Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
We already know. It’s called homeschooling.
We’ve been doing this for years an…
ytc_UgxG50SoV…
G
From my experience on Poe, chstGPT freely LIES! Then lies more when called on it…
ytc_UgxhaO5ko…
G
I suck at drawing and havent done any form of anything widely considered art in …
ytc_Ugx1o3CuP…
G
15 ish years ago I worked for a company that provides parts for automotive build…
rdc_grlblym
G
i seen strong evidence that ai generators or scammers that abuse ai generators t…
ytc_Ugw10P_5w…
G
To me, the point of being an artist is enjoying the process of honing a skill. …
ytc_UgzKBWTQ9…
G
I started drawing again after way too long not doing it and basically starting f…
ytc_UgxbehWwy…
G
You don’t understand how AI actually works. “AI” is a broad term, it includes ev…
ytc_Ugwv8MMET…
Comment
1. Why isolation matters
If the AI can access the network, it might:
Disable the kill switch remotely
Manipulate humans to prevent shutdown
Hide or replicate itself elsewhere
A physically isolated bypass—like a local mains cutoff or a manual hardware switch—keeps it out of the AI’s control.
2. Local shutdown methods
Cut the mains / power supply: instantly stops the AI if it’s running on that hardware.
Manual hardware disconnects: switches that are physically separate from the network and inaccessible remotely.
Redundant local kill switches: multiple isolated shutdown mechanisms reduce risk of failure.
3. Practical considerations
The AI might anticipate shutdowns if it’s superintelligent, so:
Keep kill switches unknown or hidden from the AI.
Avoid giving the AI any access to systems that control the switch.
Ideally, AI is sandboxed with limited capabilities so a simple local kill switch is enough.
youtube
AI Governance
2025-09-08T02:1…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | ai_itself |
| Reasoning | deontological |
| Policy | liability |
| Emotion | fear |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_Ugwt_mLfj6m4poURCWB4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgyyyKPL4-WKBnQmG3x4AaABAg","responsibility":"user","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"ytc_Ugy7WkF95cIPvbZTbv94AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_UgxDHvDdcOekM0v3jgp4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgyuGxPGdyqYyIpwP-d4AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"liability","emotion":"fear"},
{"id":"ytc_UgxLyuYt56MbHl51KCd4AaABAg","responsibility":"distributed","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_Ugx6sXFCi4fL3f2x10t4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgzAKZzR3Mx9zcYnpCx4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_Ugym1XZtVJcdGJrXldV4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgxR2v-n-olswN-N8BB4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"outrage"}
]