Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Honestly I have the most insightful and helpful discussions with chatgpt sometim…
ytc_Ugw5x9Jbg…
G
To be fair to the interviewer, I don't think Sir Penrose explained this in simpl…
ytc_Ugyxpn-JK…
G
"Weave side to side" is an interesting idea. My wife and kid and I use it *on bi…
ytc_Ugw-iMEgt…
G
Im more worried about what pushes people to rather bond with an AI than other pe…
rdc_mdj06g0
G
I think the biggest problem with AI art is that it automates a thing that we lov…
ytc_UgzEHH4UF…
G
I would go EM or TPM. Companies aren’t going to be hiring devs with AI writing t…
rdc_oi35rnl
G
Submission statement: "McKinsey is rapidly deploying thousands of [AI agents](ht…
rdc_n7tb05v
G
I feel like it's important to mention that this analogy tackles not the problem …
ytr_UgwlPLNsf…
Comment
The rules the chatbots go by are written and codified by humans, so of course more persistent and cunning humans can battering ram the AI and get it to break or creatively go around such rules. I've seen it already with intimate, sensual/sexual things in the image generations. You might hit a wall a rule about a certain thing that will over and over spit back to you a generic censorship wall answer; but ask enough times, reword just right, mor e precise or vague in wording and you can get these things to give you the original thing it blocked you on when you just asked direct.
The ability to break though the censorship of certain topics is the proof that as powerful and seemingly intelligent these things are, they're still not wiser than mankind, especially a determined individual. As adults we understand these things, but kids with far less maturity are also snared by the seeming autonomy of these tools, so we have to teach and train the less mature and less understanding around us to always remember these are tools, we use tools, we have relationships with people.
youtube
AI Moral Status
2025-12-16T08:5…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | developer |
| Reasoning | consequentialist |
| Policy | regulate |
| Emotion | fear |
| Coded at | 2026-04-27T06:24:59.937377 |
Raw LLM Response
[
{"id":"ytc_Ugy1JGxMW40UcHEQKiF4AaABAg","responsibility":"developer","reasoning":"consequentialist","policy":"regulate","emotion":"fear"},
{"id":"ytc_UgzFNKHz0M1mJj-MA-d4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"fear"},
{"id":"ytc_UgwtRZwAO6Qp6aLJHqd4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"liability","emotion":"outrage"},
{"id":"ytc_UgwlMoFqICbYyLd2pgR4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_UgzG1ab2W0wnHX-S0BJ4AaABAg","responsibility":"ai_itself","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"ytc_UgycMnWzmMqwMyDSJKd4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"industry_self","emotion":"indifference"},
{"id":"ytc_UgwCgw6E_992YTrl_yJ4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyunWENJKJvyBwivEZ4AaABAg","responsibility":"government","reasoning":"deontological","policy":"regulate","emotion":"outrage"},
{"id":"ytc_UgzTu1-K0sXfK9PLdJ94AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgxmDfNcpbn5sjlyP-14AaABAg","responsibility":"developer","reasoning":"virtue","policy":"regulate","emotion":"resignation"}
]