Raw LLM Responses

Inspect the exact model output for any coded comment.

Comment
Copilot, especially Copilot Labs, has actually gotten better at this. While it's not always correct, and knowing how to give a good prompt is critical to getting the kind of results you want, it can generate code and tests. It's especially good for working on a project where you don't know the language or framework well. I'm building a webapp in react right now, and I didn't know react when I started it. Building out features and learning on the go is pretty fluid. Labs let's you basically have a conversation, and generate implementations from features you want. But there's traps here. The same traps you would have with a junior. Code organization is still something you have to do yourself. Sometimes, the implementations are sub-par, especially when you know about a feature coming down the pipe that will make the suggested implementation more difficult to change. It doesn't always do best practices. Hell, sometimes it will just make up shit. But honestly, I'm probably plugged into it for life.
reddit AI Jobs 1687899529.0 ♥ 10
Coding Result
DimensionValue
Responsibilitynone
Reasoningunclear
Policyunclear
Emotionapproval
Coded at2026-04-25T08:33:43.502452
Raw LLM Response
[ {"id":"rdc_jprdwi2","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"outrage"}, {"id":"rdc_jprsjgs","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"rdc_jpsidqx","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"approval"}, {"id":"rdc_jprlhnu","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"indifference"}, {"id":"rdc_jpsbnl5","responsibility":"none","reasoning":"unclear","policy":"unclear","emotion":"outrage"} ]