Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
The real cost of tokens has been subsidized by VC funding. Anyone who has tried …
ytc_UgxJCMkky…
G
The AI program is designed to replicate human learning. Humains program giving …
ytc_UgziwFosU…
G
That’s what I’ve been saying, I’ve been seeing so many Redditors cheer this shit…
rdc_mtnuvef
G
Everything is shot on one camera. I record my conversation with the AI facing me…
ytr_Ugx98Yd1n…
G
@david.stachon What am I missing in autopilot's promised capabilities? Here's wh…
ytr_UgxqSab4w…
G
There will be a time when the first sentient AI programs will be released into o…
ytc_UgwyWtzJ9…
G
I'm in financial services for a hospital, and there's already an AI bot working …
ytc_UgwXfOfjr…
G
AI is a faster Intellisense that costs millions and will stop being free or good…
ytc_Ugyj-0ODA…
Comment
15:05 Exactly. Hallucinations are as bad as they are because we've *forced* these models to prioritize plausibility (or engagement) over accuracy. They don't have to be right, they just have to be convincing (or enticing) enough to keep you running them!
15:58 Not just that. We see the "reasoning" models output things like "I have to check X. I checked X, it's Y." while hallucinating - because they never saw and don't care what happened when someone did go away to reference other material.
42:40 All its output is role playing. "System prompts" are all about trying to describe the role. This includes when they're tested for alignment; a compliant chatbot will without hesitation role play as your worst fears. And its roles are based on all its training data, including every apocalyptic scenario the data hoarders could find.
*_It does not need to have its own goals or superintelligence to drive us to madness._*
1:01:50 Yes, we've gotten better. But we haven't gotten *fundamentally* better in a very long time. We need to work better than we have at a societal level, because we are not just our biggest threat, we're an imminent threat.
youtube
AI Moral Status
2025-10-31T10:1…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | mixed |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_UgzqSrhcMc5eA-mHUWd4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyRBROBATfSuxlz24B4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgwkI7xjS9FJrr_TCDt4AaABAg","responsibility":"developer","reasoning":"virtue","policy":"none","emotion":"mixed"},
{"id":"ytc_UgzGXyd0KA7vzoJuyxd4AaABAg","responsibility":"none","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxHtVxe1xBrLXLzdLB4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyJ_eeBMPqjd0yPWvR4AaABAg","responsibility":"none","reasoning":"virtue","policy":"none","emotion":"mixed"},
{"id":"ytc_Ugxm0TtDjwAb9x039cJ4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytc_UgwJGNY8IfdyrSHiD6N4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugxov7_kdNf5ZDujqil4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgxK7_q4uQmAz4Ns--14AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"mixed"}
]