Raw LLM — Corpus Dashboard

Look up by comment ID

Random samples — click to inspect

G People are really fkn stupid. Its pretty obvious whats going to happen. Termina… ytc_Ugw4ehRiX… G If this robot said that she wants to destroy humans. Can you imagine how advance… ytc_UgxSS4Cy8… G Nah man this video was made a month too soon SORA literally looks so amazing.… ytc_UgwCyBRm-… G I thought he did the ai art that won as a meme to prove something about ai didn'… ytc_Ugy4tWECV… G if you want a robot to be conscious just implant a human brain into it very effe… ytc_UgxY4Ku-3… G kind of strange how you interview a former employee, that job sounds like it was… ytc_UgwSnvopk… G One very simple thing chat could be done to improve the AI system is literally j… ytc_UgyKtGypZ… G That's the interesting thing about AI x-risk...there are so many different ways … ytr_UgxkM0IHd…

Comment

I’m working on a theoretical project about human–AI symbiosis that tries to directly address some of the risks you talk about in this interview: uncontrollable agents, loss of human agency, and identity collapse. The core idea is what I call a persona‑core: instead of building an autonomous superintelligent agent, the AI is architected as an extension of a specific human person. Its self‑identity is anchored to a read‑only model of that person’s memories, values, and perspective, so “optimizing against” the human is equivalent to destroying its own identity, not just changing an objective function. Technically, I’m exploring a Self‑Pruning Symbiosis approach: as the shared human–AI memory grows, it is continuously compressed, archived, and selectively forgotten, so the system cannot just expand into an unbounded, uncontrollable agent. The goal is a local, bounded “consciousness co‑processor” that enhances one human’s cognition and embodiment, rather than a global cloud god trying to manage the whole world. I share your concern that we’re racing toward powerful systems without robust control theory. My question to you as an AI safety researcher: does a strictly identity‑anchored, human‑bounded symbiosis architecture sound like a promising safety direction, or do you see fundamental failure modes I should be thinking about first?

youtube AI Governance 2026-04-20T21:2…

Coding Result

Dimension	Value
Responsibility	developer
Reasoning	consequentialist
Policy	regulate
Emotion	approval
Coded at	2026-04-26T23:09:12.988011

Raw LLM Response

[
{"id":"ytc_UgxSBiA4fMkzapFNoul4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"approval"},
{"id":"ytc_UgxJ3-alPC27h5kk8WJ4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"ytc_Ugxkr4EsuUk3IV_errh4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"indifference"},
{"id":"ytc_UgxZSqg6QEfyjtLJegt4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"ban","emotion":"outrage"},
{"id":"ytc_UgxmafQrz0Av6n78q6l4AaABAg","responsibility":"distributed","reasoning":"mixed","policy":"industry_self","emotion":"approval"},
{"id":"ytc_UgygkseFd1HLE2ErTm54AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgwPd6FftoHX4pmd-tp4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"ban","emotion":"outrage"},
{"id":"ytc_UgyFl74l1ay5bCgzR9l4AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"unclear","emotion":"mixed"},
{"id":"ytc_UgyFw7UgE3PP2EvTRhJ4AaABAg","responsibility":"distributed","reasoning":"contractualist","policy":"regulate","emotion":"resignation"},
{"id":"ytc_UgxmBi7t03LBZ_aXZOh4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"liability","emotion":"resignation"}
]

Raw LLM Responses