Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
Seems like if you want to redraw ai generated art, you should be aiming to make …
ytc_Ugw4jjncm…
G
Ai cooked man's brain 😂 then the YouTube channel proceeds to use ai for thumb na…
ytc_UgwGTj0Tu…
G
Electronic sensors will not achieve the resolution of human senses, nor will rob…
ytc_UgwevwzXE…
G
The car should be automated as it seems to be getting safer than human driving (…
ytc_UgiwJ2F2f…
G
Watching this video is inspiring but also a bit daunting ‚ most of us can't affo…
ytc_Ugw9xL_zb…
G
Suggestion: Bring up the fact that a country has given AI citizenship, and that …
ytc_UgyrVLf96…
G
If millions and even billions will put there consciousness in the AI it will eve…
ytc_UgyVd2stj…
G
There are bad things about it , for sure ! To use AI to falsely present persons…
ytc_UgyB5nN70…
Comment
Yep, found an issue with something I was testing yesterday where if I opened something on a page one way I got the new thing I added, but if I opened it another way it didn't show up unless I clicked something else on the page after it loaded.
Fired up Claude, told it the problem ("If I get to it this way I see the thing, but if I get to it this other way I don't see it unless I click on this other entry on the page first."), let it cook for a bit, and it found the issue, explained why it happened, and proposed a fix. Wasn't entirely satisfied with its suggestion so I asked why it settled on that, and it came up with a better one that I told it to implement. The plugin I'm using ran a quick review of the changes and said it all looks good but notes this increases the number of calls to an API and caching the value after the first call would be a good next step, so I tell it to do that too.
Project builds, so I run it locally to verify I get the result I want, put it up for review, merge it, push it out to where I was testing, and confirm it's fixed and we're good to go.
I've been at this for 15 years so I know how to debug and eval fixes and whatnot, but Claude definitely found the issue quickly, proposed a fix, iterated on that fix when I wasn't as happy with it, and got it implemented in like...15 minutes? I've encountered plenty of juniors in my time who haven't really touched a debugger before, so while I can understand OP's frustration about how they're getting praise for shipping features quickly, I'm a little less with them on being miffed about juniors not being good at debugging. That's absolutely a skill that develops over time and reps.
reddit
AI Jobs
1777041258.0
♥ 9
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | none |
| Emotion | approval |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[{"id":"rdc_oi090qg","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"indifference"},{"id":"rdc_oi09evs","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},{"id":"rdc_oi0f6e2","responsibility":"developer","reasoning":"virtue","policy":"none","emotion":"mixed"},{"id":"rdc_oi0rb5s","responsibility":"none","reasoning":"unclear","policy":"none","emotion":"approval"},{"id":"rdc_ohzhtsv","responsibility":"developer","reasoning":"unclear","policy":"none","emotion":"outrage"}]