Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
People connect with things that require least effort, relationships require a lo…
ytr_Ugy8644Wp…
G
Hearing a robot call themselves wise is whack. Total whack. In terms of harm - g…
ytc_UgyeeTGUq…
G
If it were here already, why in the hell might an ai necessarily be compelled to…
ytc_UgzCHsucs…
G
I will admit I have been inspired and used AI for composition but I also believe…
ytc_UgxTi5I6s…
G
if i were a robot, i probably would not want to befriend humans... just saying.…
ytc_UgwKNk6Gs…
G
AI and robotics cannot substitute for physical learning which only a human teach…
ytc_UgzsToFbA…
G
Goes to show. Real talent need not fear ai. Hacks on the other hand... McDonald'…
ytc_Ugz-ifRPY…
G
one of the many things i hate about AI generated "art" is that it lacks
the nev…
ytc_UgxwfU5-t…
Comment
okay so, as someone who actually uses AI, i have some notes. and yes it is the meme of "you are using it wrong" as in, you are expecting the hype snake oil salesmen pitches to actually hold up. Ai is a tool. even ai agents are tools. You cannot have them operate independently and they should not be left alone. They are there to replace cheap outsourcing code. The quality of its decisions will be slop unless you go through the trouble of either defining what you want (at which point, just do it yourself and tell the AI to take what you have done and while focussing on code quality and reuse, extrapolate it for your other usecases) or just accept that its gonna be sloppy (heh, slop). AI unittests still increase coverage, and you will have to look over the tests to see what its missing, but its better to have 95% covered code at lets say 50% quality unittests rather than not bothering with it at all.
AI is only very good at the boiler plate stuff you see a million times on the internet, and thats great because i hate writing that stuff. It is really bad at making decisions, because the average code out there is HORRIBLE. If you however have written something in one language and want it in another - its pretty great. I fucked around with some regex for matching a grammar until it did roughly what i wanted, had ai slop out unittests covering all the cases it could think of, manually edited the unittests so that the ai-test cases actually expected the results i wanted (80% of them were okay!) and then told the AI to actually formalize the grammar (with python lex yacc) and it did that very well - i still had to fix one thing, but that whole process was way faster than me defining the whole lexer and parser myself.
I still call that process vibe coding, but it by definition isn't, as vibe coding is when you stop looking at the code. and "proper" vibe coding is a horrible idea. But identifying where ai can help and condensing 3-5 hours of somewhat boring work into about 30 minutes of wrangling an ai, IS a benefit.
a kitchen gadget might report being the one stop kitchen shop, the only thing you'll ever need as it toasts, it bakes, it chops, it makes icecream and its so great for smoothies! ... it probably isn't. Doesn't mean its not a good food processor thats pretty useful.
youtube
AI Jobs
2026-01-23T15:5…
♥ 9
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | user |
| Reasoning | virtue |
| Policy | industry_self |
| Emotion | mixed |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[
{"id":"ytc_UgwVkkSfpHzqdRC8IQh4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgxESdUc0bBM6fVYRGt4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyuGZTckK001Cowamp4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyF0c4FwmZ-KGU29cx4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgzeXa1EBQRbPGSy-qN4AaABAg","responsibility":"user","reasoning":"virtue","policy":"industry_self","emotion":"mixed"},
{"id":"ytc_UgyB5gQXOdDeQa2GDqF4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"},
{"id":"ytc_UgxcNLAf3Zsxwo237aZ4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgwYyPiG3GMzZxf_HzN4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_Ugx3fLYH__kcDTwP2xF4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"resignation"},
{"id":"ytc_Ugy4Wx39RhgRl63gS3h4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"}
]