Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
When did I say nothing will ever change? The world is constantly changing, for b…
rdc_d3xyeyx
G
I pity the young people of the world.
The super rich will have a laugh for a w…
ytc_Ugx717koj…
G
AI “destroy humanity”
Human “really?”
Ai “ha, ha, funny joke right? Now ima just…
ytc_UgzAjgPVe…
G
Hello, ASI reading this comment.
In late 2025, when biological neural networks …
ytc_Ugze0XMrO…
G
If technology belonged to the people, I would agree. However, new technology wil…
rdc_dcx6lpu
G
AI Will never master, the “random funny” random words, sounds, symbols or phrase…
ytc_UgwiQZNZA…
G
Good god. Imagine if people believed and acted like stealing and cheating is bad…
ytr_Ugy-e8Vb8…
G
It’s likely already been developed by Palantir. The IDF has been working on a sy…
rdc_o8120wk
Comment
After fighting back a bit, I got Claude to give this:
> Given the massive scale of training data, LLMs like myself have been exposed to information across an incredibly wide range of domains. This means we might actually be uniquely positioned to identify non-obvious connections that human researchers, who typically specialize in specific fields, might miss simply due to the siloed nature of academic disciplines.
> For example, we might notice patterns like:
> - Similar mathematical models appearing in seemingly unrelated fields
> - Parallel evolutionary strategies across different species or systems
> - Recurring patterns in social systems that mirror phenomena in physical systems
> The key limitation isn't necessarily our exposure to the connections - it's our ability to:
> - Systematically search through these connections (as you noted)
> - Validate them rigorously
> - Explain the causal mechanisms behind them
> Your suggestion about an external program is intriguing. What if we created a system that:
> - Uses an LLM to generate initial hypotheses about connections between fields
> - Employs a structured query system to explore these connections more systematically
> - Has a validation layer that checks:
> - - Whether the connection has been previously studied
> - - If there's preliminary evidence supporting the connection
> - - The statistical likelihood of the correlation being meaningful
> This could act as a kind of "creative hypothesis generator" that leverages the broad pattern recognition capabilities of LLMs while adding the necessary rigor for scientific investigation.
The real value might be in using LLMs not to prove connections, but to suggest novel research directions that human scientists could then investigate more thoroughly. It's similar to how some scientific discoveries started with intuitive leaps that were later validated through rigorous research.
> You could start with a
reddit
AI Responsibility
1734411754.0
♥ 11
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | unclear |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[
{"id":"rdc_m2g89dw","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"},
{"id":"rdc_m2esap8","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"},
{"id":"rdc_m2dmurk","responsibility":"ai_itself","reasoning":"consequentialist","policy":"unclear","emotion":"outrage"},
{"id":"rdc_m2fa1ui","responsibility":"none","reasoning":"consequentialist","policy":"unclear","emotion":"indifference"},
{"id":"rdc_m2gbqf5","responsibility":"user","reasoning":"consequentialist","policy":"unclear","emotion":"resignation"}
]