Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Mock it now, but the technology is only going to get better with each iteration.…
ytr_UgyclakTd…
G
I make ChatGPT remake my art for fun and I’m scared someone will steal my drawin…
ytc_UgwJZJ39x…
G
Only a few people have mentioned it in various articles, but isn't LaMDA only re…
ytc_Ugx3T4nu8…
G
The ai checkers that 'work" like gptzero say papers I wrote years ago were writt…
ytc_Ugzk9slI4…
G
ai right now is making me fucking paniced low-key. i spend all my time learning …
ytc_UgxyaSusL…
G
There are no AI artists. This person featured in the story is a user of generati…
ytc_UgzEmtiZM…
G
simple, dont make robots that need human rights. and stop trying to make everyth…
ytc_UggxBv6Bh…
G
yeah, if it's only running on one computer. a sufficiently smart AI would sprea…
ytr_UgzsgaakS…
Comment
That shows a deep lack of understanding on what algorithms are, and how they work.
If I've found through studying the data that every third time Bob goes to the market it's because he's out of milk. I can use that information to learn something about his milk consumption rates and habits. If Bob wanted to keep his milk consumption (crime) rate a secret AND knew the data I was tracking to learn it, he could game the system by making more or fewer trips to market. There's nothing flawed about my computations. I would be able to very accurately predict Bobs total yearly milk usage... as long as he doesn't know how I'm doing it.
In other words, the system is only accurate precisely because people are unaware of the inputs. As soon as they become aware of the inputs it necessarily becomes inaccurate, and thus unusable.
This is the essence of a "non-flawed" algorithm. Every algorithm simply by the nature of data analysis is useless if:
1. There is some gain to be had by manipulating the data
2. The inputs are known.
So we have a grand new system that extremely accurately determines someones propensity for crime, and "all" they are asking is that we dismantle it completely and go back to guesswork, because certain voting groups benefit from having crime be easier.
reddit
AI Harm Incident
1466984907.0
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_d4p6dpy","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"rdc_j0bm09j","responsibility":"company","reasoning":"unclear","policy":"liability","emotion":"outrage"},
{"id":"rdc_j0avo9u","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"mixed"},
{"id":"rdc_j0a3oov","responsibility":"ai_itself","reasoning":"virtue","policy":"unclear","emotion":"outrage"},
{"id":"rdc_j09s952","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"indifference"}
]