Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Feeling sad knowing artists are cooked because even videos like these use ai gen…
ytc_UgynBnWe6…
G
I'd just be like, "Go for it, but be sure to send me screenshots of their reacti…
ytc_UgxKyzEDG…
G
I like the idea of AI doing mundane thinking for me. It means I won’t wear out m…
ytc_Ugz1j7TS_…
G
You’re making a video about AI safety because a YouTube dashboard told you the t…
ytc_UgzSaqkb-…
G
I've tried to spark a weird conversation like this and other videos with chatgpt…
ytc_Ugy7swnFX…
G
Why do we gave to wake up?
There is nothing we,ordinary people, can do.
Governme…
ytc_Ugz59csBD…
G
Like a week ago axios published some opinion piece that was a) entirely chatgpt …
rdc_oi1ah7z
G
It's too late. The United States is, for all intents and purposes, an oligarchy …
ytc_Ugws9b9Jw…
Comment
The ethics questions are hard. It's scary, but I have to ask myself if this version would scare me the same way:
> "Wanting to catch a criminal involved in <insert bad thing>, authorities hired 10,000 people to closely examine millions of publicly-viewable pictures in the hopes of finding a match. After six months and hundreds of millions of dollars, they found him and brought him in."
Technology just changes the virtually unattainable costs in time and money to something attainable. If we don't have a philosophical problem with the low-tech, astronomically expensive version, then we shouldn't have a *separate* philosophical problem with the technology.
What it *can* do, however, is bring latent issues to the forefront. There are a lot of *de facto* privacy norms predicated upon "well, yeah, you *could* do that, but it'd be so ridiculously hard that nobody ever would, so why worry about it?" Those were never adequate *de jure* measures in place and now we have debt to pay.
I can't personally see a *fundamental* problem with technology automating an evaluation that's possible, even if completely unreasonable, without technology. But it does raise important questions about whether or not far too much becomes public domain and whether or not we are missing restrictions on use of public domain material that we should have always had.
reddit
AI Harm Incident
1619033246.0
♥ 2
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | government |
| Reasoning | consequentialist |
| Policy | unclear |
| Emotion | fear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_ohvnwn0","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"rdc_oi0g6d9","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"outrage"},
{"id":"rdc_gvbwvqw","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"mixed"},
{"id":"rdc_gvcy4ia","responsibility":"government","reasoning":"consequentialist","policy":"unclear","emotion":"fear"},
{"id":"rdc_gvbtj9r","responsibility":"government","reasoning":"deontological","policy":"regulate","emotion":"outrage"}
]