frontpage.
newsnewestaskshowjobs

Open Source @Github

fp.

Open in hackernews

Who is cleaning up all the garbage LLMs generate?

6•kbrannigan•2d ago
So my one terabyte drive got full and I realized after months of CLI sessions I generated a lot of markdown files in a lot of images, all sort of other documents they all seem sorta kinda useful, and I too afraid to delete them because they might break something

How do you deal with that?

Comments

faykn•2d ago
I just delete it and go back to the commit if it matters in the future or generally if the code changed a lot just ask the probs newer smarter model to make the file again
kbrannigan•2d ago
Hmm, hence the issue. It’s like chopping down a tree to make a box of matches. I wish the next generations of models don’t rely too much on markdown documentations which seems to be a human way of not writing code. Sometimes the code is just a little bit more structured to be honest , often times I feel like I might as well just write the code.
CBLT•2d ago
Deleting files is easy. Decommissioning production services is not. It's too early to say for me, we've only decommissioned one vibe-coded service at my Day Job and it wasn't any worse than decommissioning any other legacy service has been. But it became legacy much faster - a couple of months.
kbrannigan•2d ago
What does that mean. Taking them down because the project failed or just replacing it?
CBLT•2d ago
Replacement. Getting all customer requirements and migrating them to a superior service implementation.
Loren_SL•1d ago
Have agents write scratch output outside the repo, while anything that matters lives in version control. That way you can purge the scratch files whenever you like.
desireco42•1h ago
Legitimate question. In my projects, where I do have a ton of files, but kind of suspect not nearly as you have, I do occasionally instruct agent to review old files and decisions and archive them or delete.

Sometimes it is useful to know how we came about some decision or feature, but in general you often don't need them and they would contain prose that would confuse agent. So you remove it.

I don't use Claude Code mostly because it always loved to generate reports and there is no way I will read it's lengthy reports. I try to have documents I can actually read if I want to.

Hope this helps, but some lifecycle of your documents and requirements is useful to have.

Ask HN: Are there AI models for generating sounds based on a text and reference?

22•onemiketwelve•1d ago•12 comments

Ask HN: What do you think about Fractional CFOs

3•mtmosestn•1h ago•3 comments

Ask HN: Why is Ask HN only showing me 14 posts?

40•Gooblebrai•5h ago•37 comments

Ask HN: What Is Your Personal Backup Methods

2•Crontab•1h ago•0 comments

Ask HN: What do you run on a $5 VPS that's worth keeping online 24/7?

11•mariocesar•3h ago•4 comments

Ask HN: What games do you religiously play on your phone browser?

6•saimiam•3h ago•13 comments

Ask HN: Do you still own a printer?

9•nunorbatista•8h ago•14 comments

Ask HN: How to deal with AI "true believer" leadership at work

5•microflash•9h ago•3 comments

Ask HN: How do you describe what your relationship brings you?

2•maxignol•7h ago•3 comments

Ask HN: What would you like to see in a new AI technology release?

2•ikishade•8h ago•2 comments

Ask HN: Vibecoding Follies

4•vegnus•10h ago•2 comments

Ask HN: How come everyone is an LLM expert?

3•delis-thumbs-7e•10h ago•3 comments

ChatGPT self-distances when admitting fault

3•chrisjj•4h ago•1 comments

Tell HN: GitHub refuses to remove cracked copies of my software after a month

54•IvanK_net•8h ago•52 comments

Ask HN: Show me your agentic SDE semantics

2•danielovichdk•12h ago•0 comments

Ask HN: Is observability broken for you?

2•tmach32•14h ago•0 comments

Ask HN: My Apple ID is locked for a week, Apple is a single point of failure

11•akg_67•15h ago•3 comments

Ask HN: Agent access to chat data should require participant consent?

2•maxwellito•17h ago•0 comments

Ask HN: Alternatives to NeurIPS, ICLR and ICML

3•john-titor•19h ago•0 comments

Ask HN: What brings you back to personal AI agents like Instinct and Muse?

3•sdrth•21h ago•1 comments

Orcah Studio: A local-first video agent that can search your videos

8•iliashad•1d ago•5 comments

Ask HN: Why does Astra compact context so frequently vs. Fable?

2•yesitcan•1d ago•0 comments

Ask HN: What do you think about AI generated slides for conferences?

4•Hixon10•1d ago•1 comments

Ask HN: How are AI budgets changing in your company?

5•bobby-cb•1d ago•2 comments

Tell HN: 2026 is the year of the Linux desktop, agent-adjusted

3•dvrp•1d ago•2 comments

Tell HN: Uceprotect is extorting website owners

101•goldenmember•1d ago•59 comments

Our digital privacy is being aggressively disbanded this week

12•Steaglsz•1d ago•1 comments

DOS Game Stunts Port to Linux,Windows,Browser,etc.

4•LowLevelMahn•2d ago•1 comments

Who is cleaning up all the garbage LLMs generate?

6•kbrannigan•2d ago•7 comments

Ask HN: Anyone accepted into OpenAI "Codex for open source"?

4•awb•1d ago•0 comments