Mastodawn

Cassandra is only carbon now

The thing I actually wanted to say about AI today, before the whole world jumped the shark yet again.

Cassandra is only carbon now 19h ago

Anyway, @zkat warned us. Talking about whether or not AI "works" was a trap, and always was. The ethical component is all that matters, and from that analysis alone, the onus is on all of us to reject and oppose AI.

Getting mired into whether or not it "works" is bad praxis in several ways: it de-emphasizes the ethics, it opens up to goalpost shifting about what it means for AI to "work," and it's easier for the boosters to Gish gallop or overwhelm with jargon.

Cassandra is only carbon now 19h ago

Sure enough, that's where we are now. I'm as guilty of that as anyone, to be sure. But like... all weekend, there have been so many new claims about AI "working," and every one takes a lot of effort to read critically and debunk. None of them change the ethical calculus.

technomancy 19h ago

@xgranade reminds me of this post I read recently about how arguing against a claim involves implicitly accepting its framing

https://anthonymoser.github.io/writing/rhetoric/framing/kirby/2026/01/28/the-kirby-frame.html

The Kirby Frame

writing is a technology, frames are a frame. bespoke human content.

moser’s frame shop

Katerina Marchán 19h ago

@technomancy @xgranade oh I've seen this before but I forgot about this particularly delicious framing

@technomancy @xgranade given how net-unhelpful the "AI safety" people have been in this struggle, i believe "our AI works great and is so powerful, it might just kill us all" is a framing we must not accept even a tiny bit, which requires pushing back on what are essentially technical claims (albeit wild and ungrounded ones) about "how well it works". and again because of the stakes and the harms already done, these become an inseparable part of the moral arguments.

@xgranade One reason many people want to avoid the ethical arguments is because many people won't actually take any action based on ethics. They may claim to have ethics, perhaps even the same ethics you do, but actually changing behavior, holding others accountable, organizing etc. simply doesn't happen.

These things only happen if other considerations come into play.

I don't know how to save our democracy/civilization/biosphere if we can't get people to act on ethics.

@xgranade Example: I begged my Dad to kill his lawn for years. I admit it was a personal vendetta, because I hated mowing it as a child, it's a very hilly property and I feel lucky I never slipped and caught a foot in the mower blades. But because I'm driven by ethics, I laid out all of the harms that lawns cause. Couldn't budge him.

Finally he got old enough he couldn't mow it himself, and no professional wanted to mow the hills either. I suggested he replace it with native plants, boom! Done.

@xgranade I'm a little proud that I finally convinced him, even if it was for the wrong reasons. But because he hasn't gotten into a similar bad spot with his other harmful behaviors, it feels like there is no way to budge him. I have to wait until the harmful behavior somehow harms him. And hope that he doesn't simply deny the self-harm.

@xgranade Just to be clear, I wouldn't have tried making these ethical arguments if my dad didn't seem to care about the same stuff I do.

He loves birds and used to put a lot of energy into birdwatching, so I explained how birds are dying out without the insect life most birds need to feed their babies, and insects depend on native plants. Loving birds was not sufficient to make him take action to help them survive in his yard, until he had a selfish motive.

Fabio Tavano 3h ago

@skyfaller @xgranade in that particular case you had time on your side, environment doesn’t and for the humans that got hurt-by-Ai (except these Who lost their jobs) I couldn’t care less.

@skyfaller @xgranade What I don’t understand is that acting ethically when it comes to AI is one of the easiest things to do. You just…don’t.

@sabrina @xgranade To be fair I don't think it's that easy, the tech industry is forcing workers to use it or get fired, and ordinary people encounter LLMs / generative AI in every product being pushed constantly. For example, every time I open Google Photos they want me to use genAI to make a cartoon of myself or some shit.

Avoiding becoming complicit in slop takes a large amount of effort, beginning with even recognizing something like Google search summaries as slop (one ex-friend refused).

Kwaze Kwaze 14h ago

@skyfaller @sabrina @xgranade

Who is looking for spiritual purity to the degree of "I've never even generated a Google search summary"?

I'm sorry I don't buy this.

I think you'd be hard pressed, even on here, to find someone that would get on the case of people unwittingly using genAI toys shoved in their face in the photos app yet stopped when the issues were explained to them.

No, the problem is people will be made aware of the problems and *continue* to use and defend the toys.

Kwaze Kwaze 13h ago

@skyfaller @sabrina @xgranade
And while you can be "forced to use it" by your employer (as someone who has actually experienced this) you don't actually need to fall in line and comply so readily the way many people do. There is no way to tell you're using it beyond token burn which is trivial to achieve.

Kwaze Kwaze 13h ago

@skyfaller @sabrina @xgranade
There are probably some dynamics here that are similar to what you described with your father but lawns are also a cultural artifact from before he was born. That's an important element you're omitting.

These are toys that got popular 3 years ago.

If someone's been made aware of the harms "Just don't use it" is a perfectly valid standard to apply.

Khleedril 4h ago

@skyfaller @xgranade Too true. How many people drive slower because it is less destructive to the planet?

human_powered 3h ago

@khleedril @skyfaller @xgranade
In 1974, there was a national (US) campaign to drive at 55 mph. It was a federal response to the alarming petrol crisis of 1973.
Jingles about driving at 55 arose!
"55 saves lives."
Trucker slang included 'double nickels' (5 cents and 5 cents -> 55).

That successful campaign -- the national speed limit -- remained in effect until 1995.

It suits my Sustainability mindset to remember 55 mph, although 'drives like an old codger' is an alternate framing.

Khleedril 2h ago

@rabbit74 @skyfaller @xgranade But that was more about economics than ethics.

Lars Marowsky-Brée 😷3h ago

@skyfaller @xgranade When you allow the focus shift away from ethics, you end up arguing against people's lived experience that these tools *do* provide them with genuine utility and allow them to tackle tasks that they previously couldn't, often more efficiently and effectively (from their POV).

Lived experience is very very hard to argue against successfully.

Lars Marowsky-Brée 😷3h ago

@xgranade I agree. That was one of my arguments months ago - GenAI rejection for ethical reasons shouldn't get side-tracked into the utility argument, because that shifts the focus to something one can genuinely argue about, and leads to the ethical argument being dismissed as "uninformed".

But the ethics are the point.

Kwaze Kwaze 19h ago

@xgranade
Tangent to this I can't think of a single harmful product/technology/system where alleged critics claim to know the harms yet feel they have to add "but it works". Has there been anything in history where someone understanding the harms feels the need to insist "it works"?

I can't imagine anyone said about thalidomide "it worked but I hated it". No one to my knowledge has ever sought to "add nuance" to the discussion of coal-fired power plants by saying "yeah but they work".

Kwaze Kwaze 19h ago

@xgranade Most of the time arguments from proponents of destructive systems argue that an *alternative* won't work: "We can't abandon DDT it'd be too costly", "We can't defund police, what's the alternative?", "We can't abandon natural gas, solar can't cover our needs", "I can't eat only vegetables I need protein"

I can't imagine someone ever saying in earnest "yeah ok but be nuanced, it at least works" about any of the above.

@kwazekwaze @xgranade

I would say capitalism

cthos 🐱18h ago

@kwazekwaze @xgranade crypto? Probably crypto.

Kwaze Kwaze 18h ago

@cthos
Aw man, it was right there

Klaus Frank 18h ago

@kwazekwaze @xgranade

Ehm, maybe bad example, but esp. for fossile fuel power plants people say that all the time actually...

@xgranade @zkat @ozzelot idk, it seems like a trap both ways. the correct criticism is that it steals from us all, in multiple ways, and that sucks and we should not be doing it

alanxoc3 3h ago

@xgranade @zkat why not shift the conversation to how we can make it work ethically? less power consumption, ethical training data, smaller models, decentralization, protections from minors... that seems like a more productive solution than flat out rejection. and that would let you actually have a conversation with the "AI Bros" to figure out a compromise.

the technology clearly has useful advantages that could benefit humanity if applied carefully.

Katerina Marchán 1h ago

@alanxoc3 @xgranade this smells a lot like "why aren't you being nice enough to the people who clearly don't care enough to listen to why we shouldn't be doing massively-damaging things?"

The onus is on them to fix their shit. Pointing out how trash their products are is a perfectly valid standpoint.

@xgranade @zkat we're well past the point where people have been harmed and even died by LLM-powered bureaucratic systems, medical advice etc, wherein "it doesn't work as advertised" becomes a supporting pillar of the larger moral argument. the ethical neglect of replacing human care with any kind of automated system becomes more pronounced when that system doesn't work, and they know it doesn't work. not disagreeing with your point i just think "it's broken" has an important supporting role.

✧✦Catherine✦✧19h ago

@xgranade what happened

Cassandra is only carbon now 19h ago

@whitequark No one thing, but a whole swarm of new posts about how *this* model is different and it all works now.

cthos 🐱19h ago

@xgranade @whitequark Like freakin' clockwork.