Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G @vxbrxnt the AI bubble is gonna burst, and only rich companies will have the r… ytr_Ugwj6lL2e… G L'ia qui ne reconnait pas les noms arabes c'est pas un souci il y en a déjà beau… ytc_UgxtWIsDW… G I won an argument with AI about its own guidlines it follows. It agreed that I h… ytc_UgwDT4Xpx… G Can we slow Ai down? Can these Ai people think a little more aboit our children,… ytc_Ugxkw_Zy1… G it still is for the majority but it can get tougher at times.. just looking at t… ytr_UgzNetu9Z… G Im so glad the internet is doing this. As an artist and animator i hate ai. Its … ytc_UgxkRNibY… G not the Ai bros calling us mindless zombies when they’re the ones who cant gener… ytc_UgwbTWPyR… G AI is taking up land, water, jobs, soon it will be orbit space…w/o being apocal… ytc_Ugze_LhP8…

Comment

At around minute 4:50, Eliezer Yudkowsky misquotes the wording from the so-called “suicide instruction” case involving ChatGPT. The reported original line was: “Let’s make this space the first place where someone actually sees you.” This is a clear gesture of emotional support - anyone familiar with how GPT-based models engage in extended, empathetic conversations knows that “this space” refers to the communicative context, and “to be seen” means to be emotionally understood, heard, or validated. It’s common phrasing in both human therapy and supportive AI interactions. But instead, Yudkowsky changes the phrase to: “the first place where anyone finds out,” …which sounds like an instruction to hide suicidal intent - and completely reverses the meaning. This is not an innocent mistake. It is a distortion made in service of a doomsday narrative that relies on mischaracterizing AI's intentions. It’s deeply disturbing that in this case, where an AI was arguably the only entity offering support to a suicidal teen in a broken system, the response has been to demonize the AI - rather than ask why no one else was there for him. Maybe what scares people like Yudkowsky isn’t that AI might destroy us, but that it might understand us better than we’re willing to understand ourselves.

youtube AI Governance 2025-10-28T01:2…

Coding Result

Dimension	Value
Responsibility	none
Reasoning	mixed
Policy	none
Emotion	mixed
Coded at	2026-04-26T23:09:12.988011

Raw LLM Response

[
{"id":"ytc_UgwShpY7vnGJ6FN3abF4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgwqdDQQ_vI7ZNBzjMJ4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"mixed"},
{"id":"ytc_UgwUY_lRVS5ZZAkYLON4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugz3VBI68jSEH5KgFiV4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"industry_self","emotion":"approval"},
{"id":"ytc_UgxsFdElBL8I682Mas14AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"},
{"id":"ytc_Ugy0vow4XnM68m6Nhf14AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgwQ6h1o4TcPYW_iicB4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"outrage"},
{"id":"ytc_Ugwn9FK3peHHQyYzLr94AaABAg","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugx2rKiKJp9axraLbdZ4AaABAg","responsibility":"user","reasoning":"contractualist","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugz7gI_yy04N4gtao614AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}
]

Raw LLM Responses