Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
The body language when he asked about the Turing test. You can tell the AI doesn…
ytc_UgzkH1g4x…
G
AI replacement is horsecrap. It’s a great search engine and a way for me to stil…
ytc_Ugz4RHNRT…
G
its absurd he doesnt get that giant point. And also the point that it doesn't ne…
ytr_Ugx2Q6YSa…
G
So, somehow picking mostly male candidates from a pool of mostly male applicants…
rdc_e7jrpn7
G
Please give AI the full access to nuclear weapons and speed up the process pleas…
ytc_Ugwk8Ro9B…
G
I imagine a group of apes creating multiple types of humans.
When the first smar…
ytc_UgwhPgl1g…
G
ETs made nuclear missile silos inactive - AI will be no more of a challenge to m…
ytc_UgxDTdBol…
G
Well if the people behind these mega AI companies are already murdering people w…
ytc_UgyRvdhAJ…
Comment
I like your idea of fighting back, but please don't advertise Nightshade. Glaze doesn't do much either. I'm someone who has done some in depth research behind Nightshade, and I hate to be the bearer of bad news, but Nightshade can be undone. If anyone has been paying attention to Stanford University in particular, they've been looking at both the pro and Anti AI ends of the spectrum, specifically in regards to Stable Diffusion. Nightshade was created by a man named Ben Zhao, and basically passed nightshade off as the end all be all to protecting images, but Stanford stepped in and decided to take a look at their coding, and decided to test if it can be undone, and it can be undone, and they undid it through nothing more than just simple forensics. And if anything Nightshade actually BENEFITS AI. It doesn't automatically turn Ballerinas into oranges which sounds like an entertaining pitch, but at the end of the day, you have to realize that it's just a pitch, and if it can be undone with simple forensics, is it actually as foolproof as we are led to believe or is Ben Zhao just incompetent? I'm thinking it's a bit of a mix of both. Nightshade was trained on stable diffusion, it is familiar with Stable diffusion's algorithm, what is discounted by people is that AI programs all have different algorithms, some may be based on other algorithms, like Deviantart's Dreamup, but even that has its own things that make it unique. I've spoken with the programmers of a British AI app called WOMBO that had problems recently with their quality of their outputs, A few friends of mine play with the app regularly, it has a prompt bar, a chart of styles to choose from, and then a create button, a space where you could put a seed image, and next to it is where your output image will be. What was noticed was when you had in the prompt "pretty anime girl" with the style Dreamland v2 chosen. The output looked like something from one of the earlier styles, implemented into the app vs what the app typically put out. Specifically a style called Realistic. Once it seemed like they got spammed enough, Wombo actually addressed the output issue, and had things back up and running within the span of Two hours. So what happened in this scenario? From there I found myself actually talking with the developers of Wombo, and about Nightshade, they were nothing but nice to me, they were aware of Nightshade, and we were able to talk things out, what happened on their end was they were trying to program in a new style, and things that should not have been crossed got crossed (wires). And what does this tell me? It tells me that the people who make these programs know what it takes to make or break their programs, it's not something like Nightshade, it's the programmers, especially if it's something as simple as getting wires crossed when they should not have been.
Also, the AI bros hate when you go out and redraw their generations, like they REALLY hate it, and I think it's funny.
youtube
2024-07-09T08:5…
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | consequentialist |
| Policy | regulate |
| Emotion | mixed |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_Ugwc1bqN13vIfgWdhpp4AaABAg","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"approval"},
{"id":"ytc_UgzKxebZ5uLVjs4fvId4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugxq_wLYULrjzfbIAO14AaABAg","responsibility":"unclear","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgwwXYkFkjxHPbXHjtd4AaABAg","responsibility":"user","reasoning":"virtue","policy":"none","emotion":"outrage"},
{"id":"ytc_UgzaO5W9MuxGTYTDsuN4AaABAg","responsibility":"company","reasoning":"consequentialist","policy":"regulate","emotion":"mixed"},
{"id":"ytc_UgyTKXhklO4ulM3q8KZ4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_Ugynb-Tb6wWsr57EH2J4AaABAg","responsibility":"user","reasoning":"consequentialist","policy":"none","emotion":"mixed"},
{"id":"ytc_UgwdDERlgM6mFwuBqlp4AaABAg","responsibility":"user","reasoning":"deontological","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyZhtXdwqU9yULVi6V4AaABAg","responsibility":"company","reasoning":"deontological","policy":"ban","emotion":"outrage"},
{"id":"ytc_UgxRxffjqf-luC0pUmh4AaABAg","responsibility":"ai_itself","reasoning":"unclear","policy":"none","emotion":"indifference"}
]