Raw LLM Responses
Inspect the exact model output for any coded comment.
Look up by comment ID
Random samples — click to inspect
G
>Additionally, how does it apply to parts of a work? Eg. Imagine all the back…
rdc_jwvc6wx
G
Every person in the comments still trying to defend the shitty handling of gener…
ytc_UgzkKHjCx…
G
Data Center Lobbyists and Car salesman, when their Lips are moving, they are Lyi…
ytc_Ugxw1lHlC…
G
But.... They're not doing the art!!! I can't draw for 💩💩 but I think AI is just …
ytc_UgxY99JPV…
G
tbh as far as I know about AI they will do a task if they get rewarded for it s…
ytc_Ugy5inbXk…
G
she got probably a huge sense of relief by walking away from trashy associates. …
ytr_UgzcWUICc…
G
38:57 I hate when Neil interrupts a solid point someone is making to ask a stupi…
ytc_UgwFHYg7H…
G
I'm a musician, I don't feel like people are required to listen to music I make …
ytc_UgyDwbPtC…
Comment
Why are AI companies obsessed with replacing software engineers?
> So why are companies so focused on replacing software engineers first?? Why are the first AI agents coming out "AI programmers"?
This is a poor way to frame the question. AI has been a field of study for decades
To answer your question though:
- There is a huge amount of data on which to train coding models
- Unlike other natural language data that models can be trained on, source code by design, shows the train of thought! The algorithms are laid out step by step rather than just giving an answer. This is gold for training reasoning models
- With code it is possible to generate synthetic data for training purposes more easily than other things
- Code is also more amenable to reinforcement learning as you can test what has been generated for correctness - does the code compile, run and pass the relevant unit test?
- There is a huge incentive for a frontier AI lab to build better coding models because it helps accelerate their own development. Coding along with AI research are two areas where once a critical threshold is met, can then unlock self-improvement
> AI is poorly suited for traditional software engineering. It lacks the ability to understand codebase context, handle complex system design, or resolve ambiguous requirements—key parts of an engineer’s job. While it performs well on well-defined tasks like coding challenges, it fails with the nuanced, iterative problem-solving real-world development requires.
That is today. With improvements in context, memory etc. these things are all possible to apply AI
> Yet, unlike many mindless desk jobs, or even traditional IT jobs, software engineers seem to be the primary target for AI replacement. Why?? It feels like they just want to get rid of us at this point imo
For decades, software engineering has been used to build software that ultimately replaced non-digital solutions and often displaced other roles. Now you get a taste of w
reddit
AI Jobs
1736784958.0
♥ 1
Coding Result
| Dimension | Value |
|---|---|
| Responsibility | none |
| Reasoning | unclear |
| Policy | none |
| Emotion | indifference |
| Coded at | 2026-04-25T08:33:43.502452 |
Raw LLM Response
[{"id":"rdc_m6xpy55","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"approval"},{"id":"rdc_m6xqp5f","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"approval"},{"id":"rdc_m6xrkfs","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"mixed"},{"id":"rdc_m6xssbp","responsibility":"company","reasoning":"consequentialist","policy":"none","emotion":"outrage"},{"id":"rdc_m6xtorr","responsibility":"none","reasoning":"consequentialist","policy":"none","emotion":"indifference"}]