Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Ai, which stands for Absolute Idiots, which is what people who rely on this stuf…
ytc_UgzNX4ieB…
G
Accept I have the complete right to do anything to or try to get away from a rol…
ytc_Ugx9SN_5S…
G
I read somewhere that there were actual people behind the robot watching and col…
ytc_Ugxx_IYBc…
G
Sophia means wisdom true but wisdom come from God, this is artificial intelligen…
ytc_UgxYuytJQ…
G
These systems are probabilistic prediction engines trained on human data, not th…
ytc_Ugz1fhq5W…
G
This video is a Lie. I asked ChatGPT what a lie is. It said- A lie is essentiall…
ytc_Ugx2k3zp-…
G
🔔 An Open Invitation to Yuval Harari
From the Circle of Natural Intelligence (N…
ytc_Ugz06r_O8…
G
Ew the movements are so weird! I was replaying it for a while going: "Is that ro…
ytc_Ugx4EXX-G…
Comment
This is not an AI issue - it is literally a stupid people using an AI not trained and without tools to argue a legal case. As stupid as it gets. An AI can well do that properly right by having a couple of issues fixed. First, it needs access to a current legal database so it can do research. Two, you must not run a chatbot but a proper AI swarm and correction loop. You can i.e. have the output fact checked (looking up all details), you can have a self-correction loop. Result is like 20 to 30 times the number of AI calls and it is done in an AI. Sadly, this is not a chat program but a proper AI with self-correcting loops. Which - incidentally - exist ;) At least in labs (anyone interesting contact me - demonstrating this in a public product in a month or so). Problem is speed and cost for now (that is really FOR NOW, I expect those to come down in 6-9 months latest). For legal, though, the access to a legal database is an issue - as in: I can provide the tools, but not the access. It is awful how stupid those lawyers were. It is awful that AI refers to training data like that - they should move to fictional training and have AI trained to rely on tools for more. Which also requires more of what is called "attention window" and a proper AI infrastructure, both NOT available in ChatGPT ;) Funny, how things go.
youtube
AI Responsibility
2023-06-10T16:0…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | user |
| Reasoning | consequentialist |
| Policy | none |
| Emotion | outrage |
| Coded at | 2026-04-27T06:26:44.938723 |
Raw LLM Response
[
{"id":"ytc_Ugyw1LJF4a-nzfrFbqN4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_UgzpKeJ69xGhoRCN8WJ4AaABAg","responsibility":"user","reasoning":"deontological","policy":"regulate","emotion":"outrage"},
{"id":"ytc_Ugz_6QuuoLuFf2JTL9x4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyQhU-dhewtt4k_9zp4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"ytc_UgwP9OX8jOeMqbCAWPR4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytc_UgxXWGCEEWv3JHpiA2d4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"},
{"id":"ytc_Ugz0n5opxZQfEMzf5BZ4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},
{"id":"ytc_Ugxdv5RNGoeZ47-mCH94AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"},
{"id":"ytc_UgybAR_CwpOaBE3wp4t4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxBtAqlXQNfCUtGMmN4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"}
]