a consequence of writing quickly and publishing fast enough to be part of the conversation before it drifts off. The result is nonetheless ahistorical apologia which erases the long, rich history of critiquing technology far beyond the role of capitalism.

I am not an academic. I do not keep a citations manager on hand - maybe I should - and I cannot readily tell you what to go read, except for Ursula K. Le Guin's magnificent A Rant About "Technology". What I am is a senior software engineer always reflecting critically upon my craft, someone who has worked with UI and disability justice long enough to not need to think twice before I identify something as a disability dongle.

Disability is only one axis. Race is another one, as pulse oxymeters that monitor blood oxygen levels - popularized during the early stages of the covid pandemic - commonly overestimating values in people with darker skin shows. I suppose you can still analyze it all as a consequence of capitalism even though neither ableism nor racism would disappear overnight were we to abolish that.

Also Molleindustria's post was an admission that the guy had not done even a modicum of thinking about the topic and House House only spilled the beans after someone else told on them so I'm not sure why we're calling those statements "incredibly thoughtful, critical, and reflective considerations of the contradictions of technology under capitalism". Every AI-critical programmer has been laughing their ass off about it all. We are, after all, extremely familiar with how we sound when trying to cover our own ass.

The House House always wins

What can one even do when the Overton window shifts so far, so fast that you find yourself an extremist simply for saying no?

Does that count as going Mad, maybe? A friend tells me there's excellent vampire fiction on the topic. Downsides of being immortal and all that.

This is about Big Walk and also it's not about Big Walk. It never was. There will be more like it. For a moment, I hoped that the statement from IGF jurors meant we could finally have the right conversations. But there, too, I seem to find myself sitting on broken glass despite never getting off my chair.

Further reading :

• This entire fucked dynamic of finding yourself on the edge because the world moved on also applies to taking covid precautions in 2026 so go read some of that I guess.
• On the Big Walk debacle and the IGF statement, Austin Walker wrote a piece that both gives all the context you'll need and is remarkably sharp.
• Following the same Big Walk conversations, writers have covered topics such as "Are you incompetent baboons incapable of understanding basics ethics?" (Bruno Dias) or "So, this is how rules work even though you do not have a giant laser to instantly enforce them" (Damon Reece).
• If you're curious about the politics of technology and disability, I have both Against Technoableism and Disabling Intelligences on my list-of-books-to-buy-and-fail-to-read.