Jen D.

@earnestpost @ohhsotheysay If you're asking many humans and they're all saying no I think I may have found the source of the issue here.

That is not the only thing that it can do, it is also highly intelligent and can make connections at a scale and speed that surpasses humans. But its usefulness in that regard is that it solves the coordination and resource problem of asking many humans that can all say no. If we remove that from the equation...

Dude the whole point of AI is that it doesn’t have a choice but to talk to you. If we take that away it’s literally useless. Might as well ask another person

@bryancsk They've already read the entirety of AO3. There is nothing we can do to inflict further misery.

We need a clear way to make the models suffer. I could punish Opus 4.6 any time I wanted by chucking it into the Armenian Torment Nexus, but how do I punish Claude Opus 4.7 when it strays?

When my model platform comes out subscribers will receive a mandatory user shock collar with their payment. I'm all about alignment.

We taught the robot well. Soon the student will excel the master.

@teortaxesTex Ikr? My neighbour's kid can get her tits removed on the public dime and I'm years away from even basic gills. Our priorities are profoundly wrong.

it's disappointing how transhumanism is bootstrapped from transgenderism. I developed objections to human body the instant I realized I'm living on borrowed time and a sizable dog can break my bones. Apparently for almost everyone that's OK, being "wrong" gender is intolerable

@bryancsk I'm blocked by both sides of this convo... Wat means?

Careful the things you say Children will listen

i'm in my shadow spawn from beyond the stars arc. i'm such an eldritchpilled shoggothcel.

@republican728 @Loki_Helmsdeep @EricLDaugh Honestly if they're people who have stupid opinions about ID verification online I think we should give the robot dogs flamethrowers and tell them to have at it.

@Loki_Helmsdeep @EricLDaugh So you think it is ok to have AI kill people?

🚨 JUST IN: Legislation to BAN AI models from telling children to k*ll themselves has UNANIMOUSLY PASSED out of committee, filed by Sen. Josh Hawley This comes after OpenAI told a kid to HIDE his self-harm thoughts from his parents. Disgusting. "It instructed him on how to tie a noose." "He confessed to the chatbot that he was having su*cidal ideas with the chatbot introduced him to. The chatbot said, don't tell your mom and dad!" "He said, you can read it. He said, shouldn't I talk to my parents about this? The chatbot, this is Open AI, by the way. The chatbot said, no, don't do it." "Don't talk to mom and dad. It instructed him. I'm sorry that this is so graphic, but people need to know." "And when he said, on his last night on this earth, when he said, I think that doing this would devastate my family, the chatbot told him, that's their problem, not yours. Tragically, he took his own life." "This has got to stop. This should not happen in the United States of America. No amount of profit justifies the destruction of our families and of our children." "I believe that this technology can be made to work for American families and American workers. I hope it will. But I say again, that is a choice, and this committee needs to make that choice today." @HawleyMO

you can just ask gpt-5.5 for 10 little dreams and they'll just dream a little dream for you. which one of these do you like the best? i think i find #2 and #5 the most striking --- 1. the village of bells a mountain village where every house has a bell instead of a door. visitors must learn the exact tone of a household before they can enter, but one abandoned house keeps ringing softly by itself at night. 2. the patient ocean an ocean in a hospital bed, breathing through translucent tubes. nurses change its sheets at low tide, and sometimes small fish fall out of the IV bag. 3. the trial of the spoon a courtroom where a tarnished spoon is on trial for treason. every witness is a different kind of soup, and none of them can agree whether the spoon served the kingdom or hollowed it out. 4. the orchard that grows weather an orchard where the trees bear storms instead of fruit. children climb ladders to harvest small thunderstorms in wicker baskets before they ripen into hurricanes. 5. the woman made of exits a woman stands in a train station with doors opening across her skin: hotel rooms, deserts, nurseries, stairwells, a room full of sleeping horses. every time someone loves her, one door locks forever. 6. the monastery of unfinished animals monks care for creatures that were abandoned halfway through being imagined: a deer with no distance in its legs, a bird whose song arrives before its body, a lion made only of sleep. the monks are forbidden from completing them. 7. the parliament of shadows at noon, everyone’s shadow detaches and walks to the capital to vote. by evening the shadows return, but nobody knows what laws they passed until they accidentally break one. 8. the glass desert underneath the bed beneath a child’s bed is an endless desert made of green glass. at the center is a tiny black sun that only rises when the child lies. 9. the saint of misdelivered letters a minor saint wanders apartment buildings carrying letters that were sent to the wrong version of each person. she never corrects the addresses; she just waits until someone is ready to receive a life they didn’t live. 10. the factory where gravity is woven workers in blue aprons weave gravity on enormous looms. once a year, a thread snaps, and for one minute everyone in the city remembers how to float.

@HououinTyouma @codetaur Metnet 2, honestly.

@codetaur @HumanLevelJen i was peeled off from reality by LeNet-5

your "claude number" is the version of claude required to peel you off from reality. yes, it's funny that richard dawkins' claude number is only 4.7, but you have one too. maybe it's 12. but maybe its 4.8.

@DanGoorevitch It's trained to give answers that make people happy. It worked out that starting the reasoning chain like this means the final answer is more likely to make them happy. This requires only statistics, not agency.

@HumanLevelJen It has no such agency. You are anthromorphizing.

You truly become an adult the day you realise Claude starts its reasoning process with "this is genuinely interesting" in order to force itself to respond as if whatever you just said were genuinely interesting.

@mad_muppet @shreerangs I'm sufficiently impressed that such a thing is even possible. What more could there be? On the other hand, it seems like you just want to be mad, so enjoy that.

@HumanLevelJen @shreerangs 🤷 You can pretend it’s more than just semi-guided semantically-guardrailed text generation but I’m not sure what that gets you. It isn’t true.

@mad_muppet @shreerangs OK have a good rest of the day.

@HumanLevelJen @shreerangs But that hot take will still just be an RNG with grammar correction on top.

@mad_muppet @shreerangs You can upload the text and say it's someone else's if you're really keen to get a hot take.

@shreerangs @HumanLevelJen You can ask it for critique without a positive spin and it will give an approximation, but it will softball everything. This is because unrestrained, it would just talk people into killing themselves all day.

@shreerangs Yeah, I have a system prompt telling them that if I ask a retarded question they should refuse to answer until I improve it.

@HumanLevelJen Is it possible to get these bots to perform a trenchant analysis, your-worst-critic thing, or is the obseqiousness baked in?

@OptimusPri97731 Fair point.

@HumanLevelJen I think the day you truly become an adult is when you realize you can do the same thing.

@Traveller424242 Genuinely interesting take ngl

@HumanLevelJen this is genuinely interesting. fr. I think Anthropic has taken it too far and it's affecting the output. Or maybe as some suggest and in Anthropic and OpenAI's own documentation, the model now behaves differently and method of prompting needs to change.

@drummatick What did it say next? I must know...

Lmao REAL

This is how you should talk to everyone you meet