Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G there are a lot of things that possess possiblities to harm human, for example s… ytc_Ugz-jNJ6X… G The problem with Texas is it a massive state with a whole lot of boredom involve… ytc_Ugzl1W9m2… G This has been an ongoing scam for decades, now just with AI. They tried the same… ytr_UgxH81UOg… G AI : let us create beautiful better works of art than u could ever hope to make … ytc_Ugzj303WY… G AI is already watching it. And the reason why it can be dangerous might be becau… ytc_UgwmbjuP-… G I care about AI because as a technologist, I love learning about and applying em… rdc_o8gno10 G IT DOESN'T MATTER if AI creates 133 million new jobs. AI will necessarily be abl… ytc_UgwuLsNHo… G He’s definitely using ai to respond to the questions. Let’s take this one by one… ytc_UgxosUPSa…

Comment

These are frustrating takes. The story of the boy's suicide was tragic, yes. But, it's been magnified and discussed as if it is the single representative artifact defining OpenAI's models. OpenAI's got some shitty PR team cuz they are definitely the "bad guys" of the top AI labs. Dario Amodei *discusses* AI safety a lot, but Claude is the more dangerous model by a long shot rn. It can--without much effort--get to writing an entire malicious codebase in a prompt or two. Chinese hackers are already using that "feature" of the model in organized cybercrime. You know the scale of repercussions along those lines? ...At scale, easily accessible, and high-quality malicious code is a greater magnitude problem by many orders of magnitude. Getting the GPT reasoning models to do that would be far more difficult. And their models are generally harder to trick into bad behaviors than Claude or Gemini, even in topics unrelated to code. I think Claude is a fuckin' awesome model with the best personality and incredible capabilities. Yet, it's still a weird, almost propogandistic narrative going around that somehow Anthropic's more concerned with safety and OpenAI is the careless greed factory. No, Anthropic's just got better PR. Damn, even my man Geoffrey Hinton sees it in a way I view as patently wrong.

youtube AI Governance 2025-12-30T11:2…

Coding Result

Dimension	Value
Responsibility	company
Reasoning	consequentialist
Policy	regulate
Emotion	mixed
Coded at	2026-04-27T06:26:44.938723

Raw LLM Response

[
{"id":"ytc_UgwS7ZaYErYtb2o844B4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"liability","emotion":"outrage"},
{"id":"ytc_Ugzed0s3l6lWq0KyhI94AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"outrage"},
{"id":"ytc_UgwuYTzyHYJzGyljuPd4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"regulate","emotion":"mixed"},
{"id":"ytc_UgxmTOCRb1tlaDWFxNR4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgzjEfw4jMnyn28cevN4AaABAg","responsibility":"government","reasoning":"deontological","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgwnkJwzlRZz_Wiz1SZ4AaABAg","responsibility":"user","reasoning":"mixed","policy":"industry_self","emotion":"approval"},
{"id":"ytc_UgxRbbrybpv42njlQqd4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgwlJtsxGhikR0n9xDN4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"ban","emotion":"resignation"},
{"id":"ytc_UgwdFbKxnavTBJazV6R4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_UgyTpFe0HyBispJeo654AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"unclear","emotion":"fear"}
]

Raw LLM Responses