Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
I hate Elon but these accidents and errors provide vital data. Similar to space …
ytc_UgzK7M1kv…
G
As, Seth Rogen recently said, AI won’t replace comedy riders until you jam a bun…
rdc_jk8jb0r
G
Open AI and Google will benefit immensely from AI regulation because it will red…
ytc_UgwYHiaSv…
G
Fun fact. Luddites weren't anti technology, they were against certain types auto…
ytc_Ugwr4peOm…
G
every time someones says LLM's "reason" instead of process or generating we get …
ytr_UgwhJHE0X…
G
So, we can't have self driving cars that could save hundreds of lives and just m…
ytc_UghK9JWzz…
G
I cannot believe this guy is the head of Ted he is so terrible at interviewing! …
ytc_UgzWSN8lh…
G
This was such an interesting watch! The way AI is changing how we work and think…
ytc_Ugx1Nxt-j…
Comment
Overall, this generation of ML has a lot of trouble with para-realism, absurdism, or intentional surrealism. It generates unsettling shit all the time by accident, and we interpret it as surreal. But trying to get it to generate any kind of material that goes directly against expectations is difficult or impossible specifically because _all_ these ML models know is expectations.
You can coach GPT into producing absurdities by telling it the defied expectation ("create a plausible fictional recipe for a beverage called 'dwarven ale' that includes several fantasy ingredients with fantasy names"; "describe a scene of a married couple arguing, and then some kind of unexpected monster appears")... but it's hard to get it to spontaneously violate expectations, as would be required for a joke.
Midjourney similarly struggles with para-realism. For instance, I have struggled enormously to get Midjourney to create an image of a [large boulder floating over a jungle canopy](https://forgottenrealms.fandom.com/wiki/Earthmote). I have tried _extensively_, and it just absolutely insists on connecting the boulder to the ground... because it has never seen an image of a floating boulder.
reddit
AI Jobs
1683133316.0
♥ 14
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | ai_itself |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[{"id":"rdc_jipmvgi","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_jipwej6","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_jipw0oi","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"rdc_jipxhbw","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_jiqpbdd","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"}]