Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
The discussion on AI vulnerabilities hits home! We rely on Pneumatic Workflow to…
ytc_UgyVBQyAj…
G
If I use claude content directly for my dissertation will it be detected as plag…
ytc_Ugytgcl5C…
G
They're arresting people for crimes they havn't done, but an A.I. has calculated…
ytc_Ugzxv3PY-…
G
AI will be ok if we never let it control anything. Suggest data only. Don’t let …
ytc_UgzGztfQz…
G
"The universe is code", the entire universe is made up of code, and the origin o…
ytc_UgwfKbwLb…
G
I work in and around art. AI doing ideation, I can understand it as a rapid prot…
ytc_Ugx3lznwQ…
G
You were drawing Pukei Pukei here, and I think monster hunter is actually a real…
ytc_UgwNnk6ut…
G
@laurentiuvladutmanea Yeah I see both sides, I think the artist should be handso…
ytr_UgwEX7Ybs…
Comment
4:12 Not quite actually. Modern agentic LLMs are a lot more defined by code than you would think. For example, lets consider a chess playing agentic AI. The agentic AI doesn't reason on playing a game of chess, it launches a docker container with a python fast MCP server that executes the chess python library/package. In this case almost all of the output of the AI is defined by code, the AI merely did a little bit of reasoning to orchestrate the task.
This is actually essential for future AI driven performance. For starters, it greatly improves performance for the same reason giving a human a hammer or a calculator improves their performance at a particular task, and it greatly improves reliability as well as speed. The AI is trained on how to use these deterministic coding tools, but looking at things from a system level you'd see that in many cases the code represented 99% of the actual work with a little bit of AI reasoning filling in the gap, representing the glue that ties in the pieces together into a coherent single structure.
youtube
AI Moral Status
2025-12-11T04:2…
♥ 3
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | mixed |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-27T06:24:53.388235 |
Raw LLM Response
[
{"id":"ytc_UgzfkJjjmVroB0IM8LF4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugw4tnbRxkkmSrEfjLx4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgztoMsIJds3l5aPyIl4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyITxXnLlFhOAHKGBJ4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"mixed"},
{"id":"ytc_Ugy9KFesjcJsMSNVfIt4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgzfcWx1_nPsEF855VB4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgzFKMJ4CLvmP3uxEOJ4AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgzWYagWJhiFibWNY9B4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyQ1L00vAevLsr3o6Z4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"regulate","emotion":"approval"},
{"id":"ytc_UgwKzdJA9OgIy7LnGJ54AaABAg","responsibility":"company","reasoning":"virtue","policy":"liability","emotion":"outrage"}
]