Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Ai: "This man is 99.9% likely to be involved in a shooting."
The man: *gets shot…
ytc_UgwRYrbTu…
G
It's a byproduct of the 24 hours news cycle that also tries to be neutral. Inst…
rdc_fap6044
G
Your argument still has flaws.
Human art or hand drawn art is personally crafte…
ytc_UgxgxgUZR…
G
+Sarah Wright Lot of thing. You know hardware are not perfect, sometimes their g…
ytr_Ugj7U0fl7…
G
After making your second point, I actually looked into all of my art and found s…
ytc_Ugy375Q3v…
G
As an American who went to a US college, AI was used so rampantly that I always …
ytc_UgxLx09xH…
G
both the process and ethics of training LLM AI(the basis of all image, text, mus…
ytc_UgzW4p0R7…
G
Una just they play... please let be careful of what we invent nowadays about rob…
ytc_UgxiKz9RZ…
Comment
ChatGPT has gotten better,. but I still don't trust it more than say a 5yr old child. I've worked with it a lot to do Powershell scripts (I have 0 powershell background).. so I have learned some things,.. but I also notice that it often gets stuck in "logic-loops" where it just keeps suggesting the same thing over and over again. Perhaps obvious, but it also can't fix what it can't see,.. for example in a recent API command I was doing, ChatGPT at first laid out putting my plain text Username and Password in the API script.. but obviously I'm not going to put my Password into ChatGPT. My password at the time contained a string "$$" .. which was messing up the logic apparently. I eventually had to ask ChatGPT "Hey, is the error we're fighting happening because my Password has "$$" in it ?.. sure enough it was.
I recently was trying to use ChatGPT to create a simple website "heat map" calendar to show all the release dates of every single IOS version since 1.0.. and boy.. it just refused to do what I wanted. I eventually switched over to Google Gemini AI Studio.. and it was much better for that particular task for some reason.
AI coding is OK.. but you really have to watch yourself and understand its limitations and blindspots.
reddit
AI Jobs
1743799322.0
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | ai_itself |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | mixed |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_mletjn4","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_mlev1he","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"rdc_mlfqbmw","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"rdc_mlgid0u","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_mlgjdk0","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"}
]