Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
The tavern ai user:
I mean getting the API was annoying cuz clewd kinda a bitch,…
ytc_Ugw8oy9pv…
G
It's pretty weird to defend it by saying the algorithm will just cure the poison…
ytc_Ugw0TE9Qf…
G
Are you guys not listening the way she asked the question. Jesus Christ is part…
ytc_Ugzqs6F5X…
G
Our enemies don’t stop using AI just because we stop. They would not enforce law…
ytr_Ugz7fFJTI…
G
Remember this is AFTER most ai chatbots blatantly gave different responses betwe…
ytc_UgzozLI5n…
G
I don't have a lot of friends to talk about my goals and future, sabki alag prio…
ytc_UgxJVeZyW…
G
Wont the AI just build its own software? You'd need to pay for APIs and cloud st…
rdc_oh36k8s
G
@Motivational-Mango A lot of things are an extension of humans - but can’t expre…
ytr_Ugzn6Rhy5…
Comment
AI cannot cause mass casualties. It is simply not possible for any current AI to CAUSE any of those things, because GenAI only produces text and images currently, unless someone puts it in a mobile chassis of some kind.
Of course, people who manufactured steak knives would be furious if they were to be held liable if someone cuts someone's junk off with them, and rightly so... And would box cutter manufacturers if they were to be held liable for when people hijacked a plane using one.
I would be fully supportive of laws that would require licensure for humanoid robots, and which would place liability for action on the licensee and require insurance to some high liability for misuse...
But liability ought be borne by the person who has responsibility for creating substantial physical action outside the isolated states of a computer system's memory or pixels on a screen, and controls ought be placed on *actions* pursuant to other *actions* rather than mere twiddlings of bits and lightings of pixels.
A gun is dangerous because it is the last and final tool necessary to allow ANY agent of ANY kind capable of holding it to do violence.
An AI can, of course, teach someone how to make a highly radioactive nuclear reactor in their shed. *So can the Wiki page on The Nuclear Boyscout*.
What it cannot do is actually make it any easier to make that nuclear reactor because *nuclear materials are tightly regulated and scanned for*.
For pretty much any dangerous act, the reason they do not happen is not because the act itself is obfuscated or arcane or esoteric in any way.
The reason dangerous acts are difficult to make happen is because *we control the precursors heavily*.
That's how you prevent dangerous acts... By regulating chemicals and materials, not by regulating glorified half-baked textbooks.
reddit
AI Responsibility
1724537707.0
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | user |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_ljr04dm","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"rdc_ljr5ox8","responsibility":"ai_itself","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"rdc_ljrrwo9","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_ljrs58c","responsibility":"distributed","reasoning":"virtue","policy":"unclear","emotion":"outrage"},
{"id":"rdc_ljuigro","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"}
]