Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G AI takes every fear and threat to mankind/the world and makes them all a real po… ytc_UgxIKY7Qg… G @teopalafox Those non-everyday people who can draw weren't born knowing how to d… ytr_UgwXZzPvj… G Yeah, this problem is less about AI by itself and more about our ourselves and t… ytr_Ugy1nDsZf… G I enjoy your shows with the information you talk about. It's nice to hear your c… ytc_UgxgsAZ-v… G No, you need teachers because a student can have a question. You need teachers b… ytc_Ugz_3Cp3_… G I think some people are idolizing AI and dont see anything else - like they are … ytc_UgyFh3u2Y… G Elon musk still got potential goodness, but "AI" is not human being loaded with … ytc_Ugy5jefjx… G “Most ‘AI takeover’ scenarios ignore how AI actually works in practice—it's not … ytc_UgyK-FT92…
Comment
Absolutely loved the predictions from Sabine, Although they are predictions and may not be 100% correct, but mostly to the point, and I love that we get something conclusive from this channel, you have to be brave to predict. Counter points I would add is, though not human like intelligence but text being the default median of our professions and work, LLMs do have potential in tasks which are non critical (medical or financial) and have verification loops, they can be automated and are being, and yes 90+% of work 90+% of time is not innovating but reproducing, so LLMs can help there as well.. and also when working with trusted sources, like for my own work or APIs I own/trust, there is no risk of prompt injection as with external communication. I am not denying any of Sabine points, just showing a way that LLMs can still be productive and effective if not entirely (which is good, as we keep our white color jobs). Also one point which Sabine didn't mentioned is the limited context window and context rot, which means LLMs don't have infinite memory and all the local context like us and instead we have to load relevant context every time we need them to work on each task, which is a major limitation.
youtube 2025-12-24T02:2…
Coding Result
DimensionValue
Responsibilitynone
Reasoningmixed
Policynone
Emotionapproval
Coded at2026-04-26T23:09:12.988011
Raw LLM Response
[{"id":"ytc_Ugxq7DOfILXcmtB9wGJ4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"indifference"},{"id":"ytc_UgwxBaR_hDLHUlGjbVZ4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"},{"id":"ytc_UgwEtG8t04RNKy6oS3J4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},{"id":"ytc_Ugx7pufsNm1it4fVXDR4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"mixed"},{"id":"ytc_UgyEmFB0_08wWRFkMrR4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},{"id":"ytc_UgzuexsL6QizlmwQSTl4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},{"id":"ytc_UgyuIpkqQ0lJDSWcsER4AaABAg","responsibility":"developer","reasoning":"virtue","policy":"none","emotion":"outrage"},{"id":"ytc_UgwTY_ojOdE8e2FFtJ94AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},{"id":"ytc_Ugzh3-5JI-iIddVoxU94AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},{"id":"ytc_Ugytda5P2ZTHMxXQXO54AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"mixed"}]