I Left Port 22 Open on the Internet for 54 Days. Here's Who Showed Up
30 points by freddyb
30 points by freddyb
The data and the evaluation is interesting. It would be more interesting if this was a human's evaluation. Reads very vibe-written though :-/
Out of 7,556 attacking IPs, only 28 ever opened an interactive shell. That's 0.4%. The rest were pure automation — fingerprint and move on.
But those 31? They're the interesting ones.
Yeah, the AI can't keep its numbers straight.
Like the tags we have for PDFs or videos, a meta tag for vibe-written content is desperately needed on this site at this point.
Do we really want to defined GenAI-produced content as on topic for this site?
I'd rather AI gen content/writing not be on this site at all. But from what I've seen --- it is on here, and there are no rules preventing it from staying, and some users even defend it. It's nuanced too, sometimes some of the content is insightful but just buried under LLM prose.
I only suggested it needs a tag so that those who don't want to see it can just filter it out.
(Instead of having a comment section full of people saying "this smells like AI")
I usually flag worthless GenAI slop as spam.
This was grating to read but had enough meat on its bones to not deserve that flag from me.
If someone lets an LLM write their post (including typical LLM errors/style), doesn't release anonymized data or the code they wrote to fake the Linux environment, how can we even be sure that this whole report isn't an LLM hallucination?
Claude, write an interesting article about leaving port 22 open for 54 days using a virtual honeypot environment.
I've defended some articles in the past that involved LLM writing with a human clearly driving, but I'm flagging this one because it strikes me as insufficiently reviewed for mistakes. Assistance is OK with me; slop is not.
At some point along this curve, even topicality is secondary to matters of style. People just don't want to read LLM-flavored writing. It's repulsive.
I appreciate the article but the use of AI was very very obvious on this one. Some paragraphs were needlessly obtuse and kept reiterating the usual LLM verbal tics.
I appreciate that time was given to looking at the human invaders just poking around. I feel like I need to leave some wholesome memes with suspicious filenames on my VPSs now, just in case someone gets in.