Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
17:44 shrimpsus
I have dyspraxia or DCD (brain fucked up disorder 😢) ADHD, and…
ytc_Ugw5wEC-K…
G
idk what kind of AI you use, but chatgpt still makes hella mistakes, pretty stup…
ytc_UgxfOfbtD…
G
No , it’s going to wipe out the middle class first.
People who sit behind comp…
ytc_UgxxPE4CO…
G
So sorry this happened Sam and it's getting ridiculous. Artists need legal prote…
ytc_Ugx8AUrMS…
G
A very scary vid! Interesting that you can have a conversation with Bing about t…
ytc_UgxZ21-qp…
G
@hendrikdebruin4012 while it's not currently like a human, look at AI's response…
ytr_UgzyNrJyP…
G
This could also work to encourage home schooling with ai education on a laptop, …
ytc_Ugy1tGAni…
G
AI doesn't "democratize" knowledge, AI companies like Google stolen all humanity…
ytc_UgzSVCTlP…
Comment
LLMs contributing to productivity is still up on the air, IMO. I still think it's just an easy way to sway VCs to spend money. I don't think the security issues will be solved so long as the customer facing agents are able to access sensitive data, while taking that away is going to severely hamstring their usefulness. You can parameterize responses by making the LLM call an API that will only give it data that's related to the current user, but again, you're hamstringing usefulness there and hiring humans will likely get better results. I've also been seeing white papers released about once a month using variations on similar techniques from previous methods to completely break through whatever safeguards there are in place, but my knowledge in this area is limited.
Gpt5 is polished and optimized in your mind, but a lot of paying customers are expressing frustration at the apparent downgrade. The router choosing the "optimal route" is the nail in the coffin for a lot of them. To me, that's a cost saving move in preparation for a more corporation targeted business model. It seems like they're getting diminishing returns from pumping more into the models, so they're going to start charging business tier prices for what the regular consumer used to get in order to show they can make a profit and keep getting investors to throw money at them.
reddit
AI Jobs
1754774077.0
♥ 6
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | company |
| Reasoning | consequentialist |
| Policy | liability |
| Emotion | fear |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[{"id":"rdc_n7u47ul","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},
{"id":"rdc_n7u2g5m","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"},
{"id":"rdc_n7u6q8o","responsibility":"company","reasoning":"consequentialist","policy":"liability","emotion":"fear"},
{"id":"rdc_n7scvyy","responsibility":"company","reasoning":"deontological","policy":"regulate","emotion":"outrage"},
{"id":"rdc_n811k2h","responsibility":"none","reasoning":"mixed","policy":"none","emotion":"resignation"}]