Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Show HN: AISlop, a CLI for catching AI generated code smells

github.com

Post reply on HN
Print view View original

51–60 of 70 posts

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#51

Post by smj-edison »

cityofdelusion wrote: post #50

I’m eager to test this out. I have agent instructions to try to limit the worst of this already, but patterns still sneak through. I have a review agent run after every single edit looking for all of the following if you need more ideas for checks: - DRY principle violations, multiple definitions of the same helpers or utilities. - Changes that deviate from existing patterns and architecture already in the code, espe…

Oh my gosh, the guard code drives me crazy. In try so hard to get Kimi to put in asserts instead of silently swallowing corrupt values, but it keeps handling bad values poorly instead of crashing. I've even explicitly put in CLAUDE.md that correctness is more important then continuing to run, but it still keeps defensively programming when it should loudly crash.

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#52

Post by genghisjahn »

stymaar wrote: post #26
genghisjahn wrote: post #19

Earlier quoted context omitted.

7 and 8 are left as an exercise for the reader.

I have the implementation for languages 7 and 8 but it's too big to fit in this comment section.

I'm just joking. This is cool stuff you made.

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#54

Post by explosion-s »

Der_Einzige wrote: post #7

Related, we made "Antislop", a framework for removing repetitive phraseology from LLMs: https://arxiv.org/abs/2510.15061

This is a super cool paper! The only thing I'm a bit skeptical of is the detection of 'slop' to begin with - it would be very exciting to see some sort of way to steer a model more towards human-like output more generally. Here I imagine a bit of a challenge is that you're 1) creating the measurement for 'slop' and then 2) reducing it, so yes, you can have amazing results in the detection of the slop you've found (the methodology may be novel - a slight improvement over DPO), but this is a chicken and egg problem, as each new model would suffer differently based on training, RLHF, etc. It's not a very useful method to improve an LLM's writing quality, but rather just reducing the symptoms of said bad writing.

I'm additionally concerned about removing the LLM's voice to begin with, as though an LLM may far overuse specific words and phrases, so too do individual authors. Reducing an LLM towards the average author would result in an 'average' voice, paradoxically unlike any specific author.

I am currently doing research on something very similar over the summer but more towards the detection as opposed to the generation side - I'd love to discuss this with someone in the field if you had a few minutes sometime!

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#55

Post by Heavykenny »

cityofdelusion wrote: post #50

I’m eager to test this out. I have agent instructions to try to limit the worst of this already, but patterns still sneak through. I have a review agent run after every single edit looking for all of the following if you need more ideas for checks: - DRY principle violations, multiple definitions of the same helpers or utilities. - Changes that deviate from existing patterns and architecture already in the code, espe…

This is so good and comprehensive. I will be adding this to the list of patterns to check. Thank you so much, really appreciate

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#57

Post by Heavykenny »

beart wrote: post #32
Retr0id wrote: post #28

Earlier quoted context omitted.

Forgot to switch accounts?

OP seems to be an AI account. > Don't post generated comments or AI-edited comments. HN is for conversation between humans.

It get hard trying to proof that I’m human.

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#58

Post by vunderba »

Nice job. Somewhat related, but one "AI smell" (if you can call it that) is LLMs' desire to add layers of redundant safeguards with things like null coalescing operators and defensive fallbacks so that there’s always some default value. LLMs often seem to struggle with the distinction between happy/unhappy paths and often end up treating everything as optional and "safe".

I’m a much bigger fan of surfacing problems like missing values immediately and loudly so they bubble up, particularly during early prototyping and development.

I’ve collected a set of best practices in my AGENTS.md that covers issues like this, and that’s helped a fair bit. Obviously, you want to offload as much as possible to a deterministic linter, but this pattern is a bit tougher to catch in post.

Re: Show HN: AISlop, a CLI for catching AI generated code smells

#59

Post by stymaar »

genghisjahn wrote: post #52
stymaar wrote: post #26

Earlier quoted context omitted.

I have the implementation for languages 7 and 8 but it's too big to fit in this comment section.

I'm just joking. This is cool stuff you made.

I'm too belong to the joking crowd in this thread, I have no involvement with the team that did this project.
Post reply on HN