this post was submitted on 23 Jan 2025

119 points (100.0% liked)

Technology

37919 readers

79 users here now

A nice place to discuss rumors, happenings, innovations, and challenges in the technology sphere. We also welcome discussions on the intersections of technology and society. If it’s technological news or discussion of technology, it probably belongs here.

Remember the overriding ethos on Beehaw: Be(e) Nice. Each user you encounter here is a person, and should be treated with kindness (even if they’re wrong, or use a Linux distro you don’t like). Personal attacks will not be tolerated.

Subcommunities on Beehaw:

This community's icon was made by Aaron Schneider, under the CC-BY-NC-SA 4.0 license.

founded 3 years ago

MODERATORS

TheRtRevKaiser@kbin.social

119

Developer Creates Infinite Maze That Traps AI Training Bots (www.404media.co)

submitted 6 days ago by UrLogicFails to c/technology

15 comments fedilink hide all child comments

Archive.org link

Some key excerpts:

A pseudonymous coder has created and released an open source “tar pit” to indefinitely trap AI training web crawlers in an infinitely, randomly-generating series of pages to waste their time and computing power. The program, called Nepenthes after the genus of carnivorous pitcher plants which trap and consume their prey, can be deployed by webpage owners to protect their own content from being scraped or can be deployed “offensively” as a honeypot trap to waste AI companies’ resources.

The typical web crawler doesn't appear to have a lot of logic. It downloads a URL, and if it sees links to other URLs, it downloads those too. Nepenthes generates random links that always point back to itself - the crawler downloads those new links. Nepenthes happily just returns more and more lists of links pointing back to itself,” Aaron B, the creator of Nepenthes, told 404 Media.

Since they made and deployed a proof-of-concept, Aaron B said their pages have been hit millions of times by internet-scraping bots. On a Hacker News thread, someone claiming to be an AI company CEO said a tarpit like this is easy to avoid; Aaron B told 404 Media “If that’s, true, I’ve several million lines of access log that says even Google Almighty didn’t graduate” to avoiding the trap.

top 15 comments

sorted by: hot top controversial new old

[–] vk6flab@lemmy.radio 29 points 6 days ago* (last edited 6 days ago) (1 children)

This will not work. It sounds great, it sounds plausible, even realistic at some level, but this will not work.

Here's why.

The bot operator has more money than you do. If the efficiency of one bot decreases on one website, they'll throw another bot at it, rinse and repeat until your website stops responding because it's ground to dust.

Meta bots are good at doing this, hitting your site with thousands of requests a second, over and over again.

Meta is not alone in this, but in my experience it's the most destructive.

Source: One of my clients runs a retail website and I've been dealing with this.

At the moment the "best" - least worse is probably more accurate - "solution" is to block them as if they're malicious traffic - which essentially is what they are.

[–] G0ldenSp00n@lemmy.jacaranda.club 15 points 6 days ago

I think part of this software at least in the description on the website is easy and reliable detection of LLM bots to block. You can run it and it will generate statistics about bots that get caught in it so you can easily block big lists of them.

[–] UrLogicFails 18 points 6 days ago (2 children)

As I have gotten more and more exhausted with modern social media, I have contemplated building my own simple website (think neocities), but have been hesitant to put much of myself out as I do not want to feed the AI slop machines.

Nepenthes is encouraging since I could see it being able to protect text in a way "glazing" (ie Nightshade) cannot.

Hopefully AI companies cannot sidestep this as easily as ROBOTS.txt...

[–] vk6flab@lemmy.radio 6 points 6 days ago

If you build a static website, you can host it on AWS S3 and it will likely cost you cents to run. S3 on its own will handle a lot of traffic and you're unlikely to hit that threshold, but if you do, you can add AWS Cloud Front and have a site that can handle more traffic than you can imagine.

[–] ericjmorey 4 points 6 days ago (1 children)

What sort of things might you put on your site?

[–] UrLogicFails 5 points 6 days ago (2 children)

I don't have a full picture idea yet, but I would probably upload art/ projects/ WIPs. I also was considering adding writings, but I'm not sure if there would be enough there to have a section for writings.

I definitely don't have as much to share as others, but I would like to see the Internet become a less centralized space.

An author I follow on Bluesky has been talking about neocities a lot and it seems pretty silly and fun. You probably need to be signed in to view this, but you can see her talk about the process here.

[–] westyvw@lemm.ee 7 points 6 days ago

I still run websites. Some simple and static, others forums, yet others are blogs.

I like to have small communities around simple interests and they get enough traffic to stay interesting.

The blogs are for recording things I do and want to remember and share, like setting up postgres to do interesting things.

None of them have ads or generate revenue. I just miss the old internet and like to see these things exist.

At the same time, I don't really care if AI scrapes it. It's out there to be looked at. Of course one website is a complete farce, illogical and fake product nonsense. I enjoy it getting scraped the most.

[–] ericjmorey 5 points 6 days ago* (last edited 6 days ago) (1 children)

Seems like you'd be interested in https://solar.lowtechmagazine.com/about/

I hope you find a way to share your works.

[–] SplashJackson@lemmy.ca 2 points 5 days ago* (last edited 5 days ago)

Highly recommend this, we've got a few of the compliations from this magazine in print

[–] lvxferre@mander.xyz 16 points 5 days ago (1 children)

This looks interesting. I'd probably combine it with model poisoning - giving each page longer chunks of text, containing bullshit claims and "grammar of slightly brokenness"; so if the data is used to train a model with, the result gets worse.

[–] jherazob 9 points 5 days ago (1 children)

By now i've seen like 6 or 7 projects on either trapping or outright poison content for LLM bots, and yesterday i saw this one which outright modifies the HTML of your page making it harder to steal and instead replaces it with a random prompt, someone asks it to summarize a blogpost and instead the thing starts talking about poodles or something, while normal browser users notice nothing

[–] lvxferre@mander.xyz 4 points 5 days ago (1 children)

Yup, something like this - but for the honeypot, not for the legit pages.

[–] jherazob 5 points 5 days ago

I think many of them already do, i know Iocaine uses Markov chain text generation to spit out nonsense to poison the LLMs, do check it, at the bottom of the project page it links to others

[–] pine@lemmy.blahaj.zone 12 points 6 days ago (1 children)

I stumbled onto a page like this a few years ago and it simply left me baffled and confused. It's offline now but I finally understand what it was doing. https://web.archive.org/web/20230820024137/https://cjb.com/

[–] AstralPath@lemmy.ca 1 points 5 days ago

My fav link in there is "Poop in my freshman dorm experience is essential."