Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
It is not AI, that is just martketing. These things are LLM's (Large Language Mo…
ytc_UgwDBHS5y…
G
I'm not making any AI art and I never will, yet I'm certain that I'm unable to e…
ytc_UgzinRwKP…
G
I know this is mildly (almost completely) unrelated to the video, but every new …
ytc_UgzY9inhb…
G
Alright, he’s terrified and quit. Can I be hired by OpenAI and take over his for…
rdc_m9ijp9w
G
How do I know I am not an AI programmed to believe I am a human?…
ytc_Ugh9Koh_R…
G
9:43 This exactly the paradox I was thinking about. Elon says there will be abun…
ytc_Ugw9bVWSF…
G
Eventually AI will be used to expose the mathematical flaws in the equations sci…
ytc_UgzxKYwVP…
G
Robots and AI cannot fully replace humans. The planet doesnt have indefinite res…
ytc_Ugxyr80z1…
Comment
Intelligence doesn't require consciousness. GPT-4 is intelligent enough to serve as the backbone of AGI. We can easily design software architectures in which GPT-4 engages in inner monologues, rewrites its goals, and stores memories. However, we currently can't concentrate the necessary computing power for such architectures to be effective.
There are experiments where one GPT-4 instance communicates with another to evaluate its answers. In some scenarios, dozens of GPT-4 models propose different answers, with another instance selecting the optimal one. Imagine millions of AIs debating a question a thousand times while continuously searching for information on the internet.
LLMs are already becoming more compact. Open-source models, such as those based on Meta's LLaMA and comparable in capability to GPT-3.5, can already run on personal hardware. Instead of bigger models, we might need models that are more efficient and at least on par with GPT-4 to achieve AGI. If such an AGI manages to autonomously replicate itself on other machines and utilize their computing power, it could mark the last chapter in human history.
youtube
AI Moral Status
2023-08-20T20:2…
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-26T23:09:12.988011 |
Raw LLM Response
[
{"id":"ytc_UgwXFoN6YIzMoo6q-2B4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"ytc_UgwrJI0swBo5SATFRHl4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgwQ3xc6lgzXthOAbDN4AaABAg","responsibility":"developer","reasoning":"deontological","policy":"none","emotion":"outrage"},
{"id":"ytc_UgxcQUDVLu0Xj1N6gOl4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_Ugxm9d8aojNUrriw-Bl4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgyoouMwNH7GD6fgccd4AaABAg","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"resignation"},
{"id":"ytc_UgzgCuaOvpmqiH7Ae2F4AaABAg","responsibility":"ai_itself","reasoning":"consequentialist","policy":"none","emotion":"fear"},
{"id":"ytc_UgzAqH-TbytQyZYDELp4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},
{"id":"ytc_UgyUtAeurvHWLzHAJ3h4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"mixed"},
{"id":"ytc_UgxlrOm--qvgsOG3Jgd4AaABAg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"resignation"}
]