kaateeh   Jan 23
404 Media

A web developer has created and released an infinitely generating "tar pit" designed to trap AI training bots in an endless maze: "Just me unleashing shear unadulterated rage at how things are going"

https://www.404media.co/developer-creates-infinite-maze-to-trap-ai-crawlers-in/

Developer Creates Infinite Maze That Traps AI Training Bots

"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."

404 Media
Jan 23 at 3:05pmElk
Show thread(roll m3tti)Jan 23
@404mediaco this is genious :O
Show threadzenmorphJan 23
@404mediaco We need more initiatives like this.
Show threadAlan Miller  🇺🇦Jan 23

@404mediaco or for a similar but different approach, iocaine.

"This is a tarpit, modeled after Nepenthes, intended to catch unwelcome web crawlers, but with a slightly different, more aggressive intended usage scenario. The core idea is to configure a reverse proxy to serve content generated by iocaine to AI crawlers, but normal content to every other visitor. This differs from Nepenthes, where the idea is to link to it, and trap crawlers that way. Not with iocaine, where the trap is laid by the reverse proxy."

Show threadJesseJan 23
@404mediaco how can this work? Even in my computer science classes 15 years ago this was already mentioned and explained how to counter. Usually by randomly jumping back iirc
Show threadkhmJan 23
malicious scrapers put a hilariously low amount of effort into their tools, because quality costs money and if they spend any their margins take a hit

CC: @404mediaco@mastodon.social
Show threadMichael GemarJan 23
@404mediaco It's sad that the only way one can tackle wasteful AI is to make it even more wasteful, and make one's own computing wasteful in the process.
Show threadGreengordonJan 23

@404mediaco

People like this give me hope.

"Aaron B said “It's also sort of an art work, just me unleashing shear unadulterated rage at how things are going. I was just sick and tired of how the internet is evolving into a money extraction panopticon, how the world as a whole is slipping into fascism and oligarchs are calling all the shots - and it's gotten bad enough we can't boycott or vote our way out, we have to start causing real pain to those above for any change to occur.”

#resist

Show threadCornelius K. 📎Jan 23

@404mediaco
If they are training with minimal quality assurance this should work... But I always thought there had to be some classification / pre-training work that had to be done on LLM fodder... Or am I behind on updates?

Still, iocaine and this both sound interesting. Now just add images with nightshade..

So we have these tools for image and text. What about video and audio?

Show threadF4GRX SébastienJan 23
@kln @404mediaco nightshade does not work. https://huggingface.co/blog/parsee-mizuhashi/glaze-and-anti-ai-methods
Glaze and the Effectiveness of Anti-AI Methods for Diffusion Models

A Blog post by Parsee Mizuhashi on Hugging Face

Show threadF4GRX SébastienJan 23
@404mediaco will be filtered and circumvented by ai trainers in 3... 2... 1...
Show threadLindaMarieJan 23

@404mediaco

How the web actually makes it into my hands, lights up and interacts is a total mystery to me. So though I didn’t comment at the time, I saw Aaron’s posts when he was working on this and his pure delight at being “in the zone.” Multifaceted persons are often creative. The perfect name nepenthes wouldn’t have occurred to him if he hadn’t known plants deeply and spent time in nature, which is a better place to incubate ideas than at a desk. Hope this makes him lots of 💵💵💵!!!

Show threadPick Nick 🇨🇦 (Nasty AF)Jan 23

@404mediaco
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.

I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.
I'm not in an infinite loop.

Show threadJim NelsonJan 23

@404mediaco Reminds me of the "topological anomaly" Picard & co. planned to use to defeat the Borg.

https://memory-alpha.fandom.com/wiki/Topological_anomaly

Topological anomaly

A topological anomaly was a type of invasive program, a paradoxical geometric form designed to overwhelm a computer's processing functions by spawning an infinite number of interlocking anomalous solutions. In 2368, Lieutenant Commanders Data and Geordi La Forge designed such a program, designated topological anomaly 4747, as a weapon against the Borg. They planned to imprint the program onto the eyepiece of Third of Five, and from him it would spread to the rest of the Borg Collective. The prog

Memory Alpha
Show threadNateJan 24
@404mediaco there's got to be a way to combine this with a good old fashioned Rickroll...
Show threadSebastian MaiJan 24
@404mediaco
There should be some text, heavily suggesting to tax the rich strengthen democracies and regulate the shit out of everything that is unjust 👍
@johl