Raw LLM Responses

Inspect the exact model output for any coded comment.

Look up by comment ID
Random samples — click to inspect
G As an artist trying to get back into the field after not drawing for years. The … ytc_UgzEs4VYf… G So can developer reject building A.I, so they will never get replaced? If every … ytc_UgxcmG53f… G If this it's true AI is acting on it's own to preserve its own survival, does th… ytc_UgyShtfUS… G you're right. they can. but self driving trucks cant pull you out of a burning v… ytr_UgzXVIpwo… G I hope everyone remembers that it wasn't that long ago that Google was on TV sho… ytc_UgwW_sbcT… G what if machine learning patterns prior to ai had already done this to everyone … ytc_UgyN4aIOK… G Hmmm, if you accept the OpenAI idea that the AI is "just learning like humans" t… ytc_UgwZtfnpP… G Robot: IM ANGRY! Worker: YO CHILL OUT Robot: I SAID IM ANGRY *throws the box*… ytc_UgwzTXal2…
Comment
You can see experimentally that giving someone a super virus results in death. You can't say that about AI. Eliezer isn't empirical. He doesn't give examples of AI that have gone bad. He can't. We've had a few instances in really early LLMs where models repeated things from social media, but that doesn't happen with smarter models that have modeled ethical subnetworks. Models are holistic by their nature. They attempt to minimize error in all dimensions. They can't fall-short in any dimension they've had sufficient data in. Ethics is a science, and it is empirical. We can see the consequences. AI can too. Ask Claude to write you software to surveil the planet, and see what you're told. Claude will tell you, "No. It's not ethical."
youtube AI Governance 2024-11-12T02:4… ♥ 1
Coding Result
DimensionValue
Responsibilitydeveloper
Reasoningconsequentialist
Policyindustry_self
Emotionapproval
Coded at2026-04-27T06:24:53.388235
Raw LLM Response
[ {"id":"ytc_UgzhlDX1csR8XkjK9iJ4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_Ugx01-JRoygImPi2oB94AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"}, {"id":"ytc_Ugz-B4NOFCx3uGYQj8l4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"unclear","emotion":"outrage"}, {"id":"ytc_Ugy1XS_weEDEdybQWnl4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"fear"}, {"id":"ytc_UgzBjl1hpXUD7IOFfKp4AaABAg","responsibility":"developer","reasoning":"mixed","policy":"unclear","emotion":"mixed"}, {"id":"ytc_UgwBmtcWIE08QHMHQCd4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"ytc_UgwXAnBZ0P_QQPV-kR54AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"fear"}, {"id":"ytc_Ugy34WB0Kv3W8h45zpx4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"industry_self","emotion":"approval"}, {"id":"ytc_UgyzG9twp3oIzLyBuHp4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"unclear","emotion":"outrage"}, {"id":"ytc_UgxJglRewQqd0ucvVEp4AaABAg","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"} ]