Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
So this question never crossed anyone's mind, when they heard of self driving ca…
ytc_Ugx8wbQwU…
G
@Girito26 Thank you for your comment! Your insight on time travel in sci-fi movi…
ytr_Ugye91Dca…
G
I'm a fantasy author, and an artist - and I have a best friend who told me, to m…
ytc_UgwHkzlJM…
G
I was just gonna comment this.. he’s so uncreative that he can’t even write out …
ytr_UgzOTC2_C…
G
Yes, BUT-
We cannot proceed like this isn't going to be developed both privatel…
ytc_UgxVCGYdb…
G
I doubt an AI could get an ACOG right so i got the soldier right…
ytc_Ugy5xCO6F…
G
With both traditional and digital, you're using your own hands, fingers, muscles…
ytc_Ugz3s7K5W…
G
Thank you for your comment! In the video, Sofia indeed prioritizes efficiency ov…
ytr_UgyYKNJ01…
Comment
A lot of the addictive nature of short content is in the overwhelming quantity of content available, combined with the uncertainty of what you will get next.
It's essentially the psychology being the Skinner Box. You will be more determined to swipe if you don't know if the next video is going to be good. You watch a video, wondering if this one will scratch that boredom itch. Then when it doesn't, you watch the next. You don't want to put it down, because your brain is telling you that it could be the next video that entertains you. Because there is so much content out there, you assume at least some of them are the good ones.
This concept also applies to things like Reddit, or social media.
So how do we fix this? You could limit content available. That is plausible, but difficult. You can't exactly pretend most of the internet doesn't exist.
What you can do, is curate the options better. Remove that constant search for the next good content. AI could select videos and articles that you want to watch or read, and deliver them to you directly.
This could also tap into laziness directly, saving the effort of searching for new content. Which might not seem like a lot. But, watching a video for more than a minute didn't seem like a lot, until we were presented with other options.
This does have some caveats. First, it assumes someone only wants to watch feeds out of this Skinner Box addiction. If someone binges TikTok for other reasons, like for social reasons, then that is another issue. Although, I'm sure there are ways to make AI curated content social as well.
It also assumes people would want to move away from short form content, and actually prefer longer and more substantial content. I can't actually prove this, it's just a gut feeling based on my own experiences.
reddit
AI Responsibility
1761172371.0
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | consequentialist |
| Policy | unclear |
| Emotion | mixed |
| Coded at | 2026-04-25T08:13:13.233606 |
Raw LLM Response
[
{"id":"rdc_nkbdn33","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"rdc_nkhlkpz","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"},
{"id":"rdc_nkuzhuq","responsibility":"company","reasoning":"consequentialist","policy":"unclear","emotion":"mixed"},
{"id":"rdc_nlw5ej3","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"outrage"},
{"id":"rdc_nm1q0hj","responsibility":"company","reasoning":"consequentialist","policy":"unclear","emotion":"fear"}
]