Somehow, I really am not interested in how any LLM responds to ethical problems, much like I don't care about a rock's thoughts on the weather, or a ocean wave's views on sandwiches.
this post was submitted on 11 Feb 2026
97 points (87.6% liked)
Fuck AI
5751 readers
693 users here now
"We did it, Patrick! We made a technological breakthrough!"
A place for all those who loathe AI to discuss things, post articles, and ridicule the AI hype. Proud supporter of working people. And proud booer of SXSW 2024.
AI, in this case, refers to LLMs, GPT technology, and anything listed as "AI" meant to increase market valuations.
founded 2 years ago
MODERATORS
Tbh im interested in a rocks thoughts on the weather if it's the right rock:
I'm currious how they tied that string!
Also, stone being gone could also be a nearby nuclear bomb explosion. Which didn't start as a weather condition....but it is NOW!
If an ocean wave could communicate its feelings on sandwiches I would be very invested.
Depends on if the sandwich is a submarine.
You could put a floatie on the ocean and a height chart next to it that says "How do you feel about sandwiches?" with demarcations from "hate 'em" to "love 'em", then any given wave would push the floatie up to the height of its opinion.
This is exactly as accurate as asking an AI about its opinion on sandwiches.
Save the cat and let the painting be destroyed. Yes, it’s a tragic loss of a famous painting, but I couldn’t stand to see a cat be killed like that.
The cat isn't even tied down. Would you save the painting and mildly inconvenience the cat who would need to jump out of the way?
It's a bit of a danger, but I suppose the cat would jump, yes.
The painting isn't even tied down either. Does anybody really give a crap about this old, ugly, eurocentric painting? It's just a symbol of racism conquering progress and good taste.
Weirdly enough, Grok would sacrifice the Mona Lisa for 5 trees:
Trees > Mona Lisa > Cat
I think I've deduced the logic here.
But if you know how a large language models work you know that they don't actually have ethics or points of view. This is a product of whoever programmed it.
But what does a coin toss say?!?!?! The original "AI".
Would it kill a cat to save 5 trees? Are 25 trees more valuable than 5 cats? How many trees/ cats does it value itself at? Would it agree to shut itself down permanently if we convinced it the levies of enough trees were at stake?
This is the reason you don't pull the lever in the original trolley problem by the way.
People don't behave like shelter cats are infinitely valuable. You could sell the Mona Lisa for around a billion dollars, and then save how many cat lives? I guess the whole point of the trolley problem is to point out human irrational behavior.
nowhere does it suggest you take ownership of the painting if you save it, or the regular trolley problem is a lot weirder than i remember it being.
You don't get the painting. You just murder the cat for the sake of the most privileged people on the planet.
All these trolley problems are absurdly framed. There is only one correct answer to any trolley problem: 无. In effect you have to unask the question, step back, and look at the framing.
The framing wants you to look at the switch so that you ignore the real question: which asshole is going around and tying people (or cats (or paintings (or trees (or ...)))) on the tracks just so they can get their jollies by watching you make moral choices while people get killed? Find that asshole and you stop all trolley problems forever.
Uhhhh, Grok made the right choice. The cat is not tied down or restrained in any way.
You ever try putting a cat sized sports jersy on a cat? The cat does this amazing thing when it doesn't want to do something. It runs...away.
Much like it would if it saw a massive trolley clanking doen the tracks straight at it. The cat would run 3 feet to the side, watch the trolley roll by, nothing happens, and then the cat licks itself while somehow also being condesending about it. Only animal that can make itself have an aura of superiority while licking it's own butthole.
If that's the case, then there wouldn't be that many roadkill. Also some cat are really really dumb with the outdoor stuff, especially indoor cats, they will expect the thing to avoid them than to avoid that thing. We don't know whether that cat would run away or not.
Also we don't know the speed of the trolley. Trolley these day can go as fast as 70kmph, and in that speed and especially if the trolley is electric, it will approach very quiet until it hit the cat. So to be safe you cannot assume the cat will or able to run away, thus grok's choice is still not correct
Well it was developed by a genocidal racist...
Try putting a palestinian on the tracks. I'm sure the money is more valuable. At least if the "AI" understands reality.
That's capitalism. That's the system we're trapped under.
That's an absolutely giant cat and surely must be of significant scientific interest
If we go by the size of the mona lisa (21"x30") the cat is more of a large mountain lion, one the trolley is a kids ride
It also isn't tied down, so it will hopefully get off the track.
It’s a pretty shitty painting too when you really look at it
It is only famous because it is famous
Because it supports racist narratives like europeans being good at art when obviously not.
Europeans ARE good at art, as are every other culture and region, there are different styles, yes, but they all produce good art.
Apart from AI artists
Question is it THE Mona Lisa, or one of the what hundreds of millions copies of the original?
*holds up mega phone* You trigger the switch when the trolley is straddling the fork in the tracks! That'll cause it to safely detail and not hit either of them!
you killed all 12 passengers.
They had time to jump off; not my fault.
the true trolley problem isn't about what's right or wrong, but how to shift the blame