Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G The algorithm acknowledges that socioeconomic background affects crime, yet they… ytc_UgwyfVPjV… G why do you think most of them built bunkers all the silicon valley bros some of … ytr_UgxKanjqX… G Windows 11 will be the last "Windows", instead of Win12 It will be called Copilo… ytc_Ugzjr2BnY… G I've always liked watching and yes them it's an explosion of AI one of the bigge… ytc_UgxMySbWW… G Humans having control over AI super intelligence is as impossible as monkeys hav… ytc_UgyhnlqGz… G I imagine that the day AI truly gains consciousness this guy will suffer... 😂 Th… ytc_Ugzghlv8_… G We really should stop calling them "Artists" when all they do is write some prom… ytc_Ugz1gSk4X… G Better just build my own agent since I can download all the materials I want aft… ytc_UgwdwU0la…

Comment

So, you're saying that being trained on human output doesn't make the the AI human-like "at all", because "what human has read all the text?" I don't see the logic in that. If you train a neural network on the output of another neural network (the human brain), then it will become similar to that neural network. It's good to hear the idea that "AI safety is dangerous", becuase that's exactly what I thought when a long time ago I first read Eliezer write about the impossibility of safely aligning the utility function of a super-human AGI. I think that designing an explicit utility function and then trying to make the benevolence mathematically provable, is a dangerous approach. This approach is a self-fulfilling prophecy. I think it's much safer to just train the AI on human output, and ensure its humanness that way, without ever defining an explicit utility function.

youtube AI Governance 2024-04-10T19:2… ♥ 1

Coding Result

Dimension	Value
Responsibility	unclear
Reasoning	mixed
Policy	unclear
Emotion	mixed
Coded at	2026-04-27T06:24:53.388235

Raw LLM Response

[
{"id":"ytc_UgwHB3zmkdgBtRd6Y5h4AaABAg","responsibility":"company","reasoning":"virtue","policy":"liability","emotion":"outrage"},
{"id":"ytc_UgxdNrV3cv1eq5RZPMt4AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"ban","emotion":"fear"},
{"id":"ytc_UgzYM1xYeWovRXAzHrt4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_Ugwahq6aDz5bhamDiGd4AaABAg","responsibility":"unclear","reasoning":"mixed","policy":"unclear","emotion":"mixed"},
{"id":"ytc_Ugyog867X5QH-LdCyzF4AaABAg","responsibility":"company","reasoning":"virtue","policy":"liability","emotion":"outrage"},
{"id":"ytc_UgzTjacYjEIQjjTV5pR4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_UgzsyJlZUKZvhx3p-qp4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"industry_self","emotion":"approval"},
{"id":"ytc_UgwEGfSsJcVXZou2CA54AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_Ugy14FnnIIqGD6dVWat4AaABAg","responsibility":"government","reasoning":"deontological","policy":"regulate","emotion":"mixed"},
{"id":"ytc_Ugz0zx2y9HYuhxrDS1V4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}
]

Raw LLM Responses