Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G It's some people's job. Professional artists dedicate their whole life to improv… ytr_UgyHPMv05… G This is not new bc We've been hearing about the dangers of A.I since the 70s.… ytc_UgxENnIpi… G I disagree with this man. I think Sam Altman is the best man for the job. Mote i… ytc_UgxxWaZFx… G 1:14:00 i don’t think ai can ever have real emotions. It can say why a joke is … ytc_UgziJrhp8… G I don't like it one thing the world dosnt need is more fakeness and promoting wh… ytc_UgwRLX88s… G Unfortunately for us Corsight AI is based in Israel. It should be banned from us… rdc_nme2044 G This is a terrible use of Ai, gimmicks will not help a failing abandoned educati… ytc_Ugzdyv1Fr… G If the person in the car was actually behaving like a responsible driver and loo… ytc_Ugy-dEfUf…

Comment

So, you really need to talk to an engineer about AI. It is not "getting better every day", we've basically hit cap with what the LLMs can do, and people thinking this is like Moore's Law are just swallowing the PR bullshit Altman and the other Salespeople are throwing out. Take a look at the last few releases, and you'll see that the actual functionality on AI is not drastically improving. All they are doing is adding side-functionality or product integration "look they will recognize a dog" level shit which was already there, just not integrated into their product. AI and LLM is DEAD, because we cannot "code better" than what we already have. These models are black-boxes, we engineers feed shit in and no one knows WHAT the heck it's doing on a granular level. It's just a huge mass of code that, as Todd Howard is so fond of saying "Just works". I work with software consuming MASSIVE datasets for training data, and we've already seen shit breaking because it's consuming AI-generated content and becoming less and less effective. Anyways, rant off, I work in this field, you need to talk to a senior ENGINEER, not one of these Sales Fucks, and ask them how it will improve in 5 years. The answer is "it won't".

youtube AI Harm Incident 2024-07-29T17:5… ♥ 1

Coding Result

Dimension	Value
Responsibility	company
Reasoning	consequentialist
Policy	none
Emotion	outrage
Coded at	2026-04-27T06:24:53.388235

Raw LLM Response

[
{"id":"ytc_Ugxftqtk39vQhOnpxmR4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"outrage"},
{"id":"ytc_Ugzq0A4oVaKN5Bm-sTJ4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgwPNkAV4pY8hqle6CN4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxslErbhBV4tvHJ29V4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxRp2kEymQ_9Zp7luZ4AaABAg","responsibility":"government","reasoning":"consequentialist","policy":"regulate","emotion":"approval"},
{"id":"ytc_UgzhZrdOPJPHWI_TNlJ4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgxXcBLkIrtsSGLM0bV4AaABAg","responsibility":"none","reasoning":"virtue","policy":"none","emotion":"indifference"},
{"id":"ytc_UgwBPs3_IE-n6gS9Usx4AaABAg","responsibility":"none","reasoning":"virtue","policy":"none","emotion":"approval"},
{"id":"ytc_Ugzjgfoxmo8yayyuyFV4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"outrage"},
{"id":"ytc_Ugy_6EAe3vjDTU4CoCN4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"}
]

Raw LLM Responses